jackkuo commited on
Commit
e54c585
·
verified ·
1 Parent(s): 0977f72

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +49 -0
  2. 09E0T4oBgHgl3EQfdQDB/content/tmp_files/2301.02375v1.pdf.txt +1519 -0
  3. 09E0T4oBgHgl3EQfdQDB/content/tmp_files/load_file.txt +0 -0
  4. 0NFQT4oBgHgl3EQfDTUM/content/tmp_files/2301.13233v1.pdf.txt +1466 -0
  5. 0NFQT4oBgHgl3EQfDTUM/content/tmp_files/load_file.txt +0 -0
  6. 1NAzT4oBgHgl3EQfDfoM/vector_store/index.faiss +3 -0
  7. 1dE0T4oBgHgl3EQfdgBe/content/tmp_files/2301.02377v1.pdf.txt +1126 -0
  8. 1dE0T4oBgHgl3EQfdgBe/content/tmp_files/load_file.txt +304 -0
  9. 2NE0T4oBgHgl3EQfugGB/vector_store/index.faiss +3 -0
  10. 2dFIT4oBgHgl3EQf5Cs5/content/2301.11387v1.pdf +3 -0
  11. 2tFQT4oBgHgl3EQfGDUx/vector_store/index.pkl +3 -0
  12. 6NFKT4oBgHgl3EQf_C4s/content/tmp_files/2301.11960v1.pdf.txt +482 -0
  13. 6NFKT4oBgHgl3EQf_C4s/content/tmp_files/load_file.txt +0 -0
  14. 8NFLT4oBgHgl3EQfBC4o/content/tmp_files/2301.11968v1.pdf.txt +792 -0
  15. 8NFLT4oBgHgl3EQfBC4o/content/tmp_files/load_file.txt +0 -0
  16. 9dE3T4oBgHgl3EQfrApq/content/tmp_files/2301.04656v1.pdf.txt +0 -0
  17. 9dE3T4oBgHgl3EQfrApq/content/tmp_files/load_file.txt +0 -0
  18. 9tFQT4oBgHgl3EQfJjU4/content/2301.13256v1.pdf +3 -0
  19. C9AzT4oBgHgl3EQfwf5z/content/tmp_files/2301.01723v1.pdf.txt +846 -0
  20. C9AzT4oBgHgl3EQfwf5z/content/tmp_files/load_file.txt +457 -0
  21. CdE2T4oBgHgl3EQf9AmF/vector_store/index.pkl +3 -0
  22. DNFQT4oBgHgl3EQfPTY7/content/tmp_files/2301.13278v1.pdf.txt +1485 -0
  23. DNFQT4oBgHgl3EQfPTY7/content/tmp_files/load_file.txt +0 -0
  24. EdE0T4oBgHgl3EQfywKq/content/tmp_files/2301.02664v1.pdf.txt +636 -0
  25. EdE0T4oBgHgl3EQfywKq/content/tmp_files/load_file.txt +471 -0
  26. EdE1T4oBgHgl3EQf-gYV/content/tmp_files/2301.03568v1.pdf.txt +1867 -0
  27. EdE1T4oBgHgl3EQf-gYV/content/tmp_files/load_file.txt +0 -0
  28. F9E0T4oBgHgl3EQfRAC9/vector_store/index.pkl +3 -0
  29. FtAyT4oBgHgl3EQf5Poe/vector_store/index.faiss +3 -0
  30. FtAyT4oBgHgl3EQf5Poe/vector_store/index.pkl +3 -0
  31. G9AyT4oBgHgl3EQfS_dT/content/2301.00096v1.pdf +3 -0
  32. HNAyT4oBgHgl3EQfrvkK/vector_store/index.faiss +3 -0
  33. INAzT4oBgHgl3EQfxv6_/content/tmp_files/2301.01744v1.pdf.txt +0 -0
  34. INAzT4oBgHgl3EQfxv6_/content/tmp_files/load_file.txt +0 -0
  35. KdE2T4oBgHgl3EQfAgZ0/content/tmp_files/2301.03592v1.pdf.txt +1427 -0
  36. KdE2T4oBgHgl3EQfAgZ0/content/tmp_files/load_file.txt +0 -0
  37. KtAyT4oBgHgl3EQf6Prq/content/2301.00820v1.pdf +3 -0
  38. KtAyT4oBgHgl3EQfsfmL/vector_store/index.faiss +3 -0
  39. KtE1T4oBgHgl3EQfsgWq/content/tmp_files/load_file.txt +540 -0
  40. LdAyT4oBgHgl3EQf6fqP/content/2301.00823v1.pdf +3 -0
  41. LdAyT4oBgHgl3EQf6fqP/vector_store/index.pkl +3 -0
  42. LtE1T4oBgHgl3EQftAUX/content/tmp_files/2301.03371v1.pdf.txt +3133 -0
  43. LtE1T4oBgHgl3EQftAUX/content/tmp_files/load_file.txt +0 -0
  44. MdE4T4oBgHgl3EQfjA06/content/2301.05138v1.pdf +3 -0
  45. MdE4T4oBgHgl3EQfjA06/vector_store/index.faiss +3 -0
  46. MdE4T4oBgHgl3EQfjA06/vector_store/index.pkl +3 -0
  47. NNFJT4oBgHgl3EQfzi34/content/tmp_files/2301.11644v1.pdf.txt +1013 -0
  48. NNFJT4oBgHgl3EQfzi34/content/tmp_files/load_file.txt +0 -0
  49. OdAzT4oBgHgl3EQfzf5C/content/tmp_files/2301.01769v1.pdf.txt +1348 -0
  50. OdAzT4oBgHgl3EQfzf5C/content/tmp_files/load_file.txt +0 -0
.gitattributes CHANGED
@@ -2001,3 +2001,52 @@ WNAzT4oBgHgl3EQfmP1C/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -tex
2001
  8dAzT4oBgHgl3EQf-v5v/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2002
  s9AyT4oBgHgl3EQf0PmZ/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2003
  G9AyT4oBgHgl3EQfS_dT/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2001
  8dAzT4oBgHgl3EQf-v5v/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2002
  s9AyT4oBgHgl3EQf0PmZ/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2003
  G9AyT4oBgHgl3EQfS_dT/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2004
+ vdFPT4oBgHgl3EQfOTSJ/content/2301.13033v1.pdf filter=lfs diff=lfs merge=lfs -text
2005
+ fNFMT4oBgHgl3EQf1TG9/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2006
+ tNAzT4oBgHgl3EQfr_0m/content/2301.01651v1.pdf filter=lfs diff=lfs merge=lfs -text
2007
+ fNFMT4oBgHgl3EQf1TG9/content/2301.12440v1.pdf filter=lfs diff=lfs merge=lfs -text
2008
+ ctAzT4oBgHgl3EQf3P5i/content/2301.01826v1.pdf filter=lfs diff=lfs merge=lfs -text
2009
+ l9FST4oBgHgl3EQfJzgq/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2010
+ ptE1T4oBgHgl3EQf2QVA/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2011
+ 2dFIT4oBgHgl3EQf5Cs5/content/2301.11387v1.pdf filter=lfs diff=lfs merge=lfs -text
2012
+ s9AyT4oBgHgl3EQf0PmZ/content/2301.00715v1.pdf filter=lfs diff=lfs merge=lfs -text
2013
+ ptE1T4oBgHgl3EQf2QVA/content/2301.03476v1.pdf filter=lfs diff=lfs merge=lfs -text
2014
+ mNAyT4oBgHgl3EQfyvkz/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2015
+ y9FQT4oBgHgl3EQfyjYu/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2016
+ mdFKT4oBgHgl3EQfES1N/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2017
+ y9FQT4oBgHgl3EQfyjYu/content/2301.13409v1.pdf filter=lfs diff=lfs merge=lfs -text
2018
+ mNAyT4oBgHgl3EQfyvkz/content/2301.00688v1.pdf filter=lfs diff=lfs merge=lfs -text
2019
+ KtAyT4oBgHgl3EQf6Prq/content/2301.00820v1.pdf filter=lfs diff=lfs merge=lfs -text
2020
+ UdE0T4oBgHgl3EQflgEz/content/2301.02486v1.pdf filter=lfs diff=lfs merge=lfs -text
2021
+ 1NAzT4oBgHgl3EQfDfoM/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2022
+ FtAyT4oBgHgl3EQf5Poe/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2023
+ G9AyT4oBgHgl3EQfS_dT/content/2301.00096v1.pdf filter=lfs diff=lfs merge=lfs -text
2024
+ o9AzT4oBgHgl3EQfqv2W/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2025
+ tNAzT4oBgHgl3EQfr_0m/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2026
+ kdE2T4oBgHgl3EQfdgef/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2027
+ ydAzT4oBgHgl3EQf7_6d/content/2301.01898v1.pdf filter=lfs diff=lfs merge=lfs -text
2028
+ 9tFQT4oBgHgl3EQfJjU4/content/2301.13256v1.pdf filter=lfs diff=lfs merge=lfs -text
2029
+ qdFST4oBgHgl3EQfOjiX/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2030
+ Y9AyT4oBgHgl3EQfvvk3/content/2301.00635v1.pdf filter=lfs diff=lfs merge=lfs -text
2031
+ MdE4T4oBgHgl3EQfjA06/content/2301.05138v1.pdf filter=lfs diff=lfs merge=lfs -text
2032
+ MdE4T4oBgHgl3EQfjA06/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2033
+ d9E1T4oBgHgl3EQfeAT4/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2034
+ KtAyT4oBgHgl3EQfsfmL/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2035
+ stAyT4oBgHgl3EQf0Pls/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2036
+ 2NE0T4oBgHgl3EQfugGB/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2037
+ Y9AyT4oBgHgl3EQfvvk3/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2038
+ ctAzT4oBgHgl3EQf3P5i/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2039
+ wtFAT4oBgHgl3EQfjR1k/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2040
+ HNAyT4oBgHgl3EQfrvkK/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2041
+ wtFAT4oBgHgl3EQfjR1k/content/2301.08604v1.pdf filter=lfs diff=lfs merge=lfs -text
2042
+ LdAyT4oBgHgl3EQf6fqP/content/2301.00823v1.pdf filter=lfs diff=lfs merge=lfs -text
2043
+ WNAzT4oBgHgl3EQfmP1C/content/2301.01559v1.pdf filter=lfs diff=lfs merge=lfs -text
2044
+ ktFLT4oBgHgl3EQfdC-Z/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2045
+ ktFLT4oBgHgl3EQfdC-Z/content/2301.12085v1.pdf filter=lfs diff=lfs merge=lfs -text
2046
+ s9E2T4oBgHgl3EQf1wgV/content/2301.04153v1.pdf filter=lfs diff=lfs merge=lfs -text
2047
+ ydAzT4oBgHgl3EQf7_6d/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2048
+ f9AzT4oBgHgl3EQfof2V/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
2049
+ PNFOT4oBgHgl3EQf4TQo/content/2301.12949v1.pdf filter=lfs diff=lfs merge=lfs -text
2050
+ s9AzT4oBgHgl3EQfPfs9/content/2301.01183v1.pdf filter=lfs diff=lfs merge=lfs -text
2051
+ m9E3T4oBgHgl3EQfKgnW/content/2301.04355v1.pdf filter=lfs diff=lfs merge=lfs -text
2052
+ m9E3T4oBgHgl3EQfKgnW/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
09E0T4oBgHgl3EQfdQDB/content/tmp_files/2301.02375v1.pdf.txt ADDED
@@ -0,0 +1,1519 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Centralized Cooperative Exploration Policy for Continuous
2
+ Control Tasks
3
+ Chao Li*
4
+ Institute of Automation, Chinese
5
+ Academy of Sciences, China
6
7
+ Chen Gong∗
8
+ Institute of Automation, Chinese
9
+ Academy of Sciences, China
10
11
+ Qiang He
12
+ University of Tubingen, Germany
13
14
+ Xinwen Hou†
15
+ Institute of Automation, Chinese
16
+ Academy of Sciences, Beijing, China
17
18
+ Yu Liu
19
+ Institute of Automation, Chinese
20
+ Academy of Sciences, Beijing, China
21
22
+ ABSTRACT
23
+ The deep reinforcement learning (DRL) algorithm works brilliantly
24
+ on solving various complex control tasks. This phenomenal success
25
+ can be partly attributed to DRL encouraging intelligent agents to
26
+ sufficiently explore the environment and collect diverse experiences
27
+ during the agent training process. Therefore, exploration plays a
28
+ significant role in accessing an optimal policy for DRL. Despite
29
+ recent works making great progress in continuous control tasks,
30
+ exploration in these tasks has remained insufficiently investigated.
31
+ To explicitly encourage exploration in continuous control tasks,
32
+ we propose CCEP (Centralized Cooperative Exploration Policy),
33
+ which utilizes underestimation and overestimation of value func-
34
+ tions to maintain the capacity of exploration. CCEP first keeps two
35
+ value functions initialized with different parameters, and generates
36
+ diverse policies with multiple exploration styles from a pair of value
37
+ functions. In addition, a centralized policy framework ensures that
38
+ CCEP achieves message delivery between multiple policies, fur-
39
+ thermore contributing to exploring the environment cooperatively.
40
+ Extensive experimental results demonstrate that CCEP achieves
41
+ higher exploration capacity. Empirical analysis shows diverse ex-
42
+ ploration styles in the learned policies by CCEP, reaping benefits in
43
+ more exploration regions. And this exploration capacity of CCEP
44
+ ensures it outperforms the current state-of-the-art methods across
45
+ multiple continuous control tasks shown in experiments.
46
+ KEYWORDS
47
+ Deep Reinforcement Learning, Cooperative Exploration, Continu-
48
+ ous Control Tasks
49
+ 1
50
+ INTRODUCTION
51
+ Deep reinforcement learning (DRL) [46], which utilizes deep neu-
52
+ ral networks to learn an optimal policy, works brilliantly and has
53
+ demonstrated beyond human-level performance in solving various
54
+ challenging sequential decision-making tasks e.g., video games [17,
55
+ 34, 35, 52, 54], autonomous driving [18, 53], robotic control tasks [2,
56
+ 31], etc. In DRL settings, an agent needs to sufficiently explore the
57
+ environment and collect a set of experiences to obtain an optimal
58
+ policy. The agent aims to learn an optimal policy to maximize its ex-
59
+ pected cumulative rewards through trial and error. Therefore, DRL
60
+ ∗These two authors contributed equally.
61
+ †The corresponding author.
62
+ can be regarded as learning from reward feedback from environ-
63
+ ments. It is essential that during the training phase the agent should
64
+ be encouraged to explore the environments and gather sufficient
65
+ reward signals for well-training.
66
+ In DRL, exploration has obsessed with a critical problem: submit-
67
+ ting solutions too quickly without sufficient exploration, leading
68
+ to getting stuck at local minima or even complete failure. DRL
69
+ researchers adopt the neural network to yield the policy with sig-
70
+ nificant feature extraction and expression capabilities in a range of
71
+ continuous control tasks. Whereas this phenomenal practice has
72
+ achieved great performance, it is still obsessed with the notorious
73
+ insufficient exploration problem in continuous control tasks. Good
74
+ exploration becomes extremely difficult when the environment
75
+ is distracting or provides little feedback. Whereas existing explo-
76
+ ration methods remain a problematic drawback – lacking diversity
77
+ to explore. The classic exploration methods such as 𝜖-Greedy strate-
78
+ gies [35] or Gaussian noise [31] indirectly and implicitly change
79
+ the style of the exploration. However, in massive situations, diverse
80
+ styles of exploration are necessary. For instance, in chess games,
81
+ players should perform different styles of policies (e.g., radical, con-
82
+ servative, etc.) to keep competitive when facing various situations;
83
+ humanoid robots attempt diverse control styles and eventually learn
84
+ to walk efficiently.
85
+ Recent studies enrich diverse styles of policies by construct-
86
+ ing relationships between the distribution of policy and trajecto-
87
+ ries [1, 13, 26, 27, 43]. In Diversity is All You Need (DIAYN) [11],
88
+ authors highlight the diversity of policies plays a significant role in
89
+ well-training agents. It trains the policy that maximizes the mutual
90
+ information between the latent variable and the states, then altering
91
+ the latent variable of the policy network creates multiple policies
92
+ performing disparately. Although this interesting viewpoint has
93
+ attracted a spectrum of following works [1, 11, 13, 20, 43], the afore-
94
+ mentioned methods achieve diverse policies depending on the task
95
+ in an unsupervised way, resulting in the algorithm performing in-
96
+ sufficient generality in different tasks. In fact, the ideal algorithm
97
+ to implement the diverse policies should be general across a range
98
+ of tasks, which motivates us to design a task-oriented algorithm
99
+ working towards developing various policies. Eysenbach et al. pro-
100
+ posed [12] that the learned skills could not construct all the state
101
+ marginal distributions in the downstream tasks. For task relevance,
102
+ the mutual information is added as an intrinsic reward for em-
103
+ powerment [7, 36], but these methods change the original reward
104
+ arXiv:2301.02375v1 [cs.LG] 6 Jan 2023
105
+
106
+ function resulting in the performance being extremely sensitive to
107
+ the trade-off between original and intrinsic rewards.
108
+ Our method insights from an interesting phenomenon during
109
+ the exploration. The critic aims to approximate the accumulated
110
+ reward by bootstrapping in the actor-critic framework. However,
111
+ the different critic functions may have great differences even if they
112
+ approximate the same target due to the function approximation er-
113
+ ror. For instance, Twin Delayed Deep Deterministic policy gradient
114
+ (TD3) [16] presents that two value functions with different initial
115
+ parameters perform quite differently with identical targets. It is
116
+ knotty to measure whether a value function is exact or not, and
117
+ the gap between these two value functions is termed as controversy,
118
+ which sees a decreasing trend along with the value function updat-
119
+ ing process. Our intuition can be ascribed that controversy in the
120
+ value estimation will lead to sub-optimal policies, and these policies
121
+ have a bias toward message acquisition known as the style.
122
+ This paper highlights that controversy can be utilized to encour-
123
+ age policies to yield multiple styles, and encourages exploration
124
+ for a continuous control task by applying multi-styled policies.
125
+ Our paper contributes three aspects. (1) We first describe that the
126
+ estimation bias in double value functions can lead to various ex-
127
+ ploration styles. (2) This paper proposes the CCEP algorithm,
128
+ encouraging diverse exploration for environments by cooperation
129
+ from multi-styled policies. (3) Finally, in CCEP, we design a novel
130
+ framework, termed as the centralized value function framework,
131
+ which is updated by experience collected from all the policies and
132
+ accomplishes the message delivery mechanism between different
133
+ policies. Extensive experiments are conducted on the MuJoCo plat-
134
+ form to evaluate the effectiveness of our method. The results reveal
135
+ that the proposed CCEP approach attains substantial improvements
136
+ in both average return and sample efficiency on the baseline across
137
+ selected environments, and the average return of agents trained
138
+ using CCEP is 6.7% higher than that of the baseline. Besides, CCEP
139
+ also allows agents to explore more states during the same train-
140
+ ing time steps as the baseline. Additional analysis indicates that
141
+ message delivery leading to the cooperative multi-styled policies
142
+ further enhanced the exploration efficiency by 8.6% compared with
143
+ that of without cooperation.
144
+ We organize the rest of this paper as follows. Section 2 briefly
145
+ explains the concepts of RL. We elaborate on how to perform the
146
+ CCEP algorithm in Section 3. We introduce our experimental set-
147
+ tings in Section 4. Section 5 presents and analyzes results to validate
148
+ the effectiveness of CCEP, after which section 6 discusses related
149
+ works. Finally, we conclude our paper in Section 7. The code and
150
+ documentation are released in the link for validating reproducibil-
151
+ ity.1
152
+ 2
153
+ PRELIMINARIES
154
+ Reinforcement learning (RL) aims at training an agent to tackle the
155
+ sequential decision problems that can be formalized as a Markov
156
+ Decision Process (MDP). This process can be defined as a tuple
157
+ (S, A, 𝑃,𝑟,𝛾), where S is the state space, A is the action space,
158
+ 𝑃 : S × A × S ↦→ [0, 1] denotes the transition probability, 𝑟 (𝑠,𝑎) is
159
+ the reward function 𝑟 : S ×A ↦→ R, determining the reward agents
160
+ will receive in the state 𝑠 while executing the action 𝑎. The 𝛾 ∈
161
+ 1https://github.com/Jincate/CCEP
162
+ (0, 1) is the discount factor. The return is defined as the discounted
163
+ accumulated reward.
164
+ 𝑅 =
165
+
166
+ ∑︁
167
+ 𝑡=0
168
+ 𝛾𝑡𝑟 (𝑠𝑡,𝑎𝑡)
169
+ (1)
170
+ In the DRL community, developers usually use the neural network
171
+ parameterized with 𝜙 to indicate the policy 𝜋(𝑎|𝑠), which inputs
172
+ an observation and outputs an action. The goal of DRL is to solve
173
+ this MDP process and find the optimal policy 𝜋𝜙∗ : S ↦→ A with
174
+ parameter 𝜙∗ that maximizes the expected accumulated return.
175
+ 𝜙∗ = arg max
176
+ 𝜙
177
+ E𝑎𝑡 ∼𝜋𝜙 (·|𝑠𝑡 ),𝑠𝑡+1∼𝑃 (·|𝑠𝑡,𝑎𝑡 )
178
+ � ∞
179
+ ∑︁
180
+ 𝑡=0
181
+ 𝛾𝑡𝑟 (𝑠𝑡,𝑎𝑡)
182
+
183
+ (2)
184
+ David Silver, et al. [44] propose that solving Eq. (2) with determin-
185
+ istic policy gradient strategy,
186
+ ∇𝜙 𝐽 (𝜙) = E𝑠𝑡+1∼𝑃 (·|𝑠𝑡,𝑎𝑡 )
187
+
188
+ ∇𝑎𝑄𝜋 (𝑠,𝑎)|𝑎=𝜋 (𝑠)∇𝜙𝜋𝜙 (𝑠)
189
+
190
+ (3)
191
+ where 𝑄𝜋 (𝑠,𝑎) = E𝑎��� ∼𝜋𝜙 (·|𝑠𝑡 ),𝑠𝑡+1∼𝑃 (·|𝑠𝑡,𝑎𝑡 ) [𝑅|𝑠,𝑎] is known as the
192
+ value function, indicating how good it is for an agent to pick action
193
+ 𝑎 while being in state 𝑠. To use the gradient-based approach (e.g.,
194
+ Stochastic Gradient Descent [41]) to solve this equation, deep Q-
195
+ learning uses the neural network to approximate the value function.
196
+ The value function parameterized with 𝜃 is updated by minimizing
197
+ the temporal difference (TD) error [45] between the estimated value
198
+ of the subsequent state 𝑠′ and the current state 𝑠.
199
+ 𝜃∗ = arg min
200
+ 𝜃
201
+ E
202
+
203
+ 𝑟 (𝑠,𝑎) + 𝛾𝑄𝜋
204
+ 𝜃 (𝑠′,𝑎′) − 𝑄𝜋
205
+ 𝜃 (𝑠,𝑎)
206
+ �2
207
+ (4)
208
+ We store the trajectories of the agent exploring the environment
209
+ in a replay buffer [32] from which sample a random mini-batch of
210
+ samples, updating the parameters mentioned above.
211
+ 3
212
+ CENTRALIZED COOPERATIVE
213
+ EXPLORATION POLICY
214
+ This section details technologies of CCEP (Centralized Cooper-
215
+ ative Exploration Policy). We first analyze value estimation bias
216
+ from function approximation errors and generate multi-styled value
217
+ functions by encouraging overestimation bias and underestimation
218
+ bias for the value functions, respectively. To achieve multi-styled
219
+ exploration, we propose a multi-objective update method for train-
220
+ ing policy and randomly select one policy to explore at each time
221
+ step. These historical trajectories during exploration are stored for
222
+ training a single policy function to achieve cooperative message
223
+ delivery. We denote our policy as 𝜋(𝑠,𝑧), where 𝑧 is a one-hot label
224
+ and represents different policies. In this work, we focus on gener-
225
+ ating multiple policies with different styles to encourage diverse
226
+ exploration. We implement our method based on TD3 [16] which
227
+ maintains double critics and uses the minimum of the critics as the
228
+ target estimate.
229
+ 3.1
230
+ Function Approximation Error
231
+ This section shows that there exist approximation errors in the value
232
+ function optimization and can accumulate to substantial scales. The
233
+ accumulated approximation error will lead to value estimation bias,
234
+ which plays a significant role in policy improvement.
235
+
236
+ Sample one style per step
237
+ 𝒛~𝒑(𝒛)
238
+ 𝒂𝒕~𝝅𝝓(𝒂𝒕|𝒔𝒕, 𝒛𝒕)
239
+ STYLE
240
+ 𝒔𝒕�𝟏~𝒑 (𝒔𝒕�𝟏|𝒔𝒕, 𝒂𝒕)
241
+ ENVIRONMENT
242
+ 𝒂𝒕
243
+ REPLAY BUFFER
244
+ 𝒔𝒕�𝟏
245
+ 𝒛
246
+ 𝒛
247
+ 𝒔𝒕�𝟏
248
+ Sample a batch of N transitions
249
+ (𝒔𝒕, 𝒂𝒕, 𝒛𝒕, 𝒔𝒕�𝟏, 𝒛𝒕�𝟏, 𝒓)
250
+ UPDATE CRITICS
251
+ 𝜽𝒊
252
+ 𝒕�𝟏 ←𝐚𝐫𝐠 𝐦𝐢𝐧
253
+ 𝜽𝒊 𝑵�𝟏∑�𝒚 − 𝑸𝜽𝒊(𝒔, 𝒂)�
254
+ 𝟐
255
+ UPDATE POLICY
256
+ Generate critics by Eq. (9)
257
+ 𝝓𝒕�𝟏 ←𝐚𝐫𝐠 𝐦𝐚𝐱
258
+ 𝝓
259
+ 𝟏
260
+ 𝟒 �
261
+ 𝒌=𝟏
262
+ 𝟒
263
+ 𝑸𝒌�𝒔, 𝝅(𝒔, 𝒛𝒌)�
264
+ 𝝓𝒕�𝟏
265
+ POLICY
266
+ No.1
267
+ POLICY
268
+ No.2
269
+ POLICY
270
+ No.3
271
+ POLICY
272
+ No.4
273
+ Cooperative exploration
274
+ Centralization
275
+ Figure 1: The workflow of CCEP Algorithm. The agent 𝜋 interacts with the environment with diverse style cooperatively and
276
+ produce the transition 𝑠𝑡 → 𝑠𝑡+1. The actor and critic are updated over a mini-batch of the transition samples. A centralized
277
+ policy with four different styles is learned from the multi-styled critics.
278
+ In value-based deep reinforcement, deep neural networks ap-
279
+ proximate the value functions, and the function approximation
280
+ error exists correspondingly. One major source of the function
281
+ approximation error comes from the optimization procedure. In
282
+ this procedure, stochastic gradient descent, which uses a batch of
283
+ random samples for gradient update each time, is the mainstream
284
+ method due to the consideration of computational resources and
285
+ training efficiency. However, as [40] has indicated, a mini-batch
286
+ gradient update may have unpredictable effects on samples outside
287
+ the training batch, which leads to the function approximation error.
288
+ For explanation, we use 𝑒𝑡 to represent the approximation error of
289
+ the value function with the state-action pair(𝑠𝑡,𝑎𝑡) as input and
290
+ approximation error 𝑒𝑡 can be modeled as follows:
291
+ 𝑄𝜃 (𝑠𝑡,𝑎𝑡) = 𝑟 (𝑠𝑡,𝑎𝑡) + 𝛾E[𝑄𝜃 (𝑠𝑡+1,𝑎𝑡+1)] − 𝑒𝑡
292
+ (5)
293
+ Approximation errors influence the value estimation when using
294
+ the value function as an estimator. The estimation may be skewed
295
+ towards an overestimation, causing a wrong estimate for a given
296
+ state. This leads to a problem of an optimal action being chosen but
297
+ replaced by a sub-optimal action, owing to the overestimation of
298
+ a sub-optimal action. Thus, the overestimation bias is a common
299
+ problem in Q-Learning with discrete actions, as we choose the
300
+ seemly best action 𝑎𝑡+1 in the target value. Still, there is little chance
301
+ for the optimal state-action pair to be updated.
302
+ E[ max
303
+ 𝑎𝑡+1∈A 𝑄(𝑠𝑡+1,𝑎𝑡+1)] ≥
304
+ max
305
+ 𝑎𝑡+1∈A E[𝑄(𝑠𝑡+1,𝑎𝑡+1)]
306
+ (6)
307
+ Mentioned overestimated bias can also occur in continuous con-
308
+ trol tasks[16], since the policy approximator always provides the
309
+ optimal action at the current state based on the value function.
310
+ While this bias can be quite small in an individual update, the bias
311
+ can be accumulated to a substantial overestimation. Eq.(5) can be
312
+ expanded as follows:
313
+ 𝑄𝜃 (𝑠𝑡,𝑎𝑡) = E𝑎𝑡 ∼𝜋𝜙 (·|𝑠𝑡 ),𝑠𝑡+1∼𝑃
314
+ � ∞
315
+ ∑︁
316
+ 𝑡=0
317
+ 𝛾𝑡 (𝑟 (𝑠𝑡,𝑎𝑡) − 𝑒𝑡)
318
+
319
+ (7)
320
+ Previous works such as Double Q-learning [23] and Double DQN [24]
321
+ are proposed to alleviate value functions of underestimating. The
322
+ idea is to maintain two independent estimators in which one is
323
+ used for estimation while the other is for selecting maximal action.
324
+ Similarly, as an extension in dealing with continuous control tasks,
325
+ TD3 [16] reduce the overestimation bias by using double value func-
326
+ tions and taking the minimum between the two value functions
327
+ for an estimation which suffers from underestimation problems as
328
+ well [50, 51].
329
+ Does the estimation error influence the performance? Given
330
+ a continuous control task, we use 𝑓 to approximate the true under-
331
+ lying value function 𝑄∗, which indicates the accumulated reward
332
+ obtained by acting 𝑎 before taking optimal policy 𝜋∗ at state 𝑠. 𝑉
333
+ represents the true underlying value function(which is not known
334
+ during training). 𝑉 ∗ and 𝑉 𝜋𝑓 represent the accumulated return
335
+ obtained by adopting the optimal policy 𝜋∗ and 𝜋𝑓 in state 𝑠 re-
336
+ spectively in which 𝜋𝑓 is a learned policy by maximizing the value
337
+ function approximate 𝑓 .
338
+ Lemma 1. (Performance Gap). The performance gap of the policy
339
+ between the optimal policy 𝜋∗ and the learned policy 𝜋𝑓 is defined
340
+ by an infinity norm ∥𝑉 ∗ − 𝑉 𝜋𝑓 ∥∞ and we have
341
+ ∥𝑉 ∗ − 𝑉 𝜋𝑓 ∥∞ ≤ 2∥𝑓 − 𝑄∗∥∞
342
+ 1 − 𝛾
343
+ We provide proof detail of Lemma 1 in Supplementary A. This
344
+ inequality indicates that the performance gap of the policy can be
345
+ bounded by the estimation error of the value function and accurate
346
+ value estimate can reduce the upper bound of the performance gap
347
+
348
+ and enhance the performance.
349
+ Do overestimation bias and underestimation bias affect per-
350
+ formance in the same way? An empirical study shows that esti-
351
+ mation bias may not always be a detrimental problem while both
352
+ underestimation bias and overestimation bias may improve learn-
353
+ ing performance which depends on the environment [30]. As an
354
+ example, for an unknown area with high stochasticity, overesti-
355
+ mation bias may help to explore the overestimated area but un-
356
+ derestimation bias prevents this. However, if these areas of high
357
+ stochasticity are given low values, the overestimation bias may lead
358
+ to excess exploration in low-value regions. The fact is that we can
359
+ not choose the environment and these different circumstances can
360
+ always occur during exploration. Our method is designed to utilize
361
+ the difference in exploration behavior brought by estimation bias
362
+ to encourage multi-styled exploration.
363
+ 3.2
364
+ Multi-Style Critics: Radical, Conservative
365
+ As mentioned above, function approximation error exists in value
366
+ functions and can accumulate to substantial scales which have a
367
+ great influence on the value estimation resulting in overestimation
368
+ or underestimation bias. Estimation bias has been researched in
369
+ recent works [16, 25, 50, 51]. While these works focus on an accurate
370
+ value estimation and discussed the method to control the estimation
371
+ bias with the use of multiple value functions for auxiliary, they
372
+ just choose one of the value functions, which seems to be the
373
+ most accurate, for policy update neglecting other value functions.
374
+ However, there is no accurate value function without trial and
375
+ error. In this section, we show how to utilize the estimation bias
376
+ and introduce our method for the generalization of multi-styled
377
+ critics.
378
+ Our intuition is that there are different degrees of estimation
379
+ bias in double randomly initialized value functions when perform-
380
+ ing function approximation. However, the estimation bias can be
381
+ controlled by applying a maximum operator and minimum op-
382
+ erator, namely the maximum of the two estimates is relatively
383
+ overestimated and the minimum of the two estimates is relatively
384
+ underestimated. Two different estimates raise a controversy about
385
+ which critic gives the accurate estimate. The best way to resolve
386
+ the controversy is to follow one of the critics to explore and collect
387
+ reward messages. While controversy does not always exist because
388
+ there is only one accurate value, the critics reach an agreement
389
+ when the state value has been exactly estimated. And the existence
390
+ of controversy means more exploration is needed.
391
+ We start by maintaining double randomly initialized value func-
392
+ tions 𝑄𝜃1 and 𝑄𝜃2 with parameters 𝜃1 and 𝜃2 respectively and up-
393
+ date the value function with TD3 [16] which takes the minimum
394
+ between the two value functions as the target value estimate:
395
+ 𝑦 = 𝑟 + min
396
+ 𝑖=1,2𝑄𝜃𝑖 (𝑠′,𝑎′),𝑎′ ∼ 𝜋𝜙
397
+ (8)
398
+ But the two randomly initialized value functions potentially have
399
+ different value estimations for a given state-action pair due to the
400
+ accumulated function approximation error. This difference leads to
401
+ the result that the two critics may give two different suggestions
402
+ for the best action choice. While these estimates are relatively
403
+ overestimated or underestimated, these different criteria for a given
404
+ state-action pair may lead to a different style of action choice. It
405
+ Algorithm 1 Centralized Cooperative Exploration Policy (CCEP)
406
+ Initialize critic networks 𝑄𝜃1,𝑄𝜃2
407
+ Initialize actor network 𝜋𝜙 with random parameters 𝜃1,𝜃2,𝜙
408
+ Initialize target networks 𝜃 ′
409
+ 1 ← 𝜃1,𝜃 ′
410
+ 2 ← 𝜃2,𝜙′ ← 𝜙
411
+ Initialize replay buffer B
412
+ Initialize number of skills K
413
+ 1: for 𝑡 = 1 to 𝑇 do
414
+ 2:
415
+ Sample a skill 𝑧 from 𝑝(𝑧)
416
+ 3:
417
+ Select action with noise 𝑎 ∼ 𝜋𝜙 (𝑠,𝑎) + 𝜖,𝜖 ∼ N (0, 𝜎)
418
+ 4:
419
+ Observe a reward 𝑟 and a new state 𝑠′
420
+ 5:
421
+ Store transition tuple (𝑠,𝑧,𝑎,𝑟,𝑠′,𝑧′) in B
422
+ 6:
423
+ Sample mini-batch of 𝑁 transitions (𝑠,𝑧,𝑎,𝑟,𝑠′,𝑧′) from 𝐵
424
+ 7:
425
+ 𝑎′ ← 𝜋𝜙′(𝑠′,𝑧′) + 𝜖,𝜖 ∼ 𝑐𝑙𝑖𝑝(N (0, 𝜎), −𝑐,𝑐)
426
+ 8:
427
+ 𝑦 = 𝑟 + 𝛾 min𝑖=1,2 𝑄𝜃′
428
+ 𝑖 (𝑠′,𝑎′)
429
+ 9:
430
+ Update critics: 𝜃𝑖 ← arg min𝜃𝑖 𝑁 −1 �(𝑦 − 𝑄𝜃𝑖 (𝑠,𝑎))2
431
+ 10:
432
+ if t mod d then
433
+ 11:
434
+ Update policy:
435
+ 12:
436
+ ∇𝜙 𝐽 (𝜙) = 𝑁 −1K−1 � ∇𝑎𝑄 𝑗 (𝑠,𝑎)|𝑎=𝜋𝜙 (𝑠,𝑧)∇𝜙𝜋𝜙 (𝑠,𝑧)
437
+ 13:
438
+ Update target networks
439
+ 14:
440
+ 𝜃 ′
441
+ 𝑖 ← 𝜏𝜃𝑖 + (1 − 𝜏)𝜃 ′
442
+ 𝑖
443
+ 15:
444
+ 𝜙′
445
+ 𝑖 ← 𝜏𝜙𝑖 + (1 − 𝜏)𝜙′
446
+ 𝑖
447
+ is reasonable the estimation is radical if we choose the maximum
448
+ value of the two to estimate and the estimation is conservative if
449
+ we choose the minimum value of the two. Thus, we consider four
450
+ critics:
451
+ 𝑄 𝑗 (𝑠,𝑎) =
452
+ 
453
+ 
454
+ 𝑄𝜃1 (𝑠,𝑎)
455
+ 𝑗 = 0
456
+ 𝑄𝜃2 (𝑠,𝑎)
457
+ 𝑗 = 1
458
+ max(𝑄𝜃1 (𝑠,𝑎),𝑄𝜃2 (𝑠,𝑎))
459
+ 𝑗 = 2
460
+ min(𝑄𝜃1 (𝑠,𝑎),𝑄𝜃2 (𝑠,𝑎))
461
+ 𝑗 = 3
462
+ (9)
463
+ There exists controversy among these critics, and the controversy
464
+ can further influence the performance of the policy learned.
465
+ 3.3
466
+ Opposite Value Functions
467
+ This approach for generating diverse styles raises the problem that
468
+ the value functions may not provide sufficient difference in style
469
+ when the controversy disappear. This phenomenon is very com-
470
+ mon when the value function converges. But we don’t want this to
471
+ happen too soon, because we want the value functions to provide
472
+ more exploration for the policy. The controversy exists due to the
473
+ randomly initialized parameters of the neural networks and the
474
+ error accumulation. But actually, there is a small probability that
475
+ the two networks have great similarities, which will lead to double
476
+ consistent critics. This is not what we want, because consistent
477
+ critics mean monotonous policy. To avoid this, we try to enlarge the
478
+ controversy. The solution in this paper is to learn two opposite tar-
479
+ gets respectively for the two networks, where one of the networks
480
+ approximates the positive value and another approximates the neg-
481
+ ative one. This approach is equivalent to adding a factor -1 to the
482
+ final layer of either network. We find that the controversy is guar-
483
+ anteed with this simple network structure change. To provide some
484
+ intuition, we compared the controversy changes after the double
485
+ value functions learn the opposite target. We express the amount
486
+
487
+ 0.0
488
+ 0.2
489
+ 0.4
490
+ 0.6
491
+ 0.8
492
+ 1.0
493
+ Time Steps (1e6)
494
+ 0.0
495
+ 0.5
496
+ 1.0
497
+ 1.5
498
+ 2.0
499
+ 2.5
500
+ Average Error
501
+ (a) HalfCheetah-v3
502
+ 0.0
503
+ 0.2
504
+ 0.4
505
+ 0.6
506
+ 0.8
507
+ 1.0
508
+ Time Steps (1e6)
509
+ 0
510
+ 1
511
+ 2
512
+ 3
513
+ Average Error
514
+ Target
515
+ same target
516
+ opposite target
517
+ (b) Walker2d-v3
518
+ Figure 2: Measuring the error between double critics given
519
+ same/opposite targets in TD3 on MuJoCo environments over
520
+ 1 million time steps
521
+ of controversy between the two value functions by the errors of
522
+ state values in a batch of samples. Figure 2 shows the controversy
523
+ measuring over MuJoCo [48] environments in HalfCheetah-v3 and
524
+ Walker2d-v3. The results show that with the simple network struc-
525
+ ture change, the controversy is enlarged. But this approach will
526
+ not influence the value estimation because we just fine-tune the
527
+ structure.
528
+ 3.4
529
+ Centralized Cooperation
530
+ With four critics, we train a centralized cooperative policy to encour-
531
+ age multi-styled explorations through diverse value estimations.
532
+ We model this problem as a multi-objective optimization problem.
533
+ The target is to train multiple policies, with each policy targeting an
534
+ individual value function. We express the policy function as 𝜋(𝑠,𝑧),
535
+ with state 𝑠 and latent variable 𝑧 as input. The latent variable 𝑧,
536
+ which is a one-hot label in our method, represents different policies.
537
+ The architecture of our centralized cooperative policy is shown
538
+ in Figure 3. This idea comes from skill discovery method [11, 43],
539
+ which use the latent variable 𝑧 to express different skills. And in
540
+ skill discovery, the target is to maximize the mutual information be-
541
+ tween latent variable 𝑧 and some aspects of the trajectories, which
542
+ is a different target for a different latent variable 𝑧. Our method
543
+ encourages diverse styles of policies by different targets as well.
544
+ Particularly, we sample latent variable 𝑧 from set {0, 1, 2, 3} and
545
+ encode it in a one-hot label. For a given latent variable 𝑧, the policy
546
+ targets 𝑧-th value functions in Eq.(9). With different latent variable
547
+ 𝑧, the policy shows diverse styles due to the multi-styled targets.
548
+ We make an experiment showing that there exists different explo-
549
+ ration preferences for these policies (Section 5.2) In the exploration
550
+ procedure, we randomly sample latent variable 𝑧 and make deci-
551
+ sions by policy 𝜋(𝑠,𝑧). This approach enables diverse styles to be
552
+ applied at each time step. Broadly speaking, our exploration policy
553
+ has the following characteristics: Multi-styled, Centralized, and
554
+ Cooperative.
555
+ Multi-styled. We train four policies to accomplish the explo-
556
+ ration. These policies learn from the corresponding value function
557
+ 𝑄 𝑗:
558
+ 𝜋∗
559
+ 𝑗 = arg max
560
+ 𝜋
561
+ 𝑄 𝑗
562
+ (10)
563
+ There are four value estimators, in which two of them (𝑗 = 0, 1)
564
+ are normal but different estimators, one (𝑗 = 2) is an overestimated
565
+ Policy 1
566
+ Policy 2
567
+ Policy 3
568
+ Policy 4
569
+ Centralized
570
+ Cooperative
571
+ Policy
572
+ Environment
573
+ Label variance
574
+ Input
575
+ Output
576
+ Figure 3: The architecture of our centralized cooperative pol-
577
+ icy. The agent cooperatively explores the environments by
578
+ selecting one of the styles at each time step. The style se-
579
+ lection process is implemented by sampling latent variable
580
+ 𝑧. Policies with diverse styles exchange messages through a
581
+ centralized network.
582
+ estimator compared to the other (𝑗 = 3) and helps encourage ex-
583
+ plorations in overestimated actions, the last remaining one (𝑗 = 3)
584
+ is a conservative estimator and brings more exploitation as illus-
585
+ trated in the previous section. It is appropriate for the policies to
586
+ perform in a variety of ways given the varied estimators they use
587
+ (e.g., conservative, radical).
588
+ Centralized. Our policy is a centralized policy because we make
589
+ use of all the policies learned in each episode. At each time step𝑡, we
590
+ sample one of these policies for exploration. It allows us to generate
591
+ a variety of trajectories adopting this exploration approach as this
592
+ centralized policy can generate 4𝑛 types of trajectories theoretically
593
+ for 𝑛-step exploration, compared to using a single policy that can
594
+ only generate one. These trajectories are stored as experience and
595
+ maintain the update for a pair of centralized value functions.
596
+ Cooperative. We update the policy cooperatively. With multi-
597
+ ple policies learning their respective value functions, “knowledge”
598
+ learned by each policy cannot be shared. Our method is to learn a
599
+ single network for policies and learn cooperatively [4]. To repre-
600
+ sent different policies, we feed latent variables 𝑧 which are one-hot
601
+ labels as extra input to the network. The policy which inputs latent
602
+ variable 𝑧 and state𝑠 and outputs action𝑎 can be defined as 𝜋(𝑠,𝑧).
603
+ We sample 𝑧 to represent the sampling of different policies. Thus,
604
+ the policy can be updated by taking deterministic policy gradient.
605
+ ∇𝜙 𝐽 (𝜙) = 𝑁 −1K−1 ∑︁
606
+ ∇𝑎𝑄 𝑗 (𝑠,𝑎)|𝑎=𝜋𝜙 (𝑠,𝑧)∇𝜙𝜋𝜙 (𝑠,𝑧)
607
+ (11)
608
+ Where 𝑄 𝑗 (𝑠,𝑎) refer to the multi-styled critics in Eq.(9), K is the
609
+ number of styles which is 4 in this algorithm. The specific algorithm
610
+ is shown in Algorithm 1.
611
+ 4
612
+ EXPERIMENTAL SETTINGS
613
+ To evaluate our method, we test our algorithm on the suit of
614
+ MujoCo [48] continuous control tasks, including HalfCheetah-v3,
615
+
616
+ G(a)
617
+ (b)
618
+ (c)
619
+ (d)
620
+ Figure 4: Screenshots of MuJoCo environments. (a) Ant-v3,
621
+ (b) HalfCheetah-v3, (c) Walker2d-v3, (d) Hopper-v3
622
+ Hopper-v3, Walker2d-v3, Ant-v3, Pusher-v2 and Humanoid-v3 (the
623
+ screenshots are presented in Figure 4).
624
+ For implementation, our method builds on TD3 [16], and for com-
625
+ parison, we also establish three-layer feedforward neural networks
626
+ with 256 hidden nodes per hidden layer for both critics and actors.
627
+ Particularly, the actor takes state 𝑠 and latent variable 𝑧 concate-
628
+ nated as input, where the latent variable 𝑧 is encoded as one-hot
629
+ label. At each time step, both networks are trained with a mini-
630
+ batch of 256 samples. We apply soft updates for target networks as
631
+ well.
632
+ We compared our algorithm against some classic algorithms
633
+ such as DDPG [30], which is an efficient off-policy reinforcement
634
+ learning method for continuous tasks; PPO [42], the state-of-the-
635
+ art policy gradient algorithms; TD3 [16], which is an extension
636
+ to DDPG; SAC [22], which is an entropy-based method with high
637
+ sample efficiency. Further, we compared our algorithm with the
638
+ latest algorithm in solving the exploration problems in continuous
639
+ control tasks such as OAC [8], which makes improvements on SAC
640
+ for better exploration. We implement DDPG and PPO by OpenAI’s
641
+ baselines repository and SAC, TD3, and OAC by the github the
642
+ author provided. And we use the parameter the author recommend
643
+ for implementation. The details of the implementation are shown
644
+ in Supplementary B.
645
+ 5
646
+ EXPERIMENTAL RESULTS AND ANALYSIS
647
+ 5.1
648
+ Evaluation
649
+ To validate the performance of the CCEP algorithm, we evaluate
650
+ our algorithm in MuJoCo continuous control suites. We perform
651
+ interactions for 1 million steps in 10 different seeds and evaluate
652
+ the algorithm over 10 episodes every 5k steps. Our results report
653
+ the mean scores and standard deviations in the 10 seeds. We show
654
+ learning curves in Figure 5 and the max average return over 10 trials
655
+ of 1 million time steps in Table 1. The learning curves in 1 million
656
+ time steps show that our algorithm achieves a higher sample effi-
657
+ ciency compared with the latest algorithm. Furthermore, the results
658
+ in the Table 1 indicates that our algorithm shows superior perfor-
659
+ mance. And in HalfCheetah-v3, Walker2d-v3, Hopper-v3, Ant-v3,
660
+ Table 1: The highest average return over 10 trials of 1 million
661
+ time steps. The maximum value for each task is bolded.
662
+ Environment
663
+ Ours
664
+ OAC
665
+ SAC
666
+ TD3
667
+ DDPG
668
+ PPO
669
+ HalfCheetah
670
+ 11945
671
+ 9921
672
+ 11129
673
+ 9758
674
+ 8469
675
+ 3681
676
+ Hopper
677
+ 3636
678
+ 3364
679
+ 3357
680
+ 3479
681
+ 2709
682
+ 3365
683
+ Walker2d
684
+ 4706
685
+ 4458
686
+ 4349
687
+ 4229
688
+ 3669
689
+ 3668
690
+ Ant
691
+ 5630
692
+ 4519
693
+ 5084
694
+ 5142
695
+ 1808
696
+ 909
697
+ Pusher
698
+ -21
699
+ -25
700
+ -20
701
+ -25
702
+ -29
703
+ -21
704
+ Humanoid
705
+ 5325
706
+ 5747
707
+ 5523
708
+ 5356
709
+ 1728
710
+ 586
711
+ our algorithm outperforms all the other baselines and achieve sig-
712
+ nificant improvements. While in the Pusher-v2 task, our algorithm
713
+ show higher stability than that of TD3. For further evaluation, we
714
+ evaluate our algorithm in the state-based suite PyBullet [9] which
715
+ is considered to be harder than the suite MuJoCo. Our algorithm
716
+ still shows better performance compared to the baseline algorithms.
717
+ The corresponding results are shown in Supplementary C.1.
718
+ 5.2
719
+ Policy Style
720
+ To ensure that our proposed CCEP algorithm learns diverse styles,
721
+ we compared the distribution of explored trajectories when ex-
722
+ ploring with a single style only. We test the algorithm in Ant-v3
723
+ environment over 1𝑒6 time steps and use the states sampled to rep-
724
+ resent the trajectories. Figure 6 shows the states explored by each
725
+ style at 1𝑒5, 2𝑒5 and 3𝑒5 learning steps, and a more detailed results
726
+ are shown in Supplementary C.3. We collect the states sampled over
727
+ 10 episodes with different seeds and apply t-SNE [49] for better
728
+ visualization. The results show that while part of the states can be
729
+ gathered by all styles which implies a compromise in controversy,
730
+ there is a considerably large region of states that can only be ex-
731
+ plored by a unique style of policy. Though different styles, diverse
732
+ styles come to be in compromise as training process goes on. This
733
+ phenomenon suggests that CCEP behaves in multi-styled explo-
734
+ ration which leads to an exploration preference, and styles come
735
+ to an agreement with sufficient exploration. Another phenomenal
736
+ conclusion is that although the style tends to be consistent, new
737
+ styles are emerging which brings enduring exploration capabilities.
738
+ 5.3
739
+ Measuring Exploration Ability
740
+ The critical problem of our proposed method is whether we achieve
741
+ higher sample efficiency. Although the learning curves (Figure 5)
742
+ gives considerably convincing results, a more intuitive result has
743
+ been given in Figure 7. We compared the exploration of CCEP
744
+ with that of TD3 and SAC (which achieve the trade-off between
745
+ exploration and exploitation by entropy regularization.) over 10
746
+ episodes with different seeds (Figure 7). For a fair comparison, these
747
+ methods are trained in Ant-v3 with the same seed at half of the
748
+ training process. In order to get reliable results, the states explored
749
+ are gathered in 10 episodes with different seeds. We still apply the
750
+ same t-SNE [49] transformation to the states explored by all of the
751
+ algorithms for better visualization. While there are great differences
752
+ between the states explored by TD3 (green) and SAC (blue), the
753
+ result shows that our algorithm (red) explores a wider range of
754
+ states which even covers that TD3 and SAC explored.
755
+
756
+ 0.00
757
+ 0.25
758
+ 0.50
759
+ 0.75
760
+ 1.00
761
+ 0
762
+ 5000
763
+ 10000
764
+ Average Return
765
+ HalfCheetah-v3
766
+ 0.00
767
+ 0.25
768
+ 0.50
769
+ 0.75
770
+ 1.00
771
+ 0
772
+ 2000
773
+ 4000
774
+ Walker2d-v3
775
+ 0.00
776
+ 0.25
777
+ 0.50
778
+ 0.75
779
+ 1.00
780
+ 0
781
+ 1000
782
+ 2000
783
+ 3000
784
+ 4000
785
+ Hopper-v3
786
+ 0.00
787
+ 0.25
788
+ 0.50
789
+ 0.75
790
+ 1.00
791
+ Time Steps (1e6)
792
+ 0
793
+ 2000
794
+ 4000
795
+ Average Return
796
+ Humanoid-v3
797
+ 0.00
798
+ 0.25
799
+ 0.50
800
+ 0.75
801
+ 1.00
802
+ Time Steps (1e6)
803
+ 2000
804
+ 0
805
+ 2000
806
+ 4000
807
+ 6000
808
+ Ant-v3
809
+ 0.00
810
+ 0.25
811
+ 0.50
812
+ 0.75
813
+ 1.00
814
+ Time Steps (1e6)
815
+ 80
816
+ 60
817
+ 40
818
+ 20
819
+ Pusher-v2
820
+ Ours
821
+ OAC
822
+ PPO
823
+ DDPG
824
+ TD3
825
+ SAC
826
+ Figure 5: Learning curves for 6 MuJoCo continuous control tasks.For better visualization,the curves are smoothed uniformly.
827
+ The bolded line represents the average evaluation over 10 seeds. The shaded region represents a standard deviation of the
828
+ average evaluation over 10 seeds.
829
+ (1) Learning Steps = 100000
830
+ (2) Learning Steps = 200000
831
+ (3) Learning Steps = 300000
832
+ POLICY No.1
833
+ POLICY No.2
834
+ POLICY No.3
835
+ POLICY No.4
836
+ Figure 6: The states visited by each style. For better visualization, the states get dimension reduction by t-SNE. The points with
837
+ different color represents the states visited by the policy with the style. The distance between points represents the difference
838
+ between states.
839
+ 5.4
840
+ Ablation Study
841
+ We perform an ablation study to understand the contribution of the
842
+ cooperation between policies for message delivery. The results are
843
+ shown in Table 2 where we compare the performance of training
844
+ policies by removing policy cooperation and training them sepa-
845
+ rately. We perform interactions for 1 million time steps for each
846
+ method. The results show that without cooperation, the policy net-
847
+ work not only trains 4 times more network parameters but also
848
+ fails to reduce performance. And this performance degradation is
849
+ even more pronounced on Walker2d. Additional learning curves
850
+ can be found in Supplementary C.2.
851
+ 6
852
+ RELATED WORK
853
+ This section discusses several methods proposed recently for im-
854
+ proving the exploration of deep reinforcement learning.
855
+ A range of works take an effort in encouraging explorations with
856
+
857
+ (1) Ours
858
+ (2) TD3
859
+ (3) SAC
860
+ Figure 7: Measuring the exploration region. Comparison of exploration capabilities of TD3 (green), SAC (blue) and Ours (red).
861
+ The points represent region explored by each method in 10 episodes. All the states get dimension reduction by the same t-SNE
862
+ transformation for better visualization.
863
+ Table 2: Max Average Return over 5 trials of 1 million time
864
+ steps, comparing ablation over cooperation for message de-
865
+ livery. The maximum value for each task is bolded.
866
+ Method
867
+ HCheetah
868
+ Hopper
869
+ Walker2d
870
+ Ant
871
+ CCEP
872
+ 11969
873
+ 3672
874
+ 4789
875
+ 5488
876
+ CCEP-Cooperation
877
+ 11384
878
+ 3583
879
+ 4087
880
+ 4907
881
+ TD3
882
+ 9792
883
+ 3531
884
+ 4190
885
+ 4810
886
+ the use of randomness over model parameters [6]. Another preva-
887
+ lent series of works propose to enhance exploration by simulta-
888
+ neously maximizing the expected return and entropy of the pol-
889
+ icy [15, 21, 22, 39, 55]. Whereas, these methods do not provide
890
+ heuristic knowledge to guide the exploration, which can be consid-
891
+ ered to be insufficient and time-consuming.
892
+ To achieve effective exploration, the curiosity mechanism [19, 38]
893
+ has been proposed in recent works, e.g., the counted-based ap-
894
+ proaches [33] which quantify the “novelty” of a state by the times
895
+ visited. However, these methods maintain the state-action visitation
896
+ counts which make it challenging in solving high-dimensional or
897
+ continuous tasks. Other works rely on errors in predicting dynam-
898
+ ics, which have been used to address the difficulties in complex
899
+ environments [5, 37, 38]. Though the Intrinsic Curiosity Module
900
+ (ICM) [37] maintains a predictor on state transitions and considers
901
+ the prediction error as an intrinsic reward, Random Network Distil-
902
+ lation (RND) [5] utilizes the prediction errors of networks trained
903
+ on historical trajectories to quantify the novelty of states, which is
904
+ effective and easy to implement in real applications.
905
+ Another direction in previous work is to study exploration in
906
+ hierarchical reinforcement learning (HRL) [3, 47]. These methods
907
+ are insight from the fact that developers prefer to divide the com-
908
+ prehensive and knotty problems into several solvable sub-problems.
909
+ There are some further studies on hierarchy in terms of tasks, rep-
910
+ resentative of which are goal-based reinforcement learning and
911
+ skill discovery. The similarity of these approaches is that they both
912
+ identify different policies by utilizing latent variables. In goal-based
913
+ RL, the latent variables are defined by the policy’s goal, which
914
+ aims to complete several sub-goals and accomplish the whole task.
915
+ These methods introduce prior human knowledge, causing them
916
+ to work brilliantly on some tasks but fail when unaware of human
917
+ knowledge. Despite our method also introducing latent variables to
918
+ represent different styles of policies, all the policies share the same
919
+ objective, nevertheless differing in the road to reach the destination.
920
+ Skill discovery methods, which adopt the latent variable to repre-
921
+ sent the skill learned from the policy, introduce mutual information
922
+ to organize relationships between the latent variable 𝑧 and some
923
+ aspects of the trajectories to acquire diverse skills (also known
924
+ as style) [1, 11, 13, 20, 43]. Nevertheless, these methods train the
925
+ policy in an unsupervised way [11, 13, 43], suggesting that the
926
+ skills trained are unaware of task-driven, and they cannot rep-
927
+ resent the optimal policies when adapted to downstream tasks
928
+ illustrated in [12]. Our method avoids this issue because we train
929
+ the policy task-oriented and demonstrate the benefit brought by
930
+ the attention of these policies to the state value making them differ
931
+ considerably in exploration style. For task relevance, some related
932
+ works that learn skills by jointly learning a set of skills and a meta-
933
+ controller [3, 10, 13, 14, 28, 29]. The options of the meta-controller
934
+ control different attentions of each policy. However, these meth-
935
+ ods usually choose the best option to explore and rarely execute
936
+ sub-optimal options, leading to the drawback – the algorithm tends
937
+ to ignore sub-optimal actions that maybe fail in most states but
938
+ are effective in a few critical scenarios. Our proposed approach
939
+ randomly selects different styles of policies for directed coopera-
940
+ tive exploration, which are improved accordingly with the value
941
+ function and produce different styles due to differences in attention.
942
+ 7
943
+ CONCLUSION
944
+ In the value-based method, value estimation bias has been a com-
945
+ mon problem. While different estimation bias in double value func-
946
+ tions lead to value function controversy, the controversy can be
947
+ utilized to encourage policies to yield multiple styles. In this paper,
948
+ we aim at encouraging explorations by multi-styled policies. We
949
+
950
+ start by analysis on estimation bias during the value function train-
951
+ ing process and its effect on the exploration. We then encourage
952
+ this controversy between the value functions and generate four
953
+ critics for producing multi-styled policies. Finally, we apply these
954
+ policies with diverse styles for centralized cooperative exploration
955
+ which perform superior sample efficiency in the test environment.
956
+ Though there are a lot of works focusing on reducing the estimation
957
+ bias for an accurate value estimation, few works try to utilize these
958
+ inevitable errors to make improvements. Our results show that it
959
+ is also an option to use the errors to encourage explorations. For
960
+ future work, it is an exciting avenue for focusing on more expres-
961
+ sive policy styles. A style that can be represented as a continuous
962
+ distribution may be more efficient and more expressive.
963
+ REFERENCES
964
+ [1] Joshua Achiam, Harrison Edwards, Dario Amodei, and Pieter Abbeel. 2018.
965
+ Variational option discovery algorithms. CoRR abs/1807.10299 (2018).
966
+ http:
967
+ //arxiv.org/abs/1807.10299
968
+ [2] Marcin Andrychowicz, Filip Wolski, Alex Ray, Jonas Schneider, Rachel Fong,
969
+ Peter Welinder, Bob McGrew, Josh Tobin, OpenAI Pieter Abbeel, and Wojciech
970
+ Zaremba. 2017. Hindsight experience replay. 30 (2017), 5048–5058.
971
+ [3] Pierre-Luc Bacon, Jean Harb, and Doina Precup. 2017. The option-critic archi-
972
+ tecture. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 31.
973
+ 1726–1734.
974
+ [4] Yunpeng Bai, Chen Gong, Bin Zhang, Guoliang Fan, Xinwen Hou, and Yu Lu. 2022.
975
+ Cooperative Multi-Agent Reinforcement Learning with Hypergraph Convolution.
976
+ In 2022 International Joint Conference on Neural Networks (IJCNN). 1–8. https:
977
+ //doi.org/10.1109/IJCNN55064.2022.9891942
978
+ [5] Yuri Burda, Harrison Edwards, Amos Storkey, et al. 2019. Exploration by random
979
+ network distillation. In 7th International Conference on Learning Representations,
980
+ ICLR 2019, New Orleans, LA, USA, May 6-9, 2019. OpenReview.net.
981
+ [6] Olivier Chapelle and Lihong Li. 2011. An Empirical Evaluation of Thompson
982
+ Sampling. In Advances in Neural Information Processing Systems, J. Shawe-Taylor,
983
+ R. Zemel, P. Bartlett, F. Pereira, and K.Q. Weinberger (Eds.), Vol. 24. Curran
984
+ Associates, Inc., 2249–2257.
985
+ https://proceedings.neurips.cc/paper/2011/file/
986
+ e53a0a2978c28872a4505bdb51db06dc-Paper.pdf
987
+ [7] Jongwook Choi, Archit Sharma, Honglak Lee, Sergey Levine, and Shixiang Shane
988
+ Gu. 2021.
989
+ Variational empowerment as representation learning for goal-
990
+ conditioned reinforcement learning. In International Conference on Machine
991
+ Learning. PMLR, 1953–1963.
992
+ [8] Kamil Ciosek, Quan Vuong, Robert Loftin, and Katja Hofmann. 2019. Better
993
+ Exploration with Optimistic Actor Critic. In Advances in Neural Information Pro-
994
+ cessing Systems 32: Annual Conference on Neural Information Processing Systems
995
+ 2019, NeurIPS 2019, December 8-14, 2019, Vancouver, BC, Canada, Hanna M. Wal-
996
+ lach, Hugo Larochelle, Alina Beygelzimer, Florence d’Alché-Buc, Emily B. Fox,
997
+ and Roman Garnett (Eds.). 1785–1796. https://proceedings.neurips.cc/paper/
998
+ 2019/hash/a34bacf839b923770b2c360eefa26748-Abstract.html
999
+ [9] Erwin Coumans and Yunfei Bai. 2016. Pybullet, a python module for physics
1000
+ simulation for games, robotics and machine learning. (2016).
1001
+ [10] Peter Dayan and Geoffrey E Hinton. 1992. Feudal reinforcement learning. Ad-
1002
+ vances in neural information processing systems 5 (1992), 271–278.
1003
+ [11] Benjamin Eysenbach, Abhishek Gupta, Julian Ibarz, and Sergey Levine. 2019.
1004
+ Diversity is all you need: Learning skills without a reward function, In 7th
1005
+ International Conference on Learning Representations, ICLR 2019. arXiv preprint
1006
+ arXiv:1802.06070.
1007
+ [12] Benjamin Eysenbach, Ruslan Salakhutdinov, and Sergey Levine. 2022. The infor-
1008
+ mation geometry of unsupervised reinforcement learning. https://openreview.
1009
+ net/forum?id=3wU2UX0voE
1010
+ [13] Carlos Florensa, Yan Duan, and Pieter Abbeel. 2017. Stochastic neural networks
1011
+ for hierarchical reinforcement learning. In 5th International Conference on Learn-
1012
+ ing Representations, ICLR 2017. OpenReview.net, Toulon, France.
1013
+ [14] Kevin Frans, Jonathan Ho, Xi Chen, Pieter Abbeel, and John Schulman. 2018.
1014
+ Meta learning shared hierarchies. In 6th International Conference on Learning
1015
+ Representations, ICLR 2018. OpenReview.net. https://openreview.net/forum?id=
1016
+ SyX0IeWAW
1017
+ [15] Justin Fu, John D. Co-Reyes, and Sergey Levine. 2017. EX2: Exploration with
1018
+ Exemplar Models for Deep Reinforcement Learning. In Proceedings of the 31st In-
1019
+ ternational Conference on Neural Information Processing Systems (Long Beach, Cal-
1020
+ ifornia, USA) (NIPS’17). Curran Associates Inc., Red Hook, NY, USA, 2574–2584.
1021
+ [16] Scott Fujimoto, Herke Hoof, and David Meger. 2018. Addressing function ap-
1022
+ proximation error in actor-critic methods. In Proceedings of the 35th International
1023
+ Conference on Machine Learning, ICML 2018 (Proceedings of Machine Learning Re-
1024
+ search, Vol. 80). PMLR, PMLR, Stockholmsmässan, Stockholm, Sweden, 1582–1591.
1025
+ http://proceedings.mlr.press/v80/fujimoto18a.html
1026
+ [17] Chen Gong, Qiang He, Yunpeng Bai, Xinwen Hou, et al. 2021. Wide-Sense
1027
+ Stationary Policy Optimization with Bellman Residual on Video Games. In 2021
1028
+ IEEE International Conference on Multimedia and Expo, ICME. IEEE, 1–6.
1029
+ [18] Chen Gong, Zhou Yang, Yunpeng Bai, Junda He, Jieke Shi, Arunesh Sinha, Bowen
1030
+ Xu, Xinwen Hou, Guoliang Fan, and David Lo. 2022. Mind Your Data! Hiding
1031
+ Backdoors in Offline Reinforcement Learning Datasets. CoRR abs/2210.04688
1032
+ (2022). https://doi.org/10.48550/arXiv.2210.04688 arXiv:2210.04688
1033
+ [19] Chen Gong, Zhou Yang, Yunpeng Bai, Jieke Shi, Arunesh Sinha, Bowen Xu,
1034
+ David Lo, Xinwen Hou, and Guoliang Fan. 2022. Curiosity-Driven and Victim-
1035
+ Aware Adversarial Policies. In Proceedings of the 38th Annual Computer Security
1036
+ Applications Conference (Austin, TX, USA) (ACSAC ’22). 186–200. https://doi.org/
1037
+ 10.1145/3564625.3564636
1038
+ [20] Karol Gregor, Danilo Jimenez Rezende, and Daan Wierstra. 2017. Variational
1039
+ intrinsic control. In 5th International Conference on Learning Representations, ICLR
1040
+ 2017. OpenReview.net. https://openreview.net/forum?id=Skc-Fo4Yg
1041
+ [21] Tuomas Haarnoja, Haoran Tang, Pieter Abbeel, and Sergey Levine. 2017. Rein-
1042
+ forcement Learning with Deep Energy-Based Policies. In Proceedings of the 34th
1043
+ International Conference on Machine Learning (Sydney, NSW, Australia) (ICML’17,
1044
+ Vol. 70). JMLR.org, 1352–1361.
1045
+ [22] Tuomas Haarnoja, Aurick Zhou, Pieter Abbeel, and Sergey Levine. 2018. Soft
1046
+ actor-critic: Off-policy maximum entropy deep reinforcement learning with a
1047
+ stochastic actor. In Proceedings of the 35th International Conference on Machine
1048
+ Learning, ICML 2018 (Proceedings of Machine Learning Research, Vol. 80). PMLR,
1049
+ PMLR, Stockholmsmässan, Stockholm, Sweden, 1856–1865.
1050
+ [23] Hado Hasselt. 2010. Double Q-learning. In Advances in Neural Information Pro-
1051
+ cessing Systems, J. Lafferty, C. Williams, J. Shawe-Taylor, R. Zemel, and A. Culotta
1052
+ (Eds.), Vol. 23. Curran Associates, Inc.
1053
+ https://proceedings.neurips.cc/paper/
1054
+ 2010/file/091d584fced301b442654dd8c23b3fc9-Paper.pdf
1055
+ [24] Hado van Hasselt, Arthur Guez, and David Silver. 2016. Deep Reinforcement
1056
+ Learning with Double Q-Learning. In Proceedings of the Thirtieth AAAI Conference
1057
+ on Artificial Intelligence (Phoenix, Arizona) (AAAI’16). AAAI Press, 2094–2100.
1058
+ [25] Qiang He and Xinwen Hou. 2020. WD3: Taming the Estimation Bias in Deep
1059
+ Reinforcement Learning. In 32nd IEEE International Conference on Tools with
1060
+ Artificial Intelligence, ICTAI 2020, Baltimore, MD, USA, November 9-11, 2020. IEEE,
1061
+ 391–398. https://doi.org/10.1109/ICTAI50040.2020.00068
1062
+ [26] Qiang He, Xinwen Hou, and Yu Liu. 2022. POPO: Pessimistic Offline Policy
1063
+ Optimization. In IEEE International Conference on Acoustics, Speech and Signal
1064
+ Processing, ICASSP 2022, Virtual and Singapore, 23-27 May 2022. IEEE, 4008–4012.
1065
+ https://doi.org/10.1109/ICASSP43922.2022.9747886
1066
+ [27] Qiang He, Huangyuan Su, Jieyu Zhang, and Xinwen Hou. 2022. Representation
1067
+ Gap in Deep Reinforcement Learning. CoRR abs/2205.14557 (2022).
1068
+ https:
1069
+ //doi.org/10.48550/arXiv.2205.14557 arXiv:2205.14557
1070
+ [28] Nicolas Heess, Greg Wayne, Yuval Tassa, Timothy Lillicrap, Martin Riedmiller,
1071
+ and David Silver. 2016. Learning and transfer of modulated locomotor controllers.
1072
+ CoRR abs/1610.05182 (2016). http://arxiv.org/abs/1610.05182
1073
+ [29] Sanjay Krishnan, Roy Fox, Ion Stoica, and Ken Goldberg. 2017. Ddco: Discovery
1074
+ of deep continuous options for robot learning from demonstrations. In Conference
1075
+ on robot learning, Vol. 78. PMLR, PMLR, 418–437. http://proceedings.mlr.press/
1076
+ v78/krishnan17a.html
1077
+ [30] Qingfeng Lan, Yangchen Pan, Alona Fyshe, and Martha White. 2020. Maxmin
1078
+ Q-learning: Controlling the Estimation Bias of Q-learning. In 8th International
1079
+ Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April
1080
+ 26-30, 2020. OpenReview.net. https://openreview.net/forum?id=Bkg0u3Etwr
1081
+ [31] Timothy P Lillicrap, Jonathan J Hunt, Alexander Pritzel, Nicolas Heess, Tom Erez,
1082
+ Yuval Tassa, David Silver, and Daan Wierstra. 2016. Continuous control with
1083
+ deep reinforcement learning. (2016).
1084
+ [32] Long-Ji Lin. 1992. Self-improving reactive agents based on reinforcement learning,
1085
+ planning and teaching. Machine Learning 8, 3 (1992), 293–321. https://doi.org/
1086
+ 10.1007/BF00992699
1087
+ [33] Marlos C. Machado, Marc G. Bellemare, and Michael Bowling. 2020. Count-
1088
+ Based Exploration with the Successor Representation. Proceedings of the AAAI
1089
+ Conference on Artificial Intelligence 34, 04, 5125–5133.
1090
+ [34] Volodymyr Mnih, Adria Puigdomenech Badia, Mehdi Mirza, Alex Graves, Tim-
1091
+ othy Lillicrap, Tim Harley, David Silver, and Koray Kavukcuoglu. 2016. Asyn-
1092
+ chronous methods for deep reinforcement learning. In International conference
1093
+ on machine learning (JMLR Workshop and Conference Proceedings, Vol. 48), Maria-
1094
+ Florina Balcan and Kilian Q. Weinberger (Eds.). PMLR, 1928–1937.
1095
+ [35] Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A. Rusu, Joel Veness,
1096
+ Marc G. Bellemare, Alex Graves, Martin A. Riedmiller, Andreas Fidjeland, Georg
1097
+ Ostrovski, Stig Petersen, Charles Beattie, Amir Sadik, Ioannis Antonoglou, Helen
1098
+ King, Dharshan Kumaran, Daan Wierstra, Shane Legg, and Demis Hassabis. 2015.
1099
+ Human-level control through deep reinforcement learning. nature 518, 7540
1100
+ (2015), 529–533.
1101
+ [36] Shakir Mohamed and Danilo Jimenez Rezende. 2015. Variational information
1102
+ maximisation for intrinsically motivated reinforcement learning. Advances in
1103
+ neural information processing systems 28 (2015), 2125–2133.
1104
+
1105
+ [37] Deepak Pathak, Pulkit Agrawal, Alexei A. Efros, and Trevor Darrell. 2017.
1106
+ Curiosity-driven Exploration by Self-supervised Prediction. In Proceedings of
1107
+ the 34th International Conference on Machine Learning (Proceedings of Machine
1108
+ Learning Research, Vol. 70). PMLR, 2778–2787.
1109
+ [38] Roberta Raileanu and Tim Rocktäschel. 2020. RIDE: Rewarding Impact-Driven
1110
+ Exploration for Procedurally-Generated Environments. In 8th International Con-
1111
+ ference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30,
1112
+ 2020. OpenReview.net.
1113
+ [39] Konrad Rawlik, Marc Toussaint, and Sethu Vijayakumar. 2012. On Stochastic
1114
+ Optimal Control and Reinforcement Learning by Approximate Inference. In
1115
+ Proceedings of Robotics: Science and Systems. Sydney, Australia. https://doi.org/
1116
+ 10.15607/RSS.2012.VIII.045
1117
+ [40] Martin Riedmiller. 2005. Neural Fitted Q Iteration – First Experiences with a Data
1118
+ Efficient Neural Reinforcement Learning Method. In Machine Learning: ECML
1119
+ 2005, João Gama, Rui Camacho, Pavel B. Brazdil, Alípio Mário Jorge, and Luís
1120
+ Torgo (Eds.). Springer Berlin Heidelberg, Berlin, Heidelberg, 317–328.
1121
+ [41] Herbert Robbins and Sutton Monro. 1951. A stochastic approximation method.
1122
+ The annals of mathematical statistics (1951), 400–407.
1123
+ [42] John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov.
1124
+ 2017. Proximal Policy Optimization Algorithms. CoRR abs/1707.06347 (2017).
1125
+ arXiv:1707.06347 http://arxiv.org/abs/1707.06347
1126
+ [43] Archit Sharma, Shixiang Gu, Sergey Levine, Vikash Kumar, and Karol Hausman.
1127
+ 2020. Dynamics-aware unsupervised discovery of skills. In 8th International
1128
+ Conference on Learning Representations, ICLR 2020. OpenReview.net, Addis Ababa,
1129
+ Ethiopia. https://openreview.net/forum?id=HJgLZR4KvH
1130
+ [44] David Silver, Guy Lever, Nicolas Heess, Thomas Degris, Daan Wierstra, and Mar-
1131
+ tin Riedmiller. 2014. Deterministic Policy Gradient Algorithms. In Proceedings of
1132
+ the 31st International Conference on International Conference on Machine Learning
1133
+ - Volume 32 (Beijing, China) (ICML’14). JMLR.org, I–387–I–395.
1134
+ [45] Richard S. Sutton. 1988. Learning to predict by the methods of temporal differ-
1135
+ ences. Machine Learning 3, 1 (1988), 9–44. https://doi.org/10.1007/BF00115009
1136
+ [46] Richard S. Sutton and Andrew G. Barto. 2018. Reinforcement learning: An intro-
1137
+ duction. MIT press.
1138
+ [47] Richard S Sutton, Doina Precup, and Satinder Singh. 1999. Between MDPs and
1139
+ semi-MDPs: A framework for temporal abstraction in reinforcement learning.
1140
+ Artificial intelligence 112, 1-2 (1999), 181–211.
1141
+ [48] Emanuel Todorov, Tom Erez, and Yuval Tassa. 2012. MuJoCo: A physics engine
1142
+ for model-based control. In 2012 IEEE/RSJ International Conference on Intelligent
1143
+ Robots and Systems. 5026–5033. https://doi.org/10.1109/IROS.2012.6386109
1144
+ [49] Laurens van der Maaten and Geoffrey Hinton. 2008. Visualizing Data using
1145
+ t-SNE. Journal of Machine Learning Research 9, 86 (2008), 2579–2605.
1146
+ http:
1147
+ //jmlr.org/papers/v9/vandermaaten08a.html
1148
+ [50] Wei Wei, Yujia Zhang, Jiye Liang, Lin Li, and Yyuze Li. 2022. Controlling Under-
1149
+ estimation Bias in Reinforcement Learning via Quasi-median Operation. Proceed-
1150
+ ings of the AAAI Conference on Artificial Intelligence 36, 8 (Jun. 2022), 8621–8628.
1151
+ https://doi.org/10.1609/aaai.v36i8.20840
1152
+ [51] Dongming Wu, Xingping Dong, Jianbing Shen, and Steven C. H. Hoi. 2020.
1153
+ Reducing Estimation Bias via Triplet-Average Deep Deterministic Policy Gradient.
1154
+ IEEE Transactions on Neural Networks and Learning Systems 31, 11 (2020), 4933–
1155
+ 4945. https://doi.org/10.1109/TNNLS.2019.2959129
1156
+ [52] Zhiwei Xu, Yunpeng Bai, Dapeng Li, Bin Zhang, and Guoliang Fan. 2022. SIDE:
1157
+ State Inference for Partially Observable Cooperative Multi-Agent Reinforcement
1158
+ Learning. In Proceedings of the 21st International Conference on Autonomous Agents
1159
+ and Multiagent Systems. International Foundation for Autonomous Agents and
1160
+ Multiagent Systems (IFAAMAS), 1400–1408.
1161
+ [53] Yingjun Ye, Xiaohui Zhang, and Jian Sun. 2019. Automated vehicle’s behavior
1162
+ decision making using deep reinforcement learning and high-fidelity simulation
1163
+ environment. Transportation Research Part C: Emerging Technologies 107 (2019),
1164
+ 155–170.
1165
+ [54] Xianjie Zhang, Yu Liu, Hangyu Mao, and Chao Yu. 2022. Common belief multi-
1166
+ agent reinforcement learning based on variational recurrent models. Neurocom-
1167
+ puting 513 (2022), 341–350. https://doi.org/10.1016/j.neucom.2022.09.144
1168
+ [55] Brian D Ziebart, Andrew L Maas, J Andrew Bagnell, Anind K Dey, et al. 2008.
1169
+ Maximum entropy inverse reinforcement learning.. In Aaai, Vol. 8. Chicago, IL,
1170
+ USA, 1433–1438.
1171
+
1172
+ Supplementary Materials
1173
+ A
1174
+ PROOF OF LEMMA 1
1175
+ (Lemma 1)(Performance Gap). Let 𝑉 ∗ be the ground truth state
1176
+ value in Bellman value iterations, 𝑄∗ be the ground truth state
1177
+ action value, 𝑉 𝜋𝑓 be the state value when applying learned policy
1178
+ 𝜋𝑓 , 𝑓 be the value function approximator. The performance gap of
1179
+ the policy between the optimal policy 𝜋∗ and the learned policy 𝜋𝑓
1180
+ is defined by an infinity norm ∥𝑉 ∗ − 𝑉 𝜋𝑓 ∥∞ and we have
1181
+ ∥𝑉 ∗ − 𝑉 𝜋𝑓 ∥∞ ≤ 2∥𝑓 −𝑄∗ ∥∞
1182
+ 1−𝛾
1183
+ Proof. For any 𝑠 ∈ S
1184
+ 𝑉 ∗(𝑠) − 𝑉 𝜋𝑓 (𝑠) =𝑄∗(𝑠, 𝜋∗(𝑠)) − 𝑄∗(𝑠, 𝜋𝑓 (𝑠))
1185
+ + 𝑄∗(𝑠, 𝜋𝑓 (𝑠)) − 𝑄∗(𝑠, 𝜋𝑓 (𝑠))
1186
+ ≤𝑄∗(𝑠, 𝜋∗(𝑠)) − 𝑓 (𝑠, 𝜋∗(𝑠))
1187
+ + 𝑓 (𝑠, 𝜋𝑓 (𝑠)) − 𝑄∗(𝑠, 𝜋𝑓 (𝑠))
1188
+ + 𝛾E𝑠′∼𝑃 (𝑠,𝜋𝑓 (𝑠)) [𝑉 ∗(𝑠′) − 𝑉 𝜋𝑓 (𝑠′)]
1189
+ ≤2∥𝑓 − 𝑄∗∥∞ + 𝛾∥𝑉 ∗ − 𝑉 𝜋𝑓 ∥∞
1190
+ B
1191
+ EXPERIMENTAL DETAILS
1192
+ B.1
1193
+ Environments
1194
+ We evaluate the performance of CCEP on environments from Mu-
1195
+ joCo Control Suite [48]which can be listed as HalfCheetah-v3, Ant-
1196
+ v3, Walker2d-v3, Humanoid-v3, Hopper-v3, and Pusher-v2, and the
1197
+ specific parameters of these environments are listed in Table 3. We
1198
+ use the publicly available environments without any modification.
1199
+ B.2
1200
+ Implementation and Hyper-parameters
1201
+ Here, we describe certain implementation details of CCEP. For
1202
+ our implementation of CCEP, we follows a standard actor-critic
1203
+ framework.
1204
+ B.3
1205
+ Soft Actor-Critic Implementation Details
1206
+ For implementation of SAC, we use the code the author provided
1207
+ and use the parameters the author recommended. We use a single
1208
+ Gaussian distribution and use the environment-dependent reward
1209
+ scaling as described by the authors. For a fair comparison, we
1210
+ apply the version of soft target update and train one iteration per
1211
+ time step. We use the reward scales as the author recommended
1212
+ (except for Pusher-v2 which is not mentioned by the author in
1213
+ the article). Considering that there are similar action dimensions
1214
+ between Pusher-v2 and HalfCheetah-v3, we set the same reward
1215
+ scale for Pusher-v2. The specific reward scales for each environment
1216
+ is shown in Table 4.
1217
+ Table 3: Environment Specific Parameters
1218
+ Environment
1219
+ State Dimensions
1220
+ Action Dimensions
1221
+ Ant-v3
1222
+ 111
1223
+ 8
1224
+ HalfCheetah-v3
1225
+ 17
1226
+ 6
1227
+ Hopper-v3
1228
+ 11
1229
+ 3
1230
+ Humanoid-v3
1231
+ 376
1232
+ 17
1233
+ Pusher-v2
1234
+ 23
1235
+ 7
1236
+ Walker2d-v3
1237
+ 17
1238
+ 6
1239
+ Table 4: SAC Environment Specific Parameters
1240
+ Environment
1241
+ Reward Scale
1242
+ Ant-v3
1243
+ 5
1244
+ HalfCheetah-v3
1245
+ 5
1246
+ Hopper-v3
1247
+ 5
1248
+ Humanoid-v3
1249
+ 20
1250
+ Pusher-v2
1251
+ 5
1252
+ Walker2d
1253
+ 5
1254
+ B.4
1255
+ Optimistic Actor-Critic Implementation
1256
+ Details
1257
+ The implementation of OAC is mainly based on the open source
1258
+ code. We set the hyper-parameters the same as OAC used in MuJoCo
1259
+ which is listed in Table 5. And for fair comparison, we train with 1
1260
+ training gradient per environment step. We use the same reward
1261
+ scales as SAC, listed in Table 4.
1262
+ Table 5: SAC Environment Specific Parameters
1263
+ Parameter
1264
+ Value
1265
+ shift multiplier
1266
+
1267
+ 2𝛿
1268
+ 6.86
1269
+ 𝛽𝑈 𝐵
1270
+ 4.66
1271
+ 𝛽𝐿𝐵
1272
+ -3.65
1273
+ B.5
1274
+ Reproducing Other Baselines
1275
+ For reproduction of TD3, we use the official implementation (
1276
+ https://github.com/sfujim/TD3). For reproduction of DDPG and
1277
+ PPO we use OpenAI’s baselines repository and apply default hyper-
1278
+ parameters.
1279
+ Table 6: CCEP Parameters settings
1280
+ Parameter
1281
+ Value
1282
+ Exploration policy
1283
+ N (0, 0.1),𝑧 ∼ 𝑝(𝑧)
1284
+ Number of policy
1285
+ 4
1286
+ Variance of exploration noise
1287
+ 0.2
1288
+ Random starting exploration time steps
1289
+ 2.5 × 104
1290
+ Optimizer
1291
+ Adam[30]
1292
+ Learning rate for actor
1293
+ 3 × 10−4
1294
+ Learning rate for critic
1295
+ 3 × 10−4
1296
+ Replay buffer size
1297
+ 1 × 106
1298
+ Batch size
1299
+ 256
1300
+ Discount (𝛾)
1301
+ 0.99
1302
+ Number of hidden layers
1303
+ 2
1304
+ Number of hidden units per layer
1305
+ 256
1306
+ Activation function
1307
+ ReLU
1308
+ Iterations per time step
1309
+ 1
1310
+ Target smoothing coefficient (𝜂)
1311
+ 5 × 10−3
1312
+ Variance of target policy smoothing
1313
+ 0.2
1314
+ Noise clip range
1315
+ [−0.5, 0.5]
1316
+ Target critic update interval
1317
+ 2
1318
+
1319
+ 0.00
1320
+ 0.25
1321
+ 0.50
1322
+ 0.75
1323
+ 1.00
1324
+ Time Steps (1e6)
1325
+ 0
1326
+ 2000
1327
+ 4000
1328
+ 6000
1329
+ 8000
1330
+ 10000
1331
+ 12000
1332
+ Average Return
1333
+ (a) HalfCheetah-v3
1334
+ 0.00
1335
+ 0.25
1336
+ 0.50
1337
+ 0.75
1338
+ 1.00
1339
+ Time Steps (1e6)
1340
+ 0
1341
+ 1000
1342
+ 2000
1343
+ 3000
1344
+ 4000
1345
+ 5000
1346
+ (b) Walker2d-v3
1347
+ 0.00
1348
+ 0.25
1349
+ 0.50
1350
+ 0.75
1351
+ 1.00
1352
+ Time Steps (1e6)
1353
+ 0
1354
+ 1000
1355
+ 2000
1356
+ 3000
1357
+ (c) Hopper-v3
1358
+ 0.00
1359
+ 0.25
1360
+ 0.50
1361
+ 0.75
1362
+ 1.00
1363
+ Time Steps (1e6)
1364
+ 0
1365
+ 1000
1366
+ 2000
1367
+ 3000
1368
+ 4000
1369
+ 5000
1370
+ 6000
1371
+ Average Return
1372
+ (d) Ant-v3
1373
+ CCEP
1374
+ CCEP-Cooperation
1375
+ TD3
1376
+ Figure 8: Ablation over the use of cooperation. Comparison of CCEP, TD3 and the subtraction of cooperation (CCEP-
1377
+ cooperation).
1378
+ 0.00
1379
+ 0.25
1380
+ 0.50
1381
+ 0.75
1382
+ 1.00
1383
+ Time Steps (1e6)
1384
+ 0
1385
+ 500
1386
+ 1000
1387
+ 1500
1388
+ 2000
1389
+ Average Return
1390
+ (a) Walker2D
1391
+ 0.00
1392
+ 0.25
1393
+ 0.50
1394
+ 0.75
1395
+ 1.00
1396
+ Time Steps (1e6)
1397
+ 0
1398
+ 1000
1399
+ 2000
1400
+ 3000
1401
+ (b) Ant
1402
+ 0.00
1403
+ 0.25
1404
+ 0.50
1405
+ 0.75
1406
+ 1.00
1407
+ Time Steps (1e6)
1408
+ 0
1409
+ 500
1410
+ 1000
1411
+ 1500
1412
+ 2000
1413
+ 2500
1414
+ (c) Hopper
1415
+ 0.00
1416
+ 0.25
1417
+ 0.50
1418
+ 0.75
1419
+ 1.00
1420
+ Time Steps (1e6)
1421
+ 1000
1422
+ 0
1423
+ 1000
1424
+ 2000
1425
+ 3000
1426
+ Average Return
1427
+ (d) HalfCheetah
1428
+ TD3
1429
+ Ours
1430
+ SAC
1431
+ DDPG
1432
+ PPO
1433
+ Figure 9: Learning curves for 4 PyBullet continuous control tasks. For better visualization, the curves are smoothed uniformly.
1434
+ The bolded line represents the average evaluation over 10 seeds. The shaded region represents the standard deviation of the
1435
+ average evaluation over 10 seeds.
1436
+ Table 7: Evaluation in PyBullet control suite. The highest average return over 10 trials of 1 million time steps. The maximum
1437
+ value for each task is bolded.
1438
+ Pybullet Environment
1439
+ Ours
1440
+ SAC
1441
+ TD3
1442
+ DDPG
1443
+ PPO
1444
+ HalfCheetah
1445
+ 2670 ± 275
1446
+ 2494 ± 266
1447
+ 2415 ± 236
1448
+ 1120 ± 373
1449
+ 465 ± 30
1450
+ Hopper
1451
+ 2254 ± 186
1452
+ 2167 ± 323
1453
+ 1860 ± 288
1454
+ 1762 ± 368
1455
+ 623 ± 131
1456
+ Walker2d
1457
+ 1829 ± 418
1458
+ 1369 ± 408
1459
+ 1676 ± 342
1460
+ 929 ± 345
1461
+ 509 ± 106
1462
+ Ant
1463
+ 3175 ± 184
1464
+ 2423 ± 680
1465
+ 2711 ± 253
1466
+ 483 ± 70
1467
+ 578 ± 19
1468
+ C
1469
+ ADDITIONAL EXPERIMENTS
1470
+ C.1
1471
+ Additional Evaluation
1472
+ For an additional Evaluation, We conduct experiments on the state-
1473
+ based PyBullet [9] suite which is based on the well-known open-
1474
+ source physics engine bullet and is packaged as a Python module for
1475
+ robot simulation and learning. The suite of Pybullet is considered
1476
+ to be a harder environment than MuJoCo [48]. We choose TD3
1477
+ [16], SAC [22], PPO [42], DDPG [31] as our baselines due to their
1478
+ superior performance. We perform interactions for 1 million steps
1479
+ in 10 different seeds and evaluate the algorithm over 10 episodes
1480
+ every 5k steps. We evaluate our algorithm in HalfCheetah, Hopper,
1481
+ Walker2d and ant in the suite of pybullet. Our results report the
1482
+ mean scores and standard deviations in the 10 seeds. We show the
1483
+ learning curves in Figure 9 and the max average return over 10
1484
+ trials in Table 7.
1485
+ C.2
1486
+ Additional Ablation Results
1487
+ We compare the learning curves of CCEP, TD3 and the subtraction
1488
+ of cooperation (CCEP-cooperation) for better understanding the
1489
+ contribution of policy cooperation (Section 5.4). We perform inter-
1490
+ actions for 1 million steps in 10 different seeds and evaluate over
1491
+ 10 episodes every 5k steps. Our results report the mean scores and
1492
+ standard deviations in the 10 seeds. We show the learning curves
1493
+ in Figure 8
1494
+ C.3
1495
+ Supplementary Results
1496
+ We provide supplementary results for Section 5.2. Figure 10 shows
1497
+ the states visited by each style over 1M time steps with intervals
1498
+ of 100k. The results show that different styles get consistent but
1499
+
1500
+ new styles emerges as well, which brings enduring exploration
1501
+ capabilities.
1502
+
1503
+ (1) Learning Steps = 100000
1504
+ (2) Learning Steps = 200000
1505
+ (3) Learning Steps = 300000
1506
+ (4) Learning Steps = 400000
1507
+ (5) Learning Steps = 500000
1508
+ (6) Learning Steps = 600000
1509
+ (7) Learning Steps = 700000
1510
+ (8) Learning Steps = 800000
1511
+ (9) Learning Steps = 900000
1512
+ POLICY No.1
1513
+ POLICY No.2
1514
+ POLICY No.3
1515
+ POLICY No.4
1516
+ Figure 10: The states visited by each style. For better visualization, the states get dimension reduction by t-SNE. The points with
1517
+ different color represents the states visited by the policy with the style. The distance between points represents the difference
1518
+ between states.
1519
+
09E0T4oBgHgl3EQfdQDB/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
0NFQT4oBgHgl3EQfDTUM/content/tmp_files/2301.13233v1.pdf.txt ADDED
@@ -0,0 +1,1466 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MNRAS 000, 1–11 (2022)
2
+ Preprint 1 February 2023
3
+ Compiled using MNRAS LATEX style file v3.0
4
+ Observations of the Planetary Nebula SMP LMC 058 with the
5
+ JWST MIRI Medium Resolution Spectrometer
6
+ O. C. Jones1★ , J. Álvarez-Márquez2 , G. C. Sloan3,4 , P. J. Kavanagh5 , I. Argyriou6 ,
7
+ A. Labiano7 , D. R. Law3 , P. Patapis8 , Michael Mueller9 , Kirsten L. Larson3 ,
8
+ Stacey N. Bright3 , P. D. Klaassen1 , O. D. Fox3
9
+ 3, Danny Gasman6
10
+ V. C. Geers1 ,
11
+ Adrian M. Glauser7 , Pierre Guillard10,11 , Omnarayani Nayak3 , A. Noriega-Crespo3 ,
12
+ Michael E. Ressler12 , B. Sargent3,13 , T. Temim14 , B. Vandenbussche6 ,
13
+ Macarena García Marín3
14
+ 1 UK Astronomy Technology Centre, Royal Observatory, Blackford Hill, Edinburgh, EH9 3HJ, UK
15
+ 2 Centro de Astrobiología (CSIC-INTA), Carretera de Ajalvir, 28850 Torrejón de Ardoz, Madrid, Spain
16
+ 3Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, USA
17
+ 4Department of Physics and Astronomy, University of North Carolina, Chapel Hill, NC 27599-3255, USA
18
+ 5Dublin Institute for Advanced Studies, School of Cosmic Physics, Astronomy & Astrophysics Section, 31 Fitzwilliam Place, Dublin 2, Ireland
19
+ 6Institute of Astronomy, KU Leuven, Celestijnenlaan 200D, 3001 Leuven, Belgium
20
+ 7Telespazio UK for the European Space Agency (ESA), ESAC, Spain
21
+ 8ETH Zurich, Institute for Particle Physics and Astrophysics, Wolfgang-Paulistr. 27, CH-8093 Zurich, Switzerland
22
+ 9Kapteyn Astronomical Institute, University of Groningen, P.O. Box 800, 9700 AV Groningen, The Netherlands
23
+ 10Sorbonne Université, CNRS, UMR 7095, Institut d’Astrophysique de Paris, 98bis bd Arago,
24
+ 75014 Paris, France
25
+ 11Institut Universitaire de France, Ministére de l’Enseignement Supérieur et de la Recherche, 1 rue Descartes, 75231 Paris Cedex 05, France
26
+ 12Jet Propulsion Laboratory, California Institute of Technology,4800 Oak Grove Drive, Pasadena, CA 91109
27
+ 13Center for Astrophysical Sciences, The William H. Miller III Department of Physics and Astronomy, Johns Hopkins University, Baltimore, Maryland 21218, USA
28
+ 14Princeton University, 4 Ivy Ln, Princeton, NJ 08544, USA
29
+ Accepted XXX. Received YYY; in original form ZZZ
30
+ ABSTRACT
31
+ During the commissioning of JWST, the Medium-Resolution Spectrometer (MRS) on the
32
+ Mid-Infrared Instrument (MIRI) observed the planetary nebula SMP LMC 058 in the Large
33
+ Magellanic Cloud. The MRS was designed to provide medium resolution (R = 𝜆/Δ𝜆) 3D
34
+ spectroscopy in the whole MIRI range. SMP LMC 058 is the only source observed in JWST
35
+ commissioning that is both spatially and spectrally unresolved by the MRS and is a good test
36
+ of JWST’s capabilities. The new MRS spectra reveal a wealth of emission lines not previously
37
+ detected in this metal-poor planetary nebula. From these lines, the spectral resolving power
38
+ (𝜆/Δ𝜆) of the MRS is confirmed to be in the range R = 4000 to 1500, depending on the MRS
39
+ spectral sub-band. In addition, the spectra confirm that the carbon-rich dust emission is from
40
+ SiC grains and that there is little to no time evolution of the SiC dust and emission line strengths
41
+ over a 16-year epoch. These commissioning data reveal the great potential of the MIRI MRS.
42
+ Key words: instrumentation: spectrographs; infrared: general; Astrophysics - Instrumentation
43
+ and Methods for Astrophysics
44
+ 1
45
+ INTRODUCTION
46
+ The succession of increasingly powerful mid-infrared spectrographs
47
+ (e.g., the Short Wavelength Spectrometer (SWS) and the Infrared
48
+ Spectrograph (IRS) on board the Infrared Space Observatory and
49
+ ★ E-mail: [email protected]
50
+ the Spitzer Space Telescope) launched into space has revolutionised
51
+ our knowledge of the cool universe (e.g., Waters et al. 1996).
52
+ The Mid-Infrared Instrument (MIRI; Wright et al. (submitted))
53
+ on the James Webb Space Telescope (JWST) includes, in addition
54
+ to the imager and coronographs, both a low-resolution spectrometer
55
+ (LRS) covering wavelengths from 5 to 14 𝜇m (Kendrew et al. 2015)
56
+ and a medium-resolution spectrometer (MRS; Wells et al. 2015,
57
+ Argyriou et al. (in prep.)), which is an Integral Field Unit (IFU),
58
+ © 2022 The Authors
59
+ arXiv:2301.13233v1 [astro-ph.IM] 30 Jan 2023
60
+
61
+ 2
62
+ O. C. Jones et al.
63
+ that has a field of view ranging from 3.2′′ × 3.7′′ to 6.6′′ × 7.7′′
64
+ (Law et al. (in prep.)) and can spatially resolve spectroscopic data
65
+ between 4.9 and 27.9 𝜇m. This is the first time a mid-IR IFU has
66
+ been deployed outside our atmosphere and will enable resolved
67
+ spectroscopic studies of individual stars at the beginning and end
68
+ of their evolution, diffuse structure in galaxies and planets.
69
+ The Large Magellanic Cloud (LMC) is a gas-rich, metal-poor,
70
+ star-forming, irregular galaxy, which is a satellite of the Milky Way,
71
+ hosts ∼103 planetary nebulae (PNe) (Reid & Parker 2010; Reid
72
+ 2014), and is at a uniform distance of ∼50 kpc (Pietrzyński et al.
73
+ 2013). In lower metallicity environments like the LMC, which has
74
+ about half the metallicity of the Milky Way (Westerlund 1997;
75
+ Choudhury et al. 2016), significant dust production is expected to
76
+ occur in the outflows of asymptotic giant branch (AGB) stars with
77
+ further processing as these objects become planetary nebula (PNe).
78
+ The gas and dust ejected into the interstellar medium (ISM) by a
79
+ strong stellar wind from this phase of evolution contains elements
80
+ synthesised in the stellar interior and dredged up to the surface by
81
+ convection (e.g., Karakas & Lattanzio 2007, 2014). Their chemical
82
+ composition is expected to primarily depend upon the initial stel-
83
+ lar mass and the interstellar elemental abundance at the time the
84
+ progenitor stars were formed (Kwok 2000; Gonçalves et al. 2014;
85
+ Kwitter & Henry 2022).
86
+ As such, the infrared spectra of PNe host a rich variety of
87
+ features; forbidden emission lines arising from ionisation from the
88
+ hot central star (e.g., Stanghellini et al. 2007), complex organic
89
+ molecules (e.g., Ziurys 2006), polycyclic aromatic hydrocarbons
90
+ (PAHs), and inorganic and organic solids (e.g., Stanghellini et al.
91
+ 2007; Bernard-Salas et al. 2009; Guzman-Ramirez et al. 2011;
92
+ García-Hernández & Górny 2014) with the frequency of carbona-
93
+ ceous features higher in the LMC than in Galactic PNe. This is likely
94
+ due to the increased efficiency of third dredge-up (TDU) and the
95
+ increased C/O ratio at low metallicities (Karakas et al. 2002). Pro-
96
+ cessing by an external ambient UV radiation field which is stronger
97
+ in the LMC (Gordon et al. 2008) may also affect the circumstellar
98
+ chemistry. Detailed examination of PNe at low-metallicity, there-
99
+ fore, provides a unique insight into chemical abundances and their
100
+ effect on late-stage stellar evolution, dust production, and the for-
101
+ mation of PNe in conditions comparable to those during the epoch
102
+ of peak star formation in the Universe (Madau et al. 1996). Fur-
103
+ thermore, due to their compact nature and brightness over a broad
104
+ wavelength range, PNe are also useful calibration sources (e.g.,
105
+ Swinyard et al. 1996; Feuchtgruber et al. 1997; Perley & Butler
106
+ 2013; Brown et al. 2014).
107
+ SMP LMC 058 was observed by JWST as part of commission-
108
+ ing and calibration activities for MIRI. First identified by Sanduleak
109
+ et al. (1978), SMP LMC 058 is a carbon-rich planetary nebula (PN)
110
+ in the LMC, with a heliocentric radial velocity of 278±7 km s−1
111
+ (Margon et al. 2020). The central star of SMP 058 is a likely C ii
112
+ emitter (Margon et al. 2020), consistent with a very early Wolf-
113
+ Rayet type star on the carbon sequence (WC). Several dozen very
114
+ strong, common emission lines of PNe were also detected in its
115
+ optical spectra (Margon et al. 2020). SMP LMC 058 has also been
116
+ observed with the Spitzer Infrared Spectrograph (IRS) at both low-
117
+ resolution (R∼60–127) and high-resolution (R∼600). The Spitzer
118
+ spectra show SMP LMC 058 has unusual dust chemistry with a
119
+ strong SiC feature at ∼11.3 𝜇m (Bernard-Salas et al. 2009) and
120
+ other associated features, including emission from PAHs at 6–9
121
+ 𝜇m, and a shoulder at 18 𝜇m from an unidentified carrier. However,
122
+ the Spitzer-IRS data show no clear evidence of fullerenes (Sloan
123
+ et al. 2014). SiC is rarely seen in Galactic PNe, in spite of the
124
+ higher Si abundance in the Milky Way compared to the Magellanic
125
+ Clouds (Jones et al. 2017). Its strength may be due to photoexci-
126
+ tation, or because at a high C/O ratio SiC forms on the surface of
127
+ carbon grains (Sloan et al. 2014).
128
+ In this paper, we describe the observations and calibration of
129
+ JWST MIRI MRS commissioning data of SMP LMC 058 (Sec-
130
+ tion 2). We then present its MRS spectra in Section 3 and determine
131
+ the resolving power of the MRS in Section 4. In Section 5 we
132
+ identify and analyse the new emission lines and solid-state features
133
+ detected in this carbon-rich planetary nebula and compare this with
134
+ Spitzer IRS data. The potential of the MRS and our conclusions are
135
+ discussed in Section 6.
136
+ 2
137
+ OBSERVATIONS AND CALIBRATIONS
138
+ The observations were taken as part of the MIRI MRS commis-
139
+ sioning program, program ID 1049 (the commissioning purpose of
140
+ these observations was PSF characterization). They use the standard
141
+ MRS observing template, with 4-point dither patterns optimized for
142
+ channels 2, 3, and 4 respectively. Each dither pattern was used twice,
143
+ in the ‘positive’ and ‘negative’ direction. Target acquisition was ac-
144
+ tivated, with the science target itself serving as an acquisition target.
145
+ All three bands (SHORT, MEDIUM, LONG) in all channels were
146
+ observed in all dithers. Simultaneous MIRI imaging in filter F770W
147
+ was taken in the dither optimized for channel 2.
148
+ A dedicated background observation was taken, employing
149
+ a 2-point dither optimised for all channels, on a field roughly 3
150
+ arcmin away. The background field was chosen to be relatively
151
+ clear of astronomical sources based on archival WISE imaging data
152
+ (Wright et al. 2010).
153
+ A total of 45 FASTR1 frames were taken per integration. In
154
+ target observations, a single integration was taken per dither point.
155
+ The background observation had two integrations (to match the total
156
+ integration time on source, accounting for the use of only a two-
157
+ point dither on the background). The integration time per MRS sub-
158
+ band and complete dither were therefore 499.5s or roughly 1,500s
159
+ to cover the entire wavelength range (bands SHORT, MEDIUM,
160
+ and LONG). Between the six dithers on-target and the single back-
161
+ ground, the total integration time was approximately 2.9 hours (6.9
162
+ hr including all overheads).
163
+ The MRS observations were processed with version 1.7.3 of
164
+ the JWST calibration pipeline and context 0995 of the Calibra-
165
+ tion Reference Data System (CRDS). In general, we follow the
166
+ standard MRS pipeline procedure (Labiano et al. 2016; Bushouse
167
+ et al. 2022; and see Álvarez-Márquez et al. 2022 for an in-flight
168
+ example of MRS data calibration). The background subtraction
169
+ has been performed using the dedicated background observation.
170
+ We have generated twelve 3D spectral cubes, one for each of
171
+ the MRS channels and bands, with a spatial and spectral sam-
172
+ pling of 0.13" × 0.13" × 0.001 𝜇m, 0.17" × 0.17" × 0.002 𝜇m,
173
+ 0.20" × 0.20" × 0.003 𝜇m, and 0.35" × 0.35" × 0.006 𝜇m for chan-
174
+ nels 1, 2, 3, and 4, respectively. We have performed 1D spectral ex-
175
+ tractions individually in each of the MRS cubes using a circular aper-
176
+ ture of radius equal to 1.5 × 𝐹𝑊𝐻𝑀(𝜆), where 𝐹𝑊𝐻𝑀(𝜆) = 0.3
177
+ arcsec for 𝜆 < 8𝜇m and 𝐹𝑊𝐻𝑀(𝜆) = 0.31 × 𝜆[𝜇𝑚]/8 arcsec for
178
+ 𝜆 > 8𝜇m. The selected FWHM (𝜆) values follow the MRS PSF
179
+ Full Width at Half Maximum (FWHM). NIRCam observation (see
180
+ Figure 1), and MRS observations suggest that SMP LMC 058 is an
181
+ unresolved source. We use the MRS PSF models (Patapis et al. in
182
+ prep.) to correct the aperture losses in the 1D spectra. The percent-
183
+ age of flux that is lost out of the selected aperture is 17% for channel
184
+ 1 and increases to 30% in channel 4.
185
+ MNRAS 000, 1–11 (2022)
186
+
187
+ JWST MRS observations of SMP LMC 058
188
+ 3
189
+ 5h24m21.5s
190
+ 21.0s
191
+ 20.5s
192
+ 20.0s
193
+ 70 04′57′′
194
+ 05′00′′
195
+ 03′′
196
+ 06′′
197
+ RA (ICRS)
198
+ Dec (ICRS)
199
+ F356W
200
+ 0.5 pc
201
+ Figure 1. NIRCam F356W image of SMP LMC 058 shown in an Asinh
202
+ stretch. At this spatial resolution (0.063′′) SMP LMC 058 is an unresolved
203
+ point source.
204
+ The 12 spectral segments extracted from these cubes were
205
+ corrected for residual fringing using a post-pipeline spectral-level
206
+ correction which is a modified version of the detector-level correc-
207
+ tion available in the JWST calibration pipeline. The residual fringe
208
+ contrasts are reduced by employing an empirical multi-component
209
+ sine fitting method (e.g. Kester et al. 2003), under the assumption
210
+ that the pipeline fringe flat correction has reduced fringe contrasts
211
+ to the point where this multi-component sine approximation is valid
212
+ (Kavanagh et al., in prep.).
213
+ Finally, each of the 12 individual spectral segments was
214
+ stitched together to remove minor flux discontinuities. This was
215
+ done by determining a scaling factor between the median flux (ex-
216
+ cluding spectral lines) in the overlapping MRS segments; then ap-
217
+ plying this multiplicative factor to the longer wavelength segments,
218
+ in turn, to effectively shift the spectrum to match the flux of its
219
+ neighbouring shorter wavelength segment. This factor was typi-
220
+ cally on the order of <5 per cent. The flux data in the overlapping
221
+ spectral regions were then averaged. The final stitched spectrum was
222
+ inspected to ensure there were no remaining discontinuities which
223
+ may affect the continuum and model fitting.
224
+ 3
225
+ SMP LMC 058 SPECTRUM
226
+ Figure 2 shows the extracted spectrum of SMP LMC 058 which
227
+ exhibits a rich variety of atomic, molecular and solid state features,
228
+ including PAHs and silicon carbide, characteristic of carbon-rich
229
+ material, and a strong continuum which rises towards the longest
230
+ wavelengths. Due to the superior sensitivity and spectral resolution
231
+ (see Section 4) of the MRS, the MIRI spectrum of SMP LMC
232
+ 058 shows features that are not seen in the Spitzer IRS data (see
233
+ Section 5), notably in the number of emission lines detected.
234
+ In the spectrum presented here, there is a large amount of fine-
235
+ structure line emission present, from the strong nebular forbidden
236
+ lines of [Ar ii], [Ar iii], [S iv], [Ne ii], [Ne iii], [S iii] to weak
237
+ H recombination lines (Hi) from the Pfund and Humphreys series,
238
+ and beyond. To ensure we measure and identify all the emission
239
+ lines in the spectra we fit a pseudo-continuum to the broadband
240
+ spectral features using a piece-wise spline model. Obvious narrow
241
+ band features were identified and masked in the fitting based on
242
+ their amplitudes exceeding a threshold value. We used an outlier
243
+ rejection fitter to flag and ignore any weaker narrow-band features
244
+ that may compromise the continuum fit. After visual inspection of
245
+ the fit, it was subtracted to isolate any narrow-band features present.
246
+ Figure 3 shows the spectrum of SMP LMC 058 after subtraction
247
+ of the pseudocontinuum from the total spectrum. The spectrum is
248
+ extremely rich in emission lines. In total 51 lines were detected.
249
+ Using the 12 original MRS segments, we identified and ana-
250
+ lyzed all detected emission lines with a signal-to-noise ratio (SNR)
251
+ greater than 3 in the SMP LMC 058 spectra. Depending on the
252
+ line profiles (see Figure 4), we performed one-component and two-
253
+ component Gaussian fits, plus a second-order polynomial to simul-
254
+ taneously fit the continuum and emission line.1 The uncertainties
255
+ on the derived emission line parameters, like the line FWHM, flux,
256
+ central wavelength, etc, were estimated using a Monte Carlo simu-
257
+ lation (following the same methodology as Álvarez-Márquez et al.
258
+ 2021, 2022). Systemic velocity shifts were removed using a he-
259
+ liocentric radial velocity of 278 ± 7 km/s (Reid & Parker 2006;
260
+ Margon et al. 2020). Table 1 presents the measured wavelengths
261
+ and fluxes together with the identification of the mid-IR emission
262
+ lines in SMP LMC 058 spectra. Small wavelength offsets are con-
263
+ sistent with known errors in the MRS FLT-4 wavelength solution
264
+ (see discussion by Argyriou et al. (in prep.)) and should be reduced
265
+ further by ongoing calibration efforts later in Cycle 1. Weak lines are
266
+ more prevalent at shorter wavelengths in channels 1 and 2 where the
267
+ MRS sensitivity is higher and the uncertainties in the flux are better
268
+ constrained. Furthermore, we find that the current MRS wavelength
269
+ calibration is <40 km/s for all spectral sub-bands, better than the
270
+ FWHM of the MRS line spread function (75-200 km/s, Labiano
271
+ et al. 2021, Argyriou et al. (in prep.)).
272
+ 4
273
+ MRS RESOLVING POWER
274
+ The resolving power (R) is defined as 𝜆/Δ𝜆, where Δ𝜆 is the mini-
275
+ mum distance to distinguish two features in a spectrum. We define
276
+ the Δ𝜆 as the FWHM of an unresolved emission line. SMP LMC
277
+ 058 is the only source observed in the JWST commissioning datasets
278
+ that is considered both spatially and spectrally unresolved with the
279
+ MRS, this makes it an excellent target for determining the inflight
280
+ MRS resolving power. Here, we assume the intrinsic width of the
281
+ emission lines in SMP LMC 058 to be negligible, as we do not have
282
+ high-resolution spectroscopy to characterize its intrinsic velocity
283
+ dispersion. Nearby planetary nebula eject gas with typical velocity
284
+ dispersions of about 10–51 kms−1 (Reid & Parker 2006). If this is
285
+ the case for SMP LMC 058, then assuming a velocity of 25 kms−1
286
+ we might underestimate the MRS resolving power by up to 5% for
287
+ channel 1, and up to 1% for channel 4 (see e.g., Law et al. 2021).
288
+ The pre-launch MRS resolving power has been established
289
+ from MIRI ground-based test and calibration campaigns, using a
290
+ set of etalons which provided lines in all MRS bands. It was de-
291
+ termined to be in the range of about 4000 to 1500 (Labiano et al.
292
+ 2021). Figure 5 shows the comparison between the ground-based
293
+ MRS resolving power estimates and the inflight estimates derived
294
+ from the SMP LMC 058 spectra. The inflight MRS resolving power
295
+ has been determined using only emission lines with SNR higher
296
+ than 6, and following the FWHM results obtained in the one- and
297
+ 1 We used the mpfit (Markwardt 2009) Python routine to perform the fits,
298
+ the code is publicly available here.
299
+ MNRAS 000, 1–11 (2022)
300
+
301
+ 4
302
+ O. C. Jones et al.
303
+ 5
304
+ 10
305
+ 15
306
+ 20
307
+ 25
308
+ Wavelength (microns)
309
+ 0.0
310
+ 0.1
311
+ 0.2
312
+ 0.3
313
+ 0.4
314
+ 0.5
315
+ Flux (Jy)
316
+ MIRI MRS
317
+ Figure 2. The MIRI MRS spectrum of SMP LMC 058. Numerous emission lines, PAH features and dust features are clearly seen on a rising continuum. These
318
+ features are much better resolved in the MRS spectra due to the higher spectral resolution.
319
+ two-component Gaussian fits (see Section 3). In the case of Hi
320
+ emission lines, we used the FWHM of the narrow gaussian compo-
321
+ nent. The errors in the resolving power are, on average, larger for
322
+ the Hi emission lines due to the uncertainty in the two-component
323
+ Gaussian fit. Given the uncertainties, the inflight MRS resolving
324
+ power agrees with the ground-based estimates, and it presents a
325
+ trend followed by the equation 4401 − 112 × 𝜆[𝜇𝑚] + 10−9×𝜆[𝜇𝑚].
326
+ The ground-based estimates consider the width of the etalon
327
+ emission lines to be negligible, which could imply an underesti-
328
+ mation of the MRS resolving power by a factor of 10%. A similar
329
+ situation is potentially happening with the inflight estimations due
330
+ to the lack of the intrinsic velocity dispersion of SMP LMC 058. We
331
+ conclude that the estimations of the ground-based MRS resolving
332
+ power (Labiano et al. 2021) are valid, within a 10% of uncertainty,
333
+ for the MRS inflight performance. As JWST observes more sources
334
+ with spatially and spectrally unresolved spectral lines, their char-
335
+ acterisation will provide a more comprehensive understanding of
336
+ the inflight variations of the resolving power within each of the 12
337
+ spectral bands. As of now, the continuous "trend" curve in Figure 5
338
+ presents the state of knowledge of the MRS resolving power.
339
+ 5
340
+ DISCUSSION
341
+ 5.1
342
+ Emission Lines
343
+ Short-High and Long-High Spitzer spectroscopic data of SMP LMC
344
+ 058 were published in Bernard-Salas et al. (2008). A comparison of
345
+ the MRS line fluxes with those found by Bernard-Salas et al. (2008)
346
+ is given in Table 2. In general, there is good agreement between our
347
+ measurements of the forbidden emission line strengths of [S iv],
348
+ [Ne ii], [Ne iii], [S iii]. Furthermore, the high-excitation lines of
349
+ [Ar v] at 13.10 𝜇m and [O iv] with ionisation potentials of 60 and
350
+ 55 eV, respectively, are not detected in either spectrum. These lines
351
+ are excited by high-temperature stars with Teff between 140,000
352
+ – 180,000 K. The highest ionisation species in the MRS spectra
353
+ are [K iv] (46 eV) and [Na iii] (47 eV), these lines have not been
354
+ previously detected by Spitzer. Thus, we consider SMP LMC 058
355
+ to be a low-excitation source.
356
+ Given the superior sensitivity of JWST (Rigby et al. 2022) and
357
+ the MRS, we detect a line at 14.38 𝜇m an order of magnitude below
358
+ the upper-limit of [Ne v] reported by Bernard-Salas et al. (2008).
359
+ The ionisation potential of [Ne v] is 97 eV, thus it is unlikely given
360
+ the absence of other high-excitation lines in the MRS spectra of SMP
361
+ LMC 058, that this emission is from [Ne v], instead, we attribute
362
+ this line to [Cl ii] which has an ionisation potential of 13 eV.
363
+ As seen in Table 1 higher ionisation potential species expand
364
+ at a lower velocity than the lower ionisation potential species (e.g.,
365
+ Reid & Parker 2006). This is due to ionisation occurring at a greater
366
+ distance from the centre of the PN where velocities are greater, and
367
+ can cause lower excitation species to expand to larger radii in the
368
+ PN.
369
+ Hydrogen recombination lines are abundant in the spectrum
370
+ of SMP LMC 058, all are new detections. Hi emission lines more
371
+ closely trace the ionized regions, compared to molecular hydrogen.
372
+ As shown in Figure 4, the line profiles of the bright Hi emission lines
373
+ are asymmetric, exhibiting a blue tail, whereas the forbidden emis-
374
+ sion lines present symmetric unresolved profiles. Hi emission lines
375
+ are composed of a spectrally unresolved main component contain-
376
+ MNRAS 000, 1–11 (2022)
377
+
378
+ JWST MRS observations of SMP LMC 058
379
+ 5
380
+ 5
381
+ 6
382
+ 7
383
+ 8
384
+ 9
385
+ Wavelength (microns)
386
+ 0.002
387
+ 0.000
388
+ 0.002
389
+ 0.004
390
+ 0.006
391
+ 0.008
392
+ 0.010
393
+ Flux (Jy)
394
+ H Hu
395
+ H Hu
396
+ [ArII]
397
+ H Pf
398
+ H Hu
399
+ H10-7
400
+ [ArIII]
401
+ [NaIII]
402
+ H16-7
403
+ H15-7
404
+ H13-7
405
+ H12-7
406
+ H15-8
407
+ H13-8
408
+ 10
409
+ 12
410
+ 14
411
+ 16
412
+ 18
413
+ 20
414
+ 22
415
+ Wavelength (microns)
416
+ 0.002
417
+ 0.000
418
+ 0.002
419
+ 0.004
420
+ 0.006
421
+ 0.008
422
+ 0.010
423
+ Flux (Jy)
424
+ [SIV]
425
+ H9-7
426
+ H Hu
427
+ [NeII]
428
+ [NeV]
429
+ [NeIII]
430
+ H10-8
431
+ [SIII]
432
+ H8-7
433
+ Figure 3. Continuum-subtracted MRS spectrum of SMP LMC 058 (where the continuum includes dust and PAH features), highlighting the atomic emission
434
+ lines. The identification of key species are marked on the spectrum. The top panel shows lines in channels 1 and 2 of the MRS, the lower panels show channels
435
+ 3 and 4. The flux axis is truncated to highlight lower contrast lines.
436
+ MNRAS 000, 1–11 (2022)
437
+
438
+ 6
439
+ O. C. Jones et al.
440
+ Table 1. Measured central wavelengths, line flux, line widths, and line identification for SMP LMC 058. The systemic velocity was removed prior to calculating
441
+ the velocity shift of a line. If a line is present in multiple MRS bands, measurements are provided for each individual MRS segment. Uncertain line identifications
442
+ are denoted by a ‘?’.
443
+ Band
444
+ Line
445
+ 𝜆lab
446
+ 𝜆observed
447
+ 𝜎𝜆observed
448
+ 𝜆offset
449
+ FWHM
450
+ Flux (×10−15)
451
+ 𝜎 (×10−15)
452
+ Identification
453
+ 𝜇m
454
+ 𝜇m
455
+ 𝜇m
456
+ km s−1
457
+ nm
458
+ erg s−1 cm−2
459
+ erg s−1 cm−2
460
+ 1S
461
+ Hi 23−7
462
+ 4.924
463
+ 4.92903
464
+ 0.00004
465
+ -46.025
466
+ 50.005
467
+ 0.11
468
+ 0.02
469
+ 1S
470
+ Hi 22−7
471
+ 4.971
472
+ 4.97588
473
+ 0.00004
474
+ -23.351
475
+ 49.997
476
+ 0.09
477
+ 0.01
478
+ 1S
479
+ Hi 21−7
480
+ 5.026
481
+ 5.03086
482
+ 0.00007
483
+ -6.428
484
+ 63.362
485
+ 0.07
486
+ 0.01
487
+ 1S
488
+ Hi 20−7
489
+ 5.091
490
+ 5.09600
491
+ 0.00003
492
+ 2.418
493
+ 50.000
494
+ 0.13
495
+ 0.01
496
+ 1S
497
+ Hi 10−6
498
+ 5.129
499
+ 5.13342
500
+ 0.00001
501
+ -0.165
502
+ 94.556
503
+ 1.67
504
+ 0.02
505
+ 1S
506
+ Hi 19−7
507
+ 5.169
508
+ 5.17401
509
+ 0.00002
510
+ 4.175
511
+ 51.268
512
+ 0.16
513
+ 0.02
514
+ 1S
515
+ Hi 18−7
516
+ 5.264
517
+ 5.26856
518
+ 0.00006
519
+ 0.558
520
+ 77.922
521
+ 0.16
522
+ 0.01
523
+ 1S
524
+ [Fe ii]
525
+ 5.340
526
+ 5.34504
527
+ 0.00011
528
+ 5.043
529
+ 92.348
530
+ 0.06
531
+ 0.01
532
+ 1S
533
+ Hi 17−7
534
+ 5.380
535
+ 5.38499
536
+ 0.00002
537
+ -12.237
538
+ 77.919
539
+ 0.18
540
+ 0.01
541
+ 1S
542
+ Hi 16−7
543
+ 5.525
544
+ 5.53072
545
+ 0.00006
546
+ -21.881
547
+ 92.047
548
+ 0.27
549
+ 0.02
550
+ 1S
551
+ Hi 15−7
552
+ 5.711
553
+ 5.71692
554
+ 0.00002
555
+ -8.044
556
+ 50.000
557
+ 0.28
558
+ 0.02
559
+ 1M
560
+ Hi 15−7
561
+ 5.711
562
+ 5.71726
563
+ 0.00005
564
+ -26.023
565
+ 89.264
566
+ 0.28
567
+ 0.02
568
+ 1M
569
+ Hi 9−6
570
+ 5.908
571
+ 5.91340
572
+ 0.00001
573
+ 15.046
574
+ 99.683
575
+ 2.31
576
+ 0.03
577
+ 1M
578
+ Hi 14−7
579
+ 5.957
580
+ 5.96199
581
+ 0.00002
582
+ 19.244
583
+ 73.871
584
+ 0.33
585
+ 0.02
586
+ 1M
587
+ [K iv]
588
+ 5.982
589
+ 5.98750
590
+ 0.0001
591
+ 2.606
592
+ 110.497
593
+ 0.12
594
+ 0.01
595
+ 1M
596
+ Hi 13−7
597
+ 6.292
598
+ 6.29807
599
+ 0.00005
600
+ -14.725
601
+ 74.449
602
+ 0.43
603
+ 0.04
604
+ 1L
605
+ Hi 12−7
606
+ 6.772
607
+ 6.77852
608
+ 0.00002
609
+ -10.975
610
+ 85.586
611
+ 0.58
612
+ 0.02
613
+ 1L
614
+ Hi 21−8
615
+ 6.826
616
+ 6.83241
617
+ 0.00013
618
+ -8.356
619
+ 77.331
620
+ 0.06
621
+ 0.02
622
+ 1L
623
+ H2(0,0) S(5)
624
+ 6.910
625
+ 6.91588
626
+ 0.00013
627
+ 2.424
628
+ 95.459
629
+ 0.13
630
+ 0.02
631
+ 1L
632
+ Hi 20−8
633
+ 6.947
634
+ 6.95306
635
+ 0.00009
636
+ 6.515
637
+ 75.574
638
+ 0.08
639
+ 0.02
640
+ 1L
641
+ [Ar ii]
642
+ 6.985
643
+ 6.99181
644
+ 0.00001
645
+ -2.238
646
+ 89.682
647
+ 1.17
648
+ 0.01
649
+ 1L
650
+ Hi 19−8
651
+ 7.093
652
+ 7.09935
653
+ 0.00017
654
+ -2.424
655
+ 76.633
656
+ 0.11
657
+ 0.01
658
+ 1L
659
+ Hi 18−8
660
+ 7.272
661
+ 7.27877
662
+ 0.00011
663
+ -14.968
664
+ 69.767
665
+ 0.10
666
+ 0.02
667
+ 1L
668
+ [Na iii]
669
+ 7.318
670
+ 7.32485
671
+ 0.00006
672
+ -14.817
673
+ 136.625
674
+ 0.46
675
+ 0.02
676
+ 1L
677
+ Hi 6−5
678
+ 7.460
679
+ 7.46690
680
+ 0
681
+ -4.607
682
+ 79.653
683
+ 11.51
684
+ 0.04
685
+ 1L
686
+ Hi 8−6
687
+ 7.502
688
+ 7.50949
689
+ 0.00001
690
+ -1.396
691
+ 79.577
692
+ 12.33
693
+ 0.07
694
+ 1L
695
+ Hi 11−7
696
+ 7.508
697
+ 7.51515
698
+ 0.00003
699
+ -2.922
700
+ 74.553
701
+ 0.69
702
+ 0.02
703
+ 2S
704
+ Hi 15−8
705
+ 8.155
706
+ 8.16375
707
+ 0.00053
708
+ -47.407
709
+ 49.995
710
+ 0.15
711
+ 0.05
712
+ 2S
713
+ Hi 14−8
714
+ 8.665
715
+ 8.67220
716
+ 0.00019
717
+ 11.971
718
+ 53.004
719
+ 0.28
720
+ 0.03
721
+ 2M
722
+ Hi 10−7
723
+ 8.760
724
+ 8.76815
725
+ 0.00006
726
+ 1.510
727
+ 110.130
728
+ 1.04
729
+ 0.05
730
+ 2M
731
+ [Ar iii]
732
+ 8.991
733
+ 8.99859
734
+ 0
735
+ 37.731
736
+ 101.834
737
+ 20.61
738
+ 0.07
739
+ 2M
740
+ Hi 13−8
741
+ 9.392
742
+ 9.40091
743
+ 0.00012
744
+ -5.539
745
+ 84.050
746
+ 0.25
747
+ 0.04
748
+ 2M
749
+ H2(0,0) S(3)
750
+ 9.665
751
+ 9.67502
752
+ 0.00023
753
+ -35.450
754
+ 151.677
755
+ 0.23
756
+ 0.04
757
+ 2M
758
+ Hi 18−9
759
+ 9.847
760
+ 9.85887
761
+ 0.00037
762
+ -82.099
763
+ 75.445
764
+ 0.06
765
+ 0.01
766
+ 2L
767
+ [S iv]
768
+ 10.511
769
+ 10.52014
770
+ 0.00001
771
+ 3.428
772
+ 96.022
773
+ 30.07
774
+ 0.12
775
+ 2L
776
+ Hi 16−9
777
+ 10.804
778
+ 10.81253
779
+ 0.00068
780
+ 30.378
781
+ 140.073
782
+ 0.13
783
+ 0.03
784
+ 2L
785
+ Hi 9−7
786
+ 11.309
787
+ 11.31929
788
+ 0.00025
789
+ -2.583
790
+ 82.684
791
+ 1.41
792
+ 0.17
793
+ 3S
794
+ Hi 7−6
795
+ 12.372
796
+ 12.38265
797
+ 0.00003
798
+ 17.586
799
+ 94.410
800
+ 2.90
801
+ 0.03
802
+ 3S
803
+ Hi 11−8
804
+ 12.387
805
+ 12.39758
806
+ 0.00005
807
+ 26.303
808
+ 107.676
809
+ 0.68
810
+ 0.02
811
+ 3S
812
+ Hi 14−9
813
+ 12.587
814
+ 12.59863
815
+ 0.00023
816
+ 3.125
817
+ 93.158
818
+ 0.15
819
+ 0.03
820
+ 3S
821
+ [Ne ii]
822
+ 12.814
823
+ 12.82631
824
+ 0
825
+ -20.232
826
+ 98.379
827
+ 17.63
828
+ 0.02
829
+ 3M
830
+ Hi 13−9
831
+ 14.183
832
+ 14.19616
833
+ 0.00035
834
+ 1.955
835
+ 69.993
836
+ 0.13
837
+ 0.05
838
+ 3M
839
+ [Cl ii]?
840
+ 14.368
841
+ 14.38034
842
+ 0.00019
843
+ 16.544
844
+ 82.249
845
+ 0.34
846
+ 0.02
847
+ 3M
848
+ Hi 16−10
849
+ 14.962
850
+ 14.97556
851
+ 0.00064
852
+ 11.773
853
+ 50.002
854
+ 0.04
855
+ 0.02
856
+ 3L
857
+ [Ne iii]
858
+ 15.555
859
+ 15.56957
860
+ 0.00001
861
+ -0.617
862
+ 130.542
863
+ 179.26
864
+ 0.55
865
+ 3L
866
+ Hi 10−8
867
+ 16.209
868
+ 16.22334
869
+ 0.00014
870
+ 14.869
871
+ 97.309
872
+ 0.60
873
+ 0.08
874
+ 3L
875
+ Hi 12−9
876
+ 16.881
877
+ 16.89664
878
+ 0.0002
879
+ -6.106
880
+ 91.592
881
+ 0.29
882
+ 0.04
883
+ 4S
884
+ [S iii]
885
+ 18.713
886
+ 18.72914
887
+ 0.00002
888
+ 19.709
889
+ 136.931
890
+ 11.34
891
+ 0.07
892
+ 4S
893
+ Hi 8−7
894
+ 19.062
895
+ 19.07898
896
+ 0.00005
897
+ 9.654
898
+ 148.238
899
+ 2.40
900
+ 0.05
901
+ 4M
902
+ [Ar iii]
903
+ 21.830
904
+ 21.85033
905
+ 0.00032
906
+ 1.852
907
+ 155.363
908
+ 0.82
909
+ 0.06
910
+ 4M
911
+ Hi 13−10+Hi 11−9
912
+ 22.340
913
+ 22.35612
914
+ 0.00055
915
+ 68.040
916
+ 194.363
917
+ 0.64
918
+ 0.05
919
+ ing the majority of the line flux (>95%)), and a spectrally resolved
920
+ blue-shifted component possibly due to thermal broadening (Chu
921
+ et al. 1984) or from condensation outside the main core which may
922
+ be evident as a marginally resolved envelope like structure in the
923
+ MRS cube at 7.466𝜇m.
924
+ Two molecular hydrogen lines (H2) have been detected in the
925
+ MRS data, the ortho-H2 𝑣 = 0–0 S(3) and S(5) lines. The S(1) line at
926
+ 17.055 𝜇m may also be present, although this is not easily discerned
927
+ above the continuum and we do not measure its flux. The S(3) and
928
+ S(5) rotational line emission probably originate from irradiated,
929
+ and perhaps also shocked, dense molecular clumps, torus structures
930
+ (e.g., Kastner et al. 1996; Hora et al. 1999; Akras et al. 2017;
931
+ MNRAS 000, 1–11 (2022)
932
+
933
+ JWST MRS observations of SMP LMC 058
934
+ 7
935
+ 7.455
936
+ 7.460
937
+ 7.465
938
+ 7.470
939
+ 7.475
940
+ Wavelength (microns)
941
+ 0.00
942
+ 0.02
943
+ 0.04
944
+ 0.06
945
+ 0.08
946
+ 0.10
947
+ Flux (Jy)
948
+ Pf
949
+ 15.54
950
+ 15.56
951
+ 15.58
952
+ 15.60
953
+ Wavelength (microns)
954
+ 0.00
955
+ 0.25
956
+ 0.50
957
+ 0.75
958
+ 1.00
959
+ 1.25
960
+ 1.50
961
+ 1.75
962
+ 2.00
963
+ Flux (Jy)
964
+ [Ne III]
965
+ Figure 4. Top: The Pf 𝛼 H i emission line profile shows a spatially unre-
966
+ solved main component and a weaker spectrally resolved blue-shifted wing.
967
+ Bottom: The [Ne iii] line profile is spectrally unresolved and symmetric.
968
+ This shape is typical of all the forbidden emission lines in SMP LMC 058.
969
+ The dashed line marks the lines observed central wavelength.
970
+ Table 2. Comparison of SMP LMC 058 MRS line fluxes with those of
971
+ Bernard-Salas et al. (2008) taken with the high-resolution modules on the
972
+ Spitzer IRS. All line strengths reported by Bernard-Salas et al. (2008) have
973
+ a 10% error except for [S iv] which has a 10–20% error. Errors in the MRS
974
+ flux are <1% and are provided for each line in Table 1.
975
+ Line
976
+ Wavelength
977
+ MRS Flux
978
+ Spitzer Flux
979
+ Ionisation
980
+ (Rest)
981
+ ×10−15
982
+ ×10−15
983
+ potential
984
+ 𝜇m
985
+ erg s−1 cm−2
986
+ erg s−1 cm−2
987
+ (eV)
988
+ [S iv]
989
+ 10.511
990
+ 30.07
991
+ 29.2
992
+ 35
993
+ [Ne ii]
994
+ 12.814
995
+ 17.63
996
+ 20.6
997
+ 22
998
+ [Ar v]
999
+ 13.099
1000
+ <0.029
1001
+ <2.4
1002
+ 60
1003
+ [Ne v]
1004
+ 14.323
1005
+ <0.005
1006
+ <3.8
1007
+ 97
1008
+ [Ne iii]
1009
+ 15.555
1010
+ 179.26
1011
+ 200.6
1012
+ 41
1013
+ [S iii]
1014
+ 18.713
1015
+ 11.34
1016
+ 11.0
1017
+ 23
1018
+ [O iv]
1019
+ 25.883
1020
+ <0.23
1021
+ <21.6
1022
+ 55
1023
+ Fang et al. 2018), or from the outer regions of the PNe where the
1024
+ temperature is about 1000K (Aleman & Gruenwald 2004; Matsuura
1025
+ et al. 2007b).
1026
+ 5.2
1027
+ Dust and PAH Features
1028
+ The dust in SMP LMC 058 is carbon-rich. Amongst the most promi-
1029
+ nent features is the strong silicon carbide (SiC) emission at 11 𝜇m
1030
+ and the rising continuum due to the thermal emission of warm dust.
1031
+ Strong emission features from PAHs also appear in the spectrum at
1032
+ 5.2, 5.7, 6.2, 7.7, 8.6, 11.2 and 12.7 𝜇m.
1033
+ At sub-solar metallicities (∼ 0.2 − 0.5 Z⊙), SiC is commonly
1034
+ observed in PNe, yet it is rarely seen in Galactic PNe or indeed
1035
+ during the earlier AGB evolutionary phase of metal-poor carbon
1036
+ stars (Casassus et al. 2001; Zijlstra et al. 2006; Matsuura et al.
1037
+ 2007a; Stanghellini et al. 2007; Bernard-Salas et al. 2008; Woods
1038
+ et al. 2011, 2012; Sloan et al. 2014; Ruffle et al. 2015; Jones et al.
1039
+ 2017). The strength of the SiC flux in metal-poor PNe is highly
1040
+ sensitive to the radiation field (Bernard-Salas et al. 2009). This is
1041
+ likely due to a lower abundance of Si affecting the carbonaceous dust
1042
+ condensation sequence on the AGB. In this case, rather than SiC
1043
+ forming first, it instead forms in a mantle surrounding an amorphous
1044
+ carbon core (Lagadec et al. 2007; Leisenring et al. 2008). Then as
1045
+ the PNe dust becomes heated and photo-processed, the amorphous
1046
+ carbon evaporates increasing the SiC surface area and consequently
1047
+ its feature strength, until a critical ionisation potential of >55 eV
1048
+ occurs at which point the SiC features disappear (Bernard-Salas
1049
+ et al. 2009; Sloan et al. 2014).
1050
+ Following Bernard-Salas et al. (2009), we measure the strength
1051
+ of the SiC feature by integrating the flux above a continuum-
1052
+ subtracted spectrum from 9 to 13.2 𝜇m and then subtracting the flux
1053
+ contributions from the 11.2 𝜇m PAH feature and the [Ne ii] line.
1054
+ Due to the resolution of the MRS compared to the Spitzer spectra of
1055
+ SMP LMC 058, we detect several additional lines which contribute
1056
+ to the integrated flux in the SiC region; these lines include [S iv]
1057
+ and H i. Thus to obtain a reliable measurement of the SiC feature
1058
+ strength we also subtract the flux contribution from all emission
1059
+ lines in the 9 – 13.2 𝜇m region listed in Table 1. A PAH feature at
1060
+ ∼12.6 𝜇m likely contributes a small amount of flux to the measured
1061
+ SiC feature, however isolating and subtracting this emission contri-
1062
+ bution from the wing of the SiC feature is challenging even with
1063
+ the MRS spectral resolution. Additionally, an artefact at ∼12.2 𝜇m
1064
+ due to a spectral leak (e.g., Gasman et al. 2022) may also affect the
1065
+ integrated flux. Table 3 gives the measured SiC centroid and cor-
1066
+ rected feature strength. The latter agrees exceptionally well with the
1067
+ value of 29.72 ± 0.31 ×10−16 W m−2 measured by Bernard-Salas
1068
+ et al. (2009) in the Spitzer data of SMP LMC 058. This suggests
1069
+ there is little to no evolution in the SiC dust on the 16-year time
1070
+ scales between the observations. Furthermore, the agreement be-
1071
+ tween the measurements verifies the overall flux calibration of the
1072
+ MRS instrument (Gasman et al. 2022).
1073
+ In astronomical sources, the structure, wavelengths and relative
1074
+ strength of the PAHs can differ strongly between objects, with PNe
1075
+ showing the most pronounced variations in PAH profiles due to
1076
+ photoprocessing altering the ratio of aliphatics to aromatics (Peeters
1077
+ et al. 2002; Pino et al. 2008; Matsuura et al. 2014; Sloan et al. 2014;
1078
+ Jensen et al. 2022). Figure 6 shows the PAHs in SMP LMC 058. The
1079
+ PAHs in SMP LMC 058 are considered to have a class B profile by
1080
+ Bernard-Salas et al. (2009) and Sloan et al. (2014). In this schema
1081
+ devised by Peeters et al. (2002) and van Diedenhoven et al. (2004)
1082
+ the 6.2 PAH feature for class B objects has a peak between 6.24
1083
+ and 6.28 𝜇m; the dominant 7.7 PAH feature peaks between 7.8 to
1084
+ MNRAS 000, 1–11 (2022)
1085
+
1086
+ 8
1087
+ O. C. Jones et al.
1088
+ 4.5
1089
+ 5
1090
+ 6
1091
+ 7
1092
+ 8
1093
+ 9
1094
+ 10
1095
+ 12
1096
+ 15
1097
+ 20
1098
+ 25
1099
+ 30
1100
+ Wavelength [ m]
1101
+ 1000
1102
+ 1500
1103
+ 2000
1104
+ 2500
1105
+ 3000
1106
+ 3500
1107
+ 4000
1108
+ 4500
1109
+ 5000
1110
+ MRS Resolving Power
1111
+ Ground (Labiano+21)
1112
+ Fit to inflight lines
1113
+ inflight forbidden lines
1114
+ inflight HI lines
1115
+ Figure 5. Comparison between the ground-based and inflight MRS resolving power. Gray filled area and black line: ground-based MRS resolving power
1116
+ estimates (Labiano et al. 2021). Filled red circles: inflight MRS resolving power calculation using forbidden emission lines identified in this paper. Open red
1117
+ circles: inflight MRS resolving power calculation using Hi emission lines.
1118
+ 6
1119
+ 8
1120
+ 10
1121
+ 12
1122
+ 14
1123
+ Wavelength (microns)
1124
+ 0.000
1125
+ 0.025
1126
+ 0.050
1127
+ 0.075
1128
+ 0.100
1129
+ 0.125
1130
+ 0.150
1131
+ 0.175
1132
+ 0.200
1133
+ Flux (Jy)
1134
+ PAH 5.2 m
1135
+ PAH 5.7 m
1136
+ PAH 6.2 m
1137
+ PAH 7.7 m
1138
+ PAH 8.6 m
1139
+ PAH 11.2 m
1140
+ SiC
1141
+ Local continuum
1142
+ Figure 6. The SiC and PAH features are highlighted in the spectra of SMP LMC 058. A local continuum fit to the 11.3 𝜇m feature which is superimposed
1143
+ on the broad SiC emission feature is also shown. The colours highlight the spectral region for each feature, to which a local continuum was fit and the flux
1144
+ measured over.
1145
+ MNRAS 000, 1–11 (2022)
1146
+
1147
+ JWST MRS observations of SMP LMC 058
1148
+ 9
1149
+ Table 3. PAH and SiC Fluxes and Centroids.
1150
+ Centroid
1151
+ Integrated Flux
1152
+ Integrated Flux Error
1153
+ 𝜇m
1154
+ W m−2
1155
+ W m−2
1156
+ PAH
1157
+ 5.262
1158
+ 1.64×10−18
1159
+ 7.0×10−20
1160
+ PAH
1161
+ 5.698
1162
+ 5.62×10−18
1163
+ 1.1×10−19
1164
+ PAH
1165
+ 6.274
1166
+ 9.259×10−17
1167
+ 2.9×10−19
1168
+ PAH
1169
+ 7.834
1170
+ 3.427×10−16
1171
+ 1.2×10−18
1172
+ PAH
1173
+ 8.665
1174
+ 6.231×10−17
1175
+ 8.5×10−19
1176
+ PAH
1177
+ 11.298
1178
+ 1.047×10−16
1179
+ 2.0×10−18
1180
+ SiC
1181
+ 11.097
1182
+ 3.067×10−15
1183
+ 4.5×10−18
1184
+ 8.0 𝜇m; and the 8.6 PAH band is red-shifted. These values agree
1185
+ well with our measured centroids listed in Table 3. Furthermore, the
1186
+ PAHs observed in SMP LMC 058 closely resemble those observed
1187
+ in the ISO SWS spectrum of the Galactic post-AGB star, HD 44179
1188
+ (the Red Rectangle) which also shows strong aromatic features on
1189
+ top of a continuum (Waters et al. 1998).
1190
+ The relative strength of the PAH features depends on a number
1191
+ of factors including the degree of ionisation of the radiation field
1192
+ (e.g., Allamandola et al. 1999). The strength of the PAH features
1193
+ in SMP LMC 058 was determined by integrating the flux of the
1194
+ feature above an adopted local continuum, fit to each side of the
1195
+ feature and measured using specutils line_flux. Particular care
1196
+ was taken in fitting a continuum, too, and then measuring the 11.25
1197
+ 𝜇m band (produced by the out-of-plane solo C–H bending mode)
1198
+ as this is superimposed on top of the broad SiC feature. Table 3
1199
+ presents the central wavelength of the features and the integrated
1200
+ flux. The ratio of the PAH strengths correlates with the source type
1201
+ and hence its physical conditions (Hony et al. 2001); ionized PAHs
1202
+ have strong features at 6.2, 7.7 and 8.6 𝜇m whilst the 11.2 𝜇m PAH
1203
+ feature is stronger for neutral PAHs. From the PAH line strengths
1204
+ given in Table 3 it is evident that the 7.7𝜇m feature dominates the
1205
+ total PAH emission, and thus the dust around SMP LMC 058 is
1206
+ likely experiencing a high degree of ionisation.
1207
+ Carbon-rich PNe can show a rich variety of solid-state material
1208
+ in their spectra in addition to PAHs. The C60 fullerene molecule
1209
+ typically exhibits features at ∼7.0, 8.5, 17.4 and 18.9 𝜇m, and all
1210
+ four were first identified in the spectrum of the Galactic PN TC-1
1211
+ (Cami et al. 2010). Fullerenes have since been detected in several
1212
+ other PNe (e.g., García-Hernández et al. 2010, 2011; Sloan et al.
1213
+ 2014). The still-unidentified 21 𝜇m emission feature, first detected
1214
+ by Kwok et al. 1989, can also appear in carbon-rich PNe, often
1215
+ associated with unusual PAH emission and aliphatic hydrocarbons
1216
+ (Cerrigone et al. 2011; Matsuura et al. 2014; Sloan et al. 2014; Volk
1217
+ et al. 2020).
1218
+ The spectra of SMP LMC 058 from the IRS on Spitzer did
1219
+ not show any of these unusual hydrocarbon-related features, but the
1220
+ improved spectral resolution of the MRS allows for a much more
1221
+ careful examination. Nonetheless, these additional features remain
1222
+ too weak to be detected. SMP LMC 058 presents a classic Class
1223
+ B PAH spectrum, as expected for objects which have evolved to
1224
+ the young PN stage (Sloan et al. 2014). Younger objects which
1225
+ could still be described as post-AGB objects would show the 21 𝜇m
1226
+ feature and/or aliphatics. Sloan et al. (2014) identified SMP LMC
1227
+ 058 as a member of the Big-11 group because of the combination
1228
+ of a strong SiC emission feature and the 11.2 𝜇m PAH feature and
1229
+ the absence of fullerenes. This group is actually related to the PNe
1230
+ that show fullerenes, and the presence or absence of fullerenes may
1231
+ be due to something as simple as which have a clear line of sight to
1232
+ the interior of the dust shells where the fullerenes are expected to
1233
+ be present.
1234
+ 6
1235
+ SUMMARY AND CONCLUSIONS
1236
+ We have presented MIRI/MRS spectra of the carbon-rich planetary
1237
+ nebula SMP LMC 058 located in the metal-poor Large Magellanic
1238
+ Cloud. SMP LMC 058 is a point source in the MRS data and
1239
+ its spectrum contains the only spatially and spectrally unresolved
1240
+ emission lines observed during the commissioning of the JWST
1241
+ Medium-Resolution Spectrometer. In the MRS spectrum, we de-
1242
+ tected 51 emission lines, of which 47 were previously undetected
1243
+ in this source. The strongest emission lines were used to determine
1244
+ the spectral resolutions of the MIRI MRS instrument. The resolving
1245
+ power is R > 3960 in channel 1, R > 3530 in channel 2, R > 3200
1246
+ in channel 3, and R > 1920 in channel 4. This on-sky performance
1247
+ is comparable to the resolution determined from the ground cali-
1248
+ bration of the MRS which provides resolving powers from 4000 at
1249
+ channel 1 to 1500 at channel 4. Furthermore, a comparison of the
1250
+ line strengths and spectral continuum to previous observations of
1251
+ SMP LMC 058 with the IRS on the Spitzer was used to verify the
1252
+ absolute flux calibration of the MRS instrument. The MRS spectra
1253
+ confirm that the carbon-rich dust emission is from grains and not
1254
+ isolated molecules and that there is little to no time evolution of the
1255
+ SiC dust and emission line strengths in the 16 years between the
1256
+ observations. The PAH emission is dominated by the 7.7𝜇m feature.
1257
+ The strong PAHs and SiC in the spectra are consistent with the lack
1258
+ of high-excitation lines detected in the spectra, which if present,
1259
+ would indicate a hard radiation field that would likely destroy these
1260
+ grains. These commissioning data reveal the great potential and
1261
+ resolving power of the MIRI MRS to study line, molecular and
1262
+ solid-state features in individual sources in nearby galaxies.
1263
+ ACKNOWLEDGEMENTS
1264
+ We thank Kay Justtanont for her insights, comments and dis-
1265
+ cussions. This work is based on observations made with the
1266
+ NASA/ESA/CSA James Webb Space Telescope. The data were ob-
1267
+ tained from the Mikulski Archive for Space Telescopes at the Space
1268
+ Telescope Science Institute, which is operated by the Association of
1269
+ Universities for Research in Astronomy, Inc., under NASA contract
1270
+ NAS 5-03127 for JWST. These observations are associated with
1271
+ program #1049. This work is based in part on observations made
1272
+ with the Spitzer Space Telescope, which was operated by the Jet
1273
+ Propulsion Laboratory, California Institute of Technology under a
1274
+ contract with NASA
1275
+ O.C.J acknowledge support from an STFC Webb fellowship.
1276
+ J.A.M. and A.L acknowledge support by grant PIB2021-127718NB-
1277
+ 100 by the Spanish Ministry of Science and Innovation/State Agency
1278
+ of Research (MCIN/AEI). P.J.K acknowledges financial support
1279
+ from the Science Foundation Ireland/Irish Research Council Path-
1280
+ way programme under Grant Number 21/PATH-S/9360. I.A., D.G.,
1281
+ and B.V. thank the European Space Agency (ESA) and the Belgian
1282
+ Federal Science Policy Office (BELSPO) for their support in the
1283
+ framework of the PRODEX Programme. PG would like to thank
1284
+ the University Pierre and Marie Curie, the Institut Universitaire de
1285
+ France, the Centre National d’Etudes Spatiales (CNES), the "Pro-
1286
+ gramme National de Cosmologie and Galaxies" (PNCG) and the
1287
+ "Physique Chimie du Milieu Interstellaire" (PCMI) programs of
1288
+ MNRAS 000, 1–11 (2022)
1289
+
1290
+ 10
1291
+ O. C. Jones et al.
1292
+ CNRS/INSU, with INC/INP co-funded by CEA and CNES, for
1293
+ there financial supports.
1294
+ MIRI draws on the scientific and technical expertise of the
1295
+ following organisations: Ames Research Center, USA; Airbus De-
1296
+ fence and Space, UK; CEA-Irfu, Saclay, France; Centre Spatial
1297
+ de Liége, Belgium; Consejo Superior de Investigaciones Científi-
1298
+ cas, Spain; Carl Zeiss Optronics, Germany; Chalmers University of
1299
+ Technology, Sweden; Danish Space Research Institute, Denmark;
1300
+ Dublin Institute for Advanced Studies, Ireland; European Space
1301
+ Agency, Netherlands; ETCA, Belgium; ETH Zurich, Switzerland;
1302
+ Goddard Space Flight Center, USA; Institute d’Astrophysique Spa-
1303
+ tiale, France; Instituto Nacional de Técnica Aeroespacial, Spain; In-
1304
+ stitute for Astronomy, Edinburgh, UK; Jet Propulsion Laboratory,
1305
+ USA; Laboratoire d’Astrophysique de Marseille (LAM), France;
1306
+ Leiden University, Netherlands; Lockheed Advanced Technology
1307
+ Center (USA); NOVA Opt-IR group at Dwingeloo, Netherlands;
1308
+ Northrop Grumman, USA; Max-Planck Institut für Astronomie
1309
+ (MPIA), Heidelberg, Germany; Laboratoire d’Etudes Spatiales et
1310
+ d’Instrumentation en Astrophysique (LESIA), France; Paul Scher-
1311
+ rer Institut, Switzerland; Raytheon Vision Systems, USA; RUAG
1312
+ Aerospace, Switzerland; Rutherford Appleton Laboratory (RAL
1313
+ Space), UK; Space Telescope Science Institute, USA; Toegepast-
1314
+ Natuurwetenschappelijk Onderzoek (TNO-TPD), Netherlands; UK
1315
+ Astronomy Technology Centre, UK; University College London,
1316
+ UK; University of Amsterdam, Netherlands; University of Arizona,
1317
+ USA; University of Bern, Switzerland; University of Cardiff, UK;
1318
+ University of Cologne, Germany; University of Ghent; University
1319
+ of Groningen, Netherlands; University of Leicester, UK; University
1320
+ of Leuven, Belgium; University of Stockholm, Sweden; Utah State
1321
+ University, USA. A portion of this work was carried out at the Jet
1322
+ Propulsion Laboratory, California Institute of Technology, under a
1323
+ contract with the National Aeronautics and Space Administration.
1324
+ The following National and International Funding Agencies
1325
+ funded and supported the MIRI development: NASA; ESA; Bel-
1326
+ gian Science Policy Office (BELSPO); Centre Nationale d’Etudes
1327
+ Spatiales (CNES); Danish National Space Centre; Deutsches Zen-
1328
+ trum fur Luftund Raumfahrt (DLR); Enterprise Ireland; Ministerio
1329
+ De Economia y Competividad; Netherlands Research School for As-
1330
+ tronomy (NOVA); Netherlands Organisation for Scientific Research
1331
+ (NWO); Science and Technology Facilities Council; Swiss Space
1332
+ Office; Swedish National Space Agency; and UK Space Agency.
1333
+ Facilities: JWST (MIRI/MRS) - James Webb Space Telescope.
1334
+ DATA AVAILABILITY
1335
+ JWST data were obtained from the Mikulski Archive for
1336
+ Space Telescopes at the Space Telescope Science Institute
1337
+ (https://archive.stsci.edu/).
1338
+ REFERENCES
1339
+ Akras S., Gonçalves D. R., Ramos-Larios G., 2017, MNRAS, 465, 1289
1340
+ Aleman I., Gruenwald R., 2004, ApJ, 607, 865
1341
+ Allamandola L. J., Hudgins D. M., Sandford S. A., 1999, ApJ, 511, L115
1342
+ Álvarez-Márquez J., Marques-Chaves R., Colina L., Pérez-Fournon I., 2021,
1343
+ A&A, 647, A133
1344
+ Álvarez-Márquez J., et al., 2022, arXiv e-prints, p. arXiv:2209.01695
1345
+ Bernard-Salas J., Pottasch S. R., Gutenkunst S., Morris P. W., Houck J. R.,
1346
+ 2008, ApJ, 672, 274
1347
+ Bernard-Salas J., Peeters E., Sloan G. C., Gutenkunst S., Matsuura M.,
1348
+ Tielens A. G. G. M., Zijlstra A. A., Houck J. R., 2009, ApJ, 699, 1541
1349
+ Brown M. J. I., Jarrett T. H., Cluver M. E., 2014, Publ. Astron. Soc. Australia,
1350
+ 31, e049
1351
+ Bushouse
1352
+ H.,
1353
+ et
1354
+ al.,
1355
+ 2022,
1356
+ spacetelescope/jwst:
1357
+ JWST
1358
+ 1.6.2,
1359
+ doi:10.5281/zenodo.6984366,
1360
+ https://doi.org/10.5281/
1361
+ zenodo.6984366
1362
+ Cami J., Bernard-Salas J., Peeters E., Malek S. E., 2010, Science, 329, 1180
1363
+ Casassus S., Roche P. F., Aitken D. K., Smith C. H., 2001, MNRAS, 327,
1364
+ 744
1365
+ Cerrigone L., Hora J. L., Umana G., Trigilio C., Hart A., Fazio G., 2011,
1366
+ ApJ, 738, 121
1367
+ Choudhury S., Subramaniam A., Cole A. A., 2016, MNRAS, 455, 1855
1368
+ Chu Y. H., Kwitter K. B., Kaler J. B., Jacoby G. H., 1984, PASP, 96, 598
1369
+ Fang X., Zhang Y., Kwok S., Hsia C.-H., Chau W., Ramos-Larios G., Guer-
1370
+ rero M. A., 2018, ApJ, 859, 92
1371
+ Feuchtgruber H., et al., 1997, ApJ, 487, 962
1372
+ García-Hernández D. A., Górny S. K., 2014, A&A, 567, A12
1373
+ García-Hernández D. A., Manchado A., García-Lario P., Stanghellini L.,
1374
+ Villaver E., Shaw R. A., Szczerba R., Perea-Calderón J. V., 2010, ApJ,
1375
+ 724, L39
1376
+ García-Hernández D. A., et al., 2011, ApJ, 737, L30
1377
+ Gasman D., et al., 2022, arXiv e-prints, p. arXiv:2212.03596
1378
+ Gonçalves D. R., Magrini L., Teodorescu A. M., Carneiro C. M., 2014,
1379
+ MNRAS, 444, 1705
1380
+ Gordon K. D., Engelbracht C. W., Rieke G. H., Misselt K. A., Smith J. D. T.,
1381
+ Kennicutt Robert C. J., 2008, ApJ, 682, 336
1382
+ Guzman-Ramirez L., Zijlstra A. A., Níchuimín R., Gesicki K., Lagadec E.,
1383
+ Millar T. J., Woods P. M., 2011, MNRAS, 414, 1667
1384
+ Hony S., Van Kerckhoven C., Peeters E., Tielens A. G. G. M., Hudgins
1385
+ D. M., Allamandola L. J., 2001, A&A, 370, 1030
1386
+ Hora J. L., Latter W. B., Deutsch L. K., 1999, ApJS, 124, 195
1387
+ Jensen P. A., Shannon M. J., Peeters E., Sloan G. C., Stock D. J., 2022,
1388
+ A&A, 665, A153
1389
+ Jones O. C., et al., 2017, MNRAS, 470, 3250
1390
+ Karakas A., Lattanzio J. C., 2007, Publ. Astron. Soc. Australia, 24, 103
1391
+ Karakas A. I., Lattanzio J. C., 2014, Publ. Astron. Soc. Australia, 31, e030
1392
+ Karakas A. I., Lattanzio J. C., Pols O. R., 2002, Publ. Astron. Soc. Australia,
1393
+ 19, 515
1394
+ Kastner J. H., Weintraub D. A., Gatley I., Merrill K. M., Probst R. G., 1996,
1395
+ ApJ, 462, 777
1396
+ Kendrew S., et al., 2015, Publications of the Astronomical Society of the
1397
+ Pacific, 127, 623
1398
+ Kester D. J. M., Beintema D. A., Lutz D., 2003, in Metcalfe L., Salama A.,
1399
+ Peschke S. B., Kessler M. F., eds, ESA Special Publication Vol. 481,
1400
+ The Calibration Legacy of the ISO Mission. p. 375
1401
+ Kwitter K. B., Henry R. B. C., 2022, PASP, 134, 022001
1402
+ Kwok S., 2000, The Origin and Evolution of Planetary Nebulae
1403
+ Kwok S., Volk K. M., Hrivnak B. J., 1989, ApJ, 345, L51
1404
+ Labiano
1405
+ A.,
1406
+ et
1407
+ al.,
1408
+ 2016,
1409
+ in
1410
+ Observatory
1411
+ Operations:
1412
+ Strategies,
1413
+ Processes, and Systems VI. p. 99102W (arXiv:1608.05312),
1414
+ doi:10.1117/12.2232554
1415
+ Labiano A., et al., 2021, A&A, 656, A57
1416
+ Lagadec E., et al., 2007, MNRAS, 376, 1270
1417
+ Law D. R., et al., 2021, AJ, 161, 52
1418
+ Leisenring J. M., Kemper F., Sloan G. C., 2008, ApJ, 681, 1557
1419
+ Madau P., Ferguson H. C., Dickinson M. E., Giavalisco M., Steidel C. C.,
1420
+ Fruchter A., 1996, MNRAS, 283, 1388
1421
+ Margon B., Massey P., Neugent K. F., Morrell N., 2020, ApJ, 898, 85
1422
+ Markwardt C. B., 2009, in Bohlender D. A., Durand D., Dowler P.,
1423
+ eds, Astronomical Society of the Pacific Conference Series Vol. 411,
1424
+ Astronomical Data Analysis Software and Systems XVIII. p. 251
1425
+ (arXiv:0902.2850)
1426
+ Matsuura M., et al., 2007a, MNRAS, 382, 1889
1427
+ Matsuura M., et al., 2007b, MNRAS, 382, 1447
1428
+ Matsuura M., et al., 2014, MNRAS, 439, 1472
1429
+ Peeters E., Hony S., Van Kerckhoven C., Tielens A. G. G. M., Allamandola
1430
+ L. J., Hudgins D. M., Bauschlicher C. W., 2002, A&A, 390, 1089
1431
+ MNRAS 000, 1–11 (2022)
1432
+
1433
+ JWST MRS observations of SMP LMC 058
1434
+ 11
1435
+ Perley R. A., Butler B. J., 2013, ApJS, 204, 19
1436
+ Pietrzyński G., et al., 2013, Nature, 495, 76
1437
+ Pino T., et al., 2008, A&A, 490, 665
1438
+ Reid W. A., 2014, MNRAS, 438, 2642
1439
+ Reid W. A., Parker Q. A., 2006, MNRAS, 373, 521
1440
+ Reid W. A., Parker Q. A., 2010, MNRAS, 405, 1349
1441
+ Rigby J., et al., 2022, arXiv e-prints, p. arXiv:2207.05632
1442
+ Ruffle P. M. E., et al., 2015, MNRAS, 451, 3504
1443
+ Sanduleak N., MacConnell D. J., Philip A. G. D., 1978, PASP, 90, 621
1444
+ Sloan G. C., et al., 2014, ApJ, 791, 28
1445
+ Stanghellini L., García-Lario P., García-Hernández D. A., Perea-Calderón
1446
+ J. V., Davies J. E., Manchado A., Villaver E., Shaw R. A., 2007, ApJ,
1447
+ 671, 1669
1448
+ Swinyard B. M., et al., 1996, A&A, 315, L43
1449
+ Volk K., Sloan G. C., Kraemer K. E., 2020, Ap&SS, 365, 88
1450
+ Waters L. B. F. M., et al., 1996, A&A, 315, L361
1451
+ Waters L. B. F. M., et al., 1998, Nature, 391, 868
1452
+ Wells M., et al., 2015, Publications of the Astronomical Society of the
1453
+ Pacific, 127, 646–664
1454
+ Westerlund B. E., 1997, Cambridge Astrophysics Series, 29
1455
+ Woods P. M., et al., 2011, MNRAS, 411, 1597
1456
+ Woods P. M., Walsh C., Cordiner M. A., Kemper F., 2012, MNRAS, 426,
1457
+ 2689
1458
+ Wright E. L., et al., 2010, AJ, 140, 1868
1459
+ Zijlstra A. A., et al., 2006, MNRAS, 370, 1961
1460
+ Ziurys L. M., 2006, Proceedings of the National Academy of Science, 103,
1461
+ 12274
1462
+ van Diedenhoven B., Peeters E., Van Kerckhoven C., Hony S., Hudgins
1463
+ D. M., Allamandola L. J., Tielens A. G. G. M., 2004, ApJ, 611, 928
1464
+ This paper has been typeset from a TEX/LATEX file prepared by the author.
1465
+ MNRAS 000, 1–11 (2022)
1466
+
0NFQT4oBgHgl3EQfDTUM/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
1NAzT4oBgHgl3EQfDfoM/vector_store/index.faiss ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69d3ff663888866b5b518e6f3246b0725f8e040048fe33bf2cf607e9e5b618c5
3
+ size 5832749
1dE0T4oBgHgl3EQfdgBe/content/tmp_files/2301.02377v1.pdf.txt ADDED
@@ -0,0 +1,1126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Communications in Mathematics n (2023), no. m, 00–12
2
+ DOI: https://doi.org/10.46298/cm.ABCD
3
+ ©2023 G´abor Rom´an
4
+ This is an open access article licensed under the CC BY-SA 4.0
5
+ 1
6
+ On square-free numbers generated from given sets of primes
7
+ II
8
+ G´abor Rom´an
9
+ Abstract. We progress with the investigation started in article [7], namely the anal-
10
+ ysis of the asymptotic behaviour of QP(x) for different sets P, where QP(x) is the
11
+ element count of the set containing those positive square-free integers, which are
12
+ smaller than-, or equal to x, and which are only divisible by the elements of P. We
13
+ study how QP(x) behaves when we require that χ(p) = 1 must hold for every p ∈ P,
14
+ where χ is a Dirichlet character.
15
+ 1
16
+ Introduction
17
+ Let’s take a set of prime numbers P, and denote with QP(x) the element count of the
18
+ set of all those positive square-free integers, which are smaller than-, or equal to x; and
19
+ which are only divisible by the elements of P.
20
+ In article [7] we examined how QP(x) behaves asymptotically based on the structure
21
+ of P in two scenarios. During the first scenario, P contained those primes which are not
22
+ greater than an x dependent bound λ(x), see [7, Prop. 1]. In the second scenario, P
23
+ contained those primes which are not greater than an x dependent bound λ(x), and which
24
+ fall into certain congruence classes modulo q, see [7, Prop. 2].
25
+ In this article, we go further, and render the structure of P more complex. We would
26
+ want to restrict ourselves to primes p, for which a certain integer a is quadratic residue
27
+ modulo p, but to generalise, we are going to require that χ(p) = 1 holds, where χ is a
28
+ Dirichlet character. This covers our goal, as the only real valued primitive non-principal
29
+ χ(n) are given for positive n by the Kronecker symbol (D|n), where D is a fundamental
30
+ discriminant, see [9].
31
+ MSC 2020: 11M06, 11M20, 11N36, 11N37, 11N69
32
+ Keywords: Square-free numbers, Combinatorial sieve, Dirichlet character, Square-free numbers in
33
+ arithmetic progressions, L-functions, Euler product
34
+ Affiliation:
35
+ G´abor Rom´an – E¨otv¨os Lor´and University, Budapest, Hungary
36
+ E-mail: [email protected]
37
+ arXiv:2301.02377v1 [math.NT] 6 Jan 2023
38
+
39
+ =P sciences2
40
+ G´abor Rom´an
41
+ The results will be of course similar to the results in article [7], heavily depending on
42
+ the conductor q(χ) of the Dirichlet character χ in context.
43
+ Proposition 1.1. Let χ be a real valued non-principal Dirichlet character. Furthermore, let
44
+ λ : R → [1, +∞) be a monotone increasing function which is in o(x1/2), and let P contain
45
+ all the primes p which are not greater than λ(x), and for which χ(p) = 1.
46
+ Then for every ε > 0, there exit real constants a1 and a2 such that
47
+ ea1
48
+ ln q(χ)
49
+ ln λ(x) x
50
+ ln x
51
+
52
+ ln λ(x)
53
+
54
+ q(χ)ε ≪ QP(x) ≪ ea2
55
+ ln q(χ)
56
+ ln λ(x) x
57
+ ln x
58
+
59
+ ln q(χ)
60
+
61
+ ln λ(x)
62
+ (1)
63
+ as x → +∞. In addition, if χ is primitive, q(χ) is big enough, and L(s, χ) has no real
64
+ zero in the interval (1 − cχ, 1), then there exists a real constant a3 such that
65
+ ea3
66
+ ln q(χ)
67
+ ln λ(x) x
68
+ ln x
69
+
70
+ ln λ(x)
71
+
72
+ ln q(χ)
73
+ ≪ QP(x)
74
+ (2)
75
+ where cχ ≤ 1 is a χ dependant positive constant.
76
+ We can see that for primitive χ, when q(χ) is big enough, and L(s, χ) doesn’t have a
77
+ real zero close to 1, then we can cancel the ln q(χ) terms in the lower bounds by choosing
78
+ a λ(x) which contains q(χ)ε, for ε > 0.
79
+ Assuming the Riemann hypothesis for L(s, χ) we can drop the primitiveness, and the
80
+ bounds are perturbed only by expressions containing ln ln q(χ) instead of ln q(χ).
81
+ Proposition 1.2. Let χ be a real valued non-principal Dirichlet character. Then choose a
82
+ function λ, and with it define the set P as in the case of proposition 1.1.
83
+ If the Riemann hypothesis holds for L(s, χ), then there exist real constants a4 and a5
84
+ such that
85
+ ea4
86
+ ln ln q(χ)
87
+ ln λ(x)
88
+ x
89
+ ln x
90
+
91
+ ln λ(x)
92
+
93
+ ln ln q(χ)
94
+ ≪ QP(x) ≪ ea5
95
+ ln ln q(χ)
96
+ ln λ(x)
97
+ x
98
+ ln x
99
+
100
+ ln ln q(χ)
101
+
102
+ ln λ(x)
103
+ (3)
104
+ as x → +∞ and q(χ) → +∞.
105
+ A natural extension of these results would be to only allow P to contain primes which
106
+ are congruent to some mi modulo q, where q > 0 is an integer, and the m1, . . . , mk naturals
107
+ are pairwise distinct relative primes to q. We are going to restrict ourselves to the case
108
+ when q = 4; and either m = 1, or m = 3. Concerning the technicalities of this restriction
109
+ see section 3.
110
+ Proposition 1.3. Let χ be a real valued primitive non-principal Dirichlet character; and
111
+ either let m = 1, or m = 3. Furthermore, let λ : R → [1, +∞) be a monotone increasing
112
+ function which is in o(x1/2), and let P contain all those primes p which are not greater
113
+ than λ(x), for which χ(p) = 1, and for which p ≡ m (mod 4) holds.
114
+
115
+ On square-free numbers generated from given sets of primes II
116
+ 3
117
+ When m = 1, then for every ε > 0, there exit real constants b1 and b2 such that
118
+ eb1
119
+ ln q(χ)
120
+ ln λ(x) x
121
+ ln x
122
+ 4�
123
+ ln λ(x)
124
+
125
+ q(χ)ε ≪ QP(x) ≪ eb2
126
+ ln q(χ)
127
+ ln λ(x) x
128
+ ln x
129
+
130
+ ln q(χ)
131
+ 4�
132
+ ln λ(x)
133
+ (4)
134
+ as x → +∞. In addition, if q(χ) is big enough, and L(s, χ) has no real zero in the interval
135
+ (1 − cχ, 1), then there exists a real constant b3 such that
136
+ eb3
137
+ ln q(χ)
138
+ ln λ(x) x
139
+ ln x
140
+ 4�
141
+ ln λ(x)
142
+
143
+ ln q(χ)
144
+ ≪ QP(x)
145
+ (5)
146
+ where cχ ≤ 1 is a χ dependant positive constant.
147
+ When m = 3, then for every ε > 0, there exit real constants b4 and b5 such that
148
+ eb4
149
+ ln q(χ)
150
+ ln λ(x) x
151
+ ln x
152
+ 4�
153
+ ln λ(x)
154
+ 4�
155
+ q(χ)ε ln q(χ)
156
+ ≪ QP(x) ≪ eb5
157
+ ln q(χ)
158
+ ln λ(x) x
159
+ ln x
160
+ 4�
161
+ q(χ)ε ln q(χ)
162
+ 4�
163
+ ln λ(x)
164
+ (6)
165
+ as x → +∞. In addition, if q(χ) is big enough, L(s, χ) has no real zero in the interval
166
+ (1 − cχ, 1), then there exists a real constant b6 such that
167
+ QP(x) ≍ eb6
168
+ ln q(χ)
169
+ ln λ(x) x
170
+ ln x
171
+ 4�
172
+ ln λ(x)
173
+ (7)
174
+ where cχ ≤ 1 is a χ dependant positive constant.
175
+ As in the previous case, we can get much better results assuming that the Riemann
176
+ hypothesis holds for L(s, χ).
177
+ Proposition 1.4. Let χ be a real valued primitive non-principal Dirichlet character; and
178
+ either let m = 1, or m = 3. Then choose a function λ, and with it define the set P as in
179
+ the case of proposition 1.3.
180
+ If the Riemann hypothesis holds for L(s, χ), and when m = 1, then there exist real
181
+ constants b7, and b8 such that
182
+ eb7
183
+ ln ln q(χ)
184
+ ln λ(x)
185
+ x
186
+ ln x
187
+ 4�
188
+ ln λ(x)
189
+
190
+ ln ln q(χ)
191
+ ≪ QP(x) ≪ eb8
192
+ ln ln q(χ)
193
+ ln λ(x)
194
+ x
195
+ ln x
196
+
197
+ ln ln q(χ)
198
+ 4�
199
+ ln λ(x)
200
+ (8)
201
+ as x → +∞ and q(χ) → +∞.
202
+ In the same setting, but with m = 3, there exists a real constant b9 such that
203
+ QP(x) ≍ eb9
204
+ ln ln q(χ)
205
+ ln λ(x)
206
+ x
207
+ ln x
208
+ 4�
209
+ ln λ(x)
210
+ (9)
211
+ as x → +∞ and q(χ) → +∞.
212
+
213
+ 4
214
+ G´abor Rom´an
215
+ 2
216
+ Proofs
217
+ Throughout the proofs, when the index of a summation is p, or the index of a product
218
+ is p, then p takes its values from the set of primes.
219
+ Lemma 2.1. Let χ be a real valued non-principal Dirichlet character.
220
+ Then we have
221
+
222
+ p≤y
223
+ χ(p) ln
224
+
225
+ 1 − 1
226
+ p
227
+
228
+ = − ln L(1, χ) + O
229
+ � 1
230
+ ln y
231
+ L′
232
+ L (1, χ)
233
+
234
+ + O(1)
235
+ as y → +∞.
236
+ Proof. Fix a real valued non-principal Dirichlet character χ. We begin the proof by showing
237
+ that
238
+
239
+ p≤y
240
+ χ(p)
241
+ p
242
+ = ln L(1, χ) + O
243
+ � 1
244
+ ln y
245
+ L′
246
+ L (1, χ)
247
+
248
+ + O(1).
249
+ (10)
250
+ holds as y → +∞.
251
+ • First we show that the equality
252
+
253
+ p
254
+ χ(p)
255
+ p
256
+ = ln L(1, χ) + O(1)
257
+ (11)
258
+ holds. Based on [5, Sec. 5.9], when χ is a non-principal (or in their case non-trivial)
259
+ character, then L(s, χ) is entire. By this, there is no pole at s = 1, so we have that
260
+ the Euler product form
261
+ L(1, χ) =
262
+
263
+ p
264
+
265
+ 1 − χ(p)
266
+ p
267
+ �−1
268
+ (12)
269
+ see [5, Sec. 5.1, Sec. 5.9], or [2, Sec. 11.5], converges. For every non-principal
270
+ Dirichlet character, L(1, χ) ̸= 0, see [2, Thm. 6.20, Lem. 7.7], and as χ is real
271
+ valued, L(1, χ) is a positive real number, so we can take the logarithm of both sides
272
+ of (12) to get
273
+ ln L(1, χ) = −
274
+
275
+ p
276
+ ln
277
+
278
+ 1 − χ(p)
279
+ p
280
+
281
+ =
282
+
283
+ p
284
+ χ(p)
285
+ p
286
+ +
287
+
288
+ p
289
+
290
+
291
+ k=2
292
+ χ(p)k
293
+ kpk
294
+ where we could use the Mercator series, see [1, 4.1.24], as |χ(p)/p| < 1. Based on
295
+ the sum of the geometric series, see [1, 3.6.10], we have
296
+
297
+ p
298
+
299
+
300
+ k=2
301
+ ����
302
+ χ(p)k
303
+ kpk
304
+ ���� ≤
305
+
306
+ p
307
+
308
+
309
+ k=2
310
+ 1
311
+ pk =
312
+
313
+ p
314
+ 1
315
+ p(p − 1) <
316
+
317
+ p
318
+ 1
319
+ p2
320
+ (13)
321
+ which is finite.
322
+
323
+ On square-free numbers generated from given sets of primes II
324
+ 5
325
+ • Next we show that
326
+
327
+ y<p
328
+ χ(p)
329
+ p
330
+
331
+ 1
332
+ ln y
333
+ L′
334
+ L (1, χ) + O(1)
335
+ (14)
336
+ holds as y → +∞. We show this by examining the expression
337
+
338
+ y<n≤z
339
+ χ(n)Λ(n)
340
+ n ln n
341
+ (15)
342
+ in two different ways. Here Λ is the Mangoldt function, see [2, Sec. 2.8].
343
+ – Relying on the definition of the Mangoldt function, there exists a natural m
344
+ such that expression (15) is equal to
345
+
346
+ y<p≤z
347
+ χ(p)
348
+ p
349
+ +
350
+ m
351
+
352
+ k=2
353
+
354
+ y<pk≤z
355
+ χ(pk)
356
+ kpk
357
+ =
358
+
359
+ y<p≤z
360
+ χ(p)
361
+ p
362
+ + O(1)
363
+ (16)
364
+ where we can apply similar reasoning as in expression (13). Note that we can
365
+ keep to positive infinity with z, the implied constant can be bounded.
366
+ – Define
367
+ S(t) :=
368
+
369
+ 1≤n≤t
370
+ χ(n)Λ(n)
371
+ n
372
+ .
373
+ Using Abel’s identity, see [2, Thm. 4.2], we get that expression (15) is equal to
374
+ S(z)
375
+ ln z − S(y)
376
+ ln y +
377
+ � z
378
+ y
379
+ S(t)
380
+ t(ln t)2 dt.
381
+ (17)
382
+ Based on [2, Sec. 7.5] we have
383
+ S(t) =
384
+
385
+ p≤t
386
+ χ(p) ln(p)
387
+ p
388
+ + O(1) ≪ L′
389
+ L (1, χ) + O(1)
390
+ as t → +∞, where the bound is due to [2, Lem. 7.5, Lem. 7.6]. Because of this,
391
+ there exists a real threshold τ; furthermore there exist real constants η1 and η2
392
+ such that for every t ≥ τ we have
393
+ |S(t)| ≤ η1
394
+ L′
395
+ L (1, χ) + η2.
396
+ Assuming that y ≥ τ holds, we can bound the integral in expression (17) by
397
+ η1
398
+ �L′
399
+ L (1, χ) + η2
400
+ � � z
401
+ y
402
+ 1
403
+ t(ln t)2 dt = η1
404
+ �L′
405
+ L (1, χ) + η2
406
+ �� 1
407
+ ln t
408
+ �z
409
+ y
410
+ (18)
411
+
412
+ 6
413
+ G´abor Rom´an
414
+ furthermore there exist real constants η3 and η4 such that the remaining terms
415
+ in expression (17) can be bounded by
416
+ η3
417
+ ln z
418
+ L′
419
+ L (1, χ) + η4
420
+ ln y
421
+ L′
422
+ L (1, χ) + O(1).
423
+ (19)
424
+ Keeping to positive infinity with z in expression (18), and expression (19), we
425
+ get that expression (17) can be bounded by some real constant times
426
+ 1
427
+ ln y
428
+ L′
429
+ L (1, χ) + O(1).
430
+ (20)
431
+ Using the right hand side of equality (16) and expression (20) we get expression (14).
432
+ Combining (11) and (14) we get equality (10). Because 0 < 1/p ≤ 1/2 for every prime p,
433
+ we can use the Mercator series to write the sum in lemma 2.1 as
434
+
435
+
436
+ p≤y
437
+ χ(p)
438
+
439
+
440
+ k=1
441
+ 1
442
+ kpk = −
443
+
444
+ p≤y
445
+ χ(p)
446
+ p
447
+
448
+
449
+ p≤y
450
+
451
+
452
+ k=2
453
+ χ(p)
454
+ kpk .
455
+ We can use equality (10) on the first sum; and relying on expression (13) the value of the
456
+ double sum on the right hand side is in O(1).
457
+ To prove proposition 1.1 and proposition 1.2 we introduce the following function.
458
+ α(y) :=
459
+
460
+ p≤y
461
+ χ(p)=1
462
+
463
+ 1 −
464
+ 1
465
+ p + 1
466
+
467
+ Lemma 2.2. Let χ be a real valued non-principal Dirichlet character.
468
+ Then we have
469
+ α(y) ≍
470
+ 1
471
+
472
+ L(1, χ)
473
+ 1
474
+ √ln yeO
475
+
476
+ 1
477
+ ln y
478
+ L′
479
+ L (1,χ)
480
+
481
+ as y → +∞.
482
+ Proof. Fix a real valued non-principal Dirichlet character χ. Observe that we can rewrite
483
+ α(y) as
484
+
485
+ p≤y
486
+ χ(p)=1
487
+
488
+ 1 −
489
+ 1
490
+ p + 1
491
+
492
+ =
493
+
494
+ p≤y
495
+ χ(p)=1
496
+
497
+ 1 − 1
498
+ p2
499
+ �−1 �
500
+ p≤y
501
+ χ(p)=1
502
+
503
+ 1 − 1
504
+ p
505
+
506
+ where the first product on the right hand side can be bounded by a small positive constant.
507
+ Taking the logarithm of the second product on the right hand side we get
508
+
509
+ p≤y
510
+ χ(p)=1
511
+ ln
512
+
513
+ 1 − 1
514
+ p
515
+
516
+ = 1
517
+ 2
518
+
519
+ p≤y
520
+ (1 + χ(p)) ln
521
+
522
+ 1 − 1
523
+ p
524
+
525
+
526
+ On square-free numbers generated from given sets of primes II
527
+ 7
528
+ where we can split the finite sum on the right hand side, and use lemma 2.1 to get
529
+ 1
530
+ 2
531
+
532
+ p≤y
533
+ ln
534
+
535
+ 1 − 1
536
+ p
537
+
538
+ − ln
539
+
540
+ L(1, χ) + O
541
+ � 1
542
+ ln y
543
+ L′
544
+ L (1, χ)
545
+
546
+ + O(1)
547
+ as y → +∞. Via exponentiation, we get
548
+ 1
549
+
550
+ L(1, χ)
551
+ 1
552
+ √ln yeO
553
+
554
+ 1
555
+ ln y
556
+ L′
557
+ L (1,χ)
558
+
559
+ +O(1)
560
+ where we have used [8, Thm. 7, Col.] stating that
561
+
562
+ p≤y
563
+
564
+ 1 − 1
565
+ p
566
+
567
+
568
+ 1
569
+ ln y
570
+ for every y > 1.
571
+ Now we prove proposition 1.1.
572
+ Proof. Fix a real valued non-principal Dirichlet character χ, and select a function λ sat-
573
+ isfying the requirements of proposition 1.1. According article [7], we have to bound the
574
+ product
575
+
576
+ p≤x1/2
577
+
578
+ 1 −
579
+ 1
580
+ p + 1
581
+
582
+ α(λ(x))−1 ≍
583
+ 1
584
+ ln x
585
+
586
+ L(1, χ)
587
+
588
+ ln λ(x)eO
589
+
590
+ 1
591
+ ln λ(x)
592
+ L′
593
+ L (1,χ)
594
+
595
+ (21)
596
+ as x → +∞, where we have used [7, Lem. 2], and lemma 2.2. When χ is a non-principal
597
+ character, then based on article [3] we have
598
+ cεq(χ)−ε < L(1, χ) < ln q(χ)
599
+ (22)
600
+ where ε is any positive number and cε is a positive number depending on ε. Also, based
601
+ on [5, Prop. 5.7] we have
602
+ L′
603
+ L (1, χ) ≪ ln q(χ)
604
+ (23)
605
+ where the implied constant being absolute. Using these bounds in expression (21) and the
606
+ method from article [7] we can get the bounds in expression (1).
607
+ Assuming that χ is primitive, q(χ) is big enough, and that there exists a positive
608
+ constant cχ ≤ 1 such that L(s, χ) has no real zero in the interval (1 − cχ, 1), we have
609
+ 1
610
+ ln q(χ) ≪ L(1, χ)
611
+ (24)
612
+ where the implied constant being positive and absolute, see article [4]. Using this bound
613
+ in expression (21) and the method from article [7] we get the bound in expression (2).
614
+
615
+ 8
616
+ G´abor Rom´an
617
+ The proof of proposition 1.2 is the following.
618
+ Proof. Fix a real valued non-principal Dirichlet character χ, and select a function λ sat-
619
+ isfying the requirements of proposition 1.2. We are going to bound expression (21), but
620
+ with bounds based on the Riemann hypothesis. If χ is a real valued non-principal Dirichlet
621
+ character, and we assume that the Riemann hypothesis holds for L(s, χ), then based on
622
+ [6, Thm. 1] we have
623
+ 1 + o(1)
624
+ ε1 ln ln q(χ) < L(1, χ) < (1 + o(1))ε1 ln ln q(χ)
625
+ (25)
626
+ as q(χ) → ∞, where ε1 and ε2 are real constants.
627
+ (As a side note, infinitely many
628
+ real primitive characters χ satisfy these inequalities without assuming that the Riemann
629
+ hypothesis holds for L(s, χ), see article [3].) In the same setting, based on [5, Thm. 5.17]
630
+ we have
631
+ L′
632
+ L (1, χ) ≪ ln ln q(χ)
633
+ (26)
634
+ where the implied constant being absolute. Using these bounds in expression (21) and the
635
+ method from article [7] we get the bounds in expression (3).
636
+ To prove proposition 1.3 and proposition 1.4 we introduce the following function.
637
+ βm(y) :=
638
+
639
+ p≤y
640
+ p≡m(4)
641
+ χ(p)=1
642
+
643
+ 1 −
644
+ 1
645
+ p + 1
646
+
647
+ Lemma 2.3. Let χ be a real valued non-principal Dirichlet character.
648
+ Then we have
649
+ β1(y) ≍
650
+ 1
651
+ 4�
652
+ L(1, χ)L(1, χχ4,3)
653
+ 1
654
+ 4√ln yeO
655
+
656
+ 1
657
+ ln y
658
+ L′
659
+ L (1,χ)
660
+
661
+ +O
662
+
663
+ 1
664
+ ln y
665
+ L′
666
+ L (1,χχ4,3)
667
+
668
+ (27)
669
+ and
670
+ β3(y) ≍
671
+ 4
672
+
673
+ L(1, χχ4,3)
674
+ L(1, χ)
675
+ 1
676
+ 4√ln yeO
677
+
678
+ 1
679
+ ln y
680
+ L′
681
+ L (1,χ)
682
+
683
+ +O
684
+
685
+ 1
686
+ ln y
687
+ L′
688
+ L (1,χχ4,3)
689
+
690
+ (28)
691
+ as y → +∞, where χ4,3 is the non-principal Dirichlet character modulo 4.
692
+ Proof. Fix a real valued non-principal Dirichlet character χ, and let either m = 1, or
693
+ m = 3. We can rewrite βm(y) as
694
+
695
+ p≤y
696
+ p≡m(4)
697
+ χ(p)=1
698
+
699
+ 1 −
700
+ 1
701
+ p + 1
702
+
703
+ =
704
+
705
+ p≤y
706
+ p≡m(4)
707
+ χ(p)=1
708
+
709
+ 1 − 1
710
+ p2
711
+ �−1 �
712
+ p≤y
713
+ p≡m(4)
714
+ χ(p)=1
715
+
716
+ 1 − 1
717
+ p
718
+
719
+
720
+ On square-free numbers generated from given sets of primes II
721
+ 9
722
+ where the first product on the right hand side can be bounded by a small positive constant.
723
+ Taking the logarithm of the second product on the right hand side we get
724
+
725
+ p≤y
726
+ p≡m(4)
727
+ χ(p)=1
728
+ ln
729
+
730
+ 1 − 1
731
+ p
732
+
733
+ = 1
734
+ 2
735
+
736
+ p≤y
737
+ p≡m(4)
738
+ (1 + χ(p)) ln
739
+
740
+ 1 − 1
741
+ p
742
+
743
+ where we can split the finite sum on the right hand side as
744
+ 1
745
+ 2
746
+
747
+ p≤y
748
+ p≡m(4)
749
+ ln
750
+
751
+ 1 − 1
752
+ p
753
+
754
+ + 1
755
+ 2
756
+
757
+ p≤y
758
+ p≡m(4)
759
+ χ(p) ln
760
+
761
+ 1 − 1
762
+ p
763
+
764
+ .
765
+ (29)
766
+ Based on [2, Thm. 6.16] we can write the second sum in expression (29) as
767
+ 1
768
+ 2
769
+
770
+ p≤y
771
+ χ(p) ln
772
+
773
+ 1 − 1
774
+ p
775
+
776
+ 1
777
+ ϕ(4)
778
+
779
+ χ4
780
+ χ4(p)χ4(m)
781
+ (30)
782
+ where the internal sum iterates through the ϕ(4) Dirichlet characters modulo 4. There
783
+ are two Dirichlet characters modulo 4; we are going to denote them as χ4,1 (the principal
784
+ character), and as χ4,3 (the non-principal character). Splitting the internal sum, we get
785
+ χ4,1(m)
786
+ 4
787
+
788
+ p≤y
789
+ χ(p)χ4,1(p) ln
790
+
791
+ 1 − 1
792
+ p
793
+
794
+ + χ4,3(m)
795
+ 4
796
+
797
+ p≤y
798
+ χ(p)χ4,3(p) ln
799
+
800
+ 1 − 1
801
+ p
802
+
803
+ .
804
+ (31)
805
+ Concerning the sum on the left hand side of expression (31), as χ4,1(m) = 1; and as
806
+ χ4,1(p) = 1 when (p, 4) = 1, otherwise χ4,1(p) = 0, we have
807
+ 1
808
+ 4
809
+
810
+ p≤y
811
+ (p,4)=1
812
+ χ(p) ln
813
+
814
+ 1 − 1
815
+ p
816
+
817
+ = 1
818
+ 4
819
+
820
+ p≤y
821
+ χ(p) ln
822
+
823
+ 1 − 1
824
+ p
825
+
826
+ + O(1)
827
+ where we can use lemma 2.1 to get
828
+ −1
829
+ 4 ln L(1, χ) + O
830
+ � 1
831
+ ln y
832
+ L′
833
+ L (1, χ)
834
+
835
+ + O(1).
836
+ Concerning the sum on the right hand side of expression (31), χχ4,3 is a real valued non-
837
+ principal character, so we can use lemma 2.1 again to get
838
+ −χ4,3(m)
839
+ 4
840
+ ln L(1, χχ4,3) + O
841
+ � 1
842
+ ln y
843
+ L′
844
+ L (1, χχ4,3)
845
+
846
+ + O(1).
847
+ Substituting these result in expression (29), and exponentiating, in the case when m = 1,
848
+ we get expression (27) as χ4,3(1) = 1, and because
849
+
850
+ p≤y
851
+ p≡m(4)
852
+
853
+ 1 − 1
854
+ p
855
+
856
+
857
+ 1
858
+ √ln y
859
+
860
+ 10
861
+ G´abor Rom´an
862
+ based on the article of Williams [11]. Similarly in the case when m = 3 we get expression
863
+ (28) as χ4,3(3) = −1.
864
+ Now we proof proposition 1.3.
865
+ Proof. Fix a real valued primitive non-principal Dirichlet character χ; and either let m = 1,
866
+ or m = 3. Furthermore select a function λ satisfying the requirements of proposition 1.3.
867
+ Based on article [7], we have to bound the product
868
+
869
+ p≤x1/2
870
+
871
+ 1 −
872
+ 1
873
+ p + 1
874
+
875
+ βm(λ(x))−1
876
+ (32)
877
+ when m = 1, and separately when m = 3.
878
+ When m = 1, then expression (32) is asymptotic to
879
+ 1
880
+ ln x
881
+ 4�
882
+ L(1, χ)L(1, χχ4,3)
883
+ 4�
884
+ ln λ(x)eO
885
+
886
+ 1
887
+ ln y
888
+ L′
889
+ L (1,χ)
890
+
891
+ +O
892
+
893
+ 1
894
+ ln y
895
+ L′
896
+ L (1,χχ4,3)
897
+
898
+ (33)
899
+ as x → +∞, where we have used [7, Lem. 2] and lemma 2.3. Using the bounds from
900
+ expression (22) and expression (23) we can bound expression (33) from below as
901
+ 1
902
+ ln x
903
+ 4�
904
+ ln λ(x)
905
+ 4�
906
+ q(χ)εq(χχ4,3)εeO
907
+
908
+ 1
909
+ ln y (ln q(χ)+ln q(χχ4,3))
910
+
911
+ and as
912
+ 1
913
+ ln x
914
+ 4�
915
+ ln q(χ)
916
+ 4�
917
+ ln q(χχ4,3)
918
+ 4�
919
+ ln λ(x)eO
920
+
921
+ 1
922
+ ln y (ln q(χ)+ln q(χχ4,3))
923
+
924
+ from above. As χ and χ4,3 are both primitive, their product χχ4,3 is primitive too, see [10,
925
+ Ch. 3]. But then q(χχ4,3) ∈ O(q(χ)), see [5, Sec. 3.3]. Based on this and on the method
926
+ in article [7] we get the bounds in expression (4).
927
+ Assuming that q(χ) is big enough, and that there exists a positive constant cχ ≤ 1 such
928
+ that L(s, χ) has no real zero in the interval (1−cχ, 1), we can use bound (24) in expression
929
+ (33) to get
930
+ 1
931
+ ln x
932
+ 4�
933
+ ln λ(x)
934
+ 4�
935
+ ln q(χ) 4�
936
+ ln q(χχ4,3)
937
+ eO
938
+
939
+ 1
940
+ ln y (ln q(χ)+ln q(χχ4,3))
941
+
942
+ from where we can get bound (5) based on the previous train of thoughts.
943
+ When m = 3, then expression (32) is asymptotic to
944
+ 1
945
+ ln x
946
+ 4
947
+
948
+ L(1, χ)
949
+ L(1, χχ4,3)
950
+ 4�
951
+ ln λ(x)eO
952
+
953
+ 1
954
+ ln y
955
+ L′
956
+ L (1,χ)
957
+
958
+ +O
959
+
960
+ 1
961
+ ln y
962
+ L′
963
+ L (1,χχ4,3)
964
+
965
+ (34)
966
+ as x → +∞, where we have used [7, Lem. 2] and lemma 2.3 again. As in the previous
967
+ case, we can bound expression (34) from below as
968
+ 1
969
+ ln x
970
+ 4
971
+
972
+ q(χ)−ε
973
+ ln q(χχ4,3)
974
+ 4�
975
+ ln λ(x)eO
976
+
977
+ 1
978
+ ln y (ln q(χ)+ln q(χχ4,3))
979
+
980
+
981
+ On square-free numbers generated from given sets of primes II
982
+ 11
983
+ and from above as
984
+ 1
985
+ ln x
986
+ 4
987
+
988
+ ln q(χ)
989
+ q(χχ4,3)−ε
990
+ 4�
991
+ ln λ(x)eO
992
+
993
+ 1
994
+ ln y (ln q(χ)+ln q(χχ4,3))
995
+
996
+ from where we get the bounds in expression (6).
997
+ Assuming that q(χ) is big enough, and that there exists a positive constant cχ ≤ 1 such
998
+ that L(s, χ) has no real zero in the interval (1−cχ, 1), we can use bound (24) in expression
999
+ (34).
1000
+ The logarithmic contribution of the terms L(1, χ) and L(1, χχ4,3) “cancel” each
1001
+ other, and we get asymptotic (7).
1002
+ And finally, the proof of proposition 1.4 is the following.
1003
+ Proof. Fix a real valued primitive non-principal Dirichlet character χ; and either let m = 1,
1004
+ or m = 3. We use the same method as in the proof of proposition 1.3, but this time we
1005
+ assume that the Riemann hypothesis holds for L(s, χ).
1006
+ When m = 1, then we can use the bounds from expression (25) and expression (26) to
1007
+ bound expression (33) from below as
1008
+ 1
1009
+ ln x
1010
+ 4�
1011
+ ln λ(x)
1012
+ 4�
1013
+ ln ln q(χ) 4�
1014
+ ln ln q(χχ4,3)
1015
+ eO
1016
+
1017
+ 1
1018
+ ln y (ln ln q(χ)+ln ln q(χχ4,3))
1019
+
1020
+ and from above as
1021
+ 1
1022
+ ln x
1023
+ 4�
1024
+ ln ln q(χ)
1025
+ 4�
1026
+ ln ln q(χχ4,3)
1027
+ 4�
1028
+ ln λ(x)eO
1029
+
1030
+ 1
1031
+ ln y (ln ln q(χ)+ln ln q(χχ4,3))
1032
+
1033
+ as x → +∞ and as q(χ) → +∞. Using the same train of thought as in the proof of
1034
+ proposition 1.3, we get the bounds in expression (8).
1035
+ When m = 3, then using the above applied bounds (25) and (26) we get the asymptotic
1036
+ (9) from asymptotic (34) via “cancellation” again.
1037
+ 3
1038
+ Remarks
1039
+ As we have already mentioned in section 1, before proposition 1.3, a natural extension
1040
+ of proposition 1.1 and proposition 1.2 would be to only allow P to contain primes which
1041
+ are congruent to some mi modulo q, where q > 0 is an integer, and the m1, . . . , mk naturals
1042
+ are pairwise distinct relative primes to q. In the case of modulo 4, the results were already
1043
+ different for distinct m, so we can expect a similar outcome for larger moduli. However a
1044
+ general strategy for the generalisation could go along the following train of thoughts. We
1045
+ would have to supply an asymptotic for the product
1046
+
1047
+ p≤y
1048
+ p≡m(q)
1049
+ χ(p)=1
1050
+
1051
+ 1 −
1052
+ 1
1053
+ p + 1
1054
+
1055
+
1056
+ 12
1057
+ G´abor Rom´an
1058
+ which could be done by the refinement of lemma 2.3, and its proof. If we follow this path,
1059
+ then expression (30) will turn into
1060
+ 1
1061
+ 2
1062
+
1063
+ p≤y
1064
+ χ(p) ln
1065
+
1066
+ 1 − 1
1067
+ p
1068
+ � 1
1069
+ ϕ(q)
1070
+
1071
+ χq
1072
+ χq(p)χq(m)
1073
+ where we can separate the principal character from the non-principal ones as
1074
+ χq,1(m)
1075
+ ϕ(q)
1076
+
1077
+ p≤y
1078
+ χ(p)χq,1(p) ln
1079
+
1080
+ 1 − 1
1081
+ p
1082
+
1083
+ and what remains is
1084
+ 1
1085
+ ϕ(q)
1086
+
1087
+ χq̸=χq,1
1088
+ χq(m)
1089
+
1090
+ p≤y
1091
+ χ(p)χq(p) ln
1092
+
1093
+ 1 − 1
1094
+ p
1095
+
1096
+ .
1097
+ Due to the fact that χq,1 is the principal character, the first sum can be handled with our
1098
+ already presented techniques. The double sum is more problematic. On the one hand, for
1099
+ the internal sum we would have to refine lemma 2.1 and its proof. We would have to make
1100
+ sure that when χ is complex valued, then we can take the logarithm of L(1, χ) and its
1101
+ product form; furthermore that the values of these two logarithms match. On the other
1102
+ hand, we would have to obtain a good estimation for the external sum.
1103
+ References
1104
+ [1] Abramowitz M. and Stegun I. A.: Handbook of Mathematical Functions with Formulas, Graphs,
1105
+ and Mathematical Tables. Dover publications (1972).
1106
+ [2] Apostol T. M.: Introduction to Analytic Number Theory. Springer-Verlag (1976).
1107
+ [3] Bateman P. T. and Chowla S. and Erd˝os P.: Remarks on the size of L(1, χ). Publ. Math.
1108
+ Debrecen 1 (2-4) (1950) 165–182.
1109
+ [4] Hoffstein J.: On the Siegel–Tatuzawa theorem. Acta. Arith. 38 (2) (1980) 168–174.
1110
+ [5] Iwaniec H. and Kowalski E.: Analytic Number Theory. A.M.S. Colloquium Publications (2004).
1111
+ [6] Littlewood J. E.: On the class-number of the corpus P(
1112
+
1113
+ −k). Proc. London Math. Soc. 27 (1)
1114
+ (1928) 358–372.
1115
+ [7] Rom´an G.: On square-free numbers generated from given sets of primes. Comm. Math. 30 (1)
1116
+ (2022) 229–237.
1117
+ [8] Rosser J. B. and Schoenfeld L.: Approximate formulas for some functions of prime numbers.
1118
+ Illinois Journal of Mathematics 6 (1) (1962) 64–94.
1119
+ [9] Walfisz A.: Zur additiven Zahlentheorie II.. Math. Z. 40 (1) (1936) 592–607.
1120
+ [10] Washington L. C.: Introduction to Cyclotomic Fields. Springer-Verlag (1982).
1121
+ [11] Williams K.S.: Mertens’ Theorem for Arithmetic Progressions. J. Number Theory 6 (5) (1974)
1122
+ 353–359.
1123
+ Received: Received date
1124
+ Accepted for publication: Accepted date
1125
+ Communicated by: Handling Editor
1126
+
1dE0T4oBgHgl3EQfdgBe/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,304 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf,len=303
2
+ page_content='Communications in Mathematics n (2023), no.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
3
+ page_content=' m, 00–12 DOI: https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
4
+ page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
5
+ page_content='46298/cm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
6
+ page_content='ABCD ©2023 G´abor Rom´an This is an open access article licensed under the CC BY-SA 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
7
+ page_content='0 1 On square-free numbers generated from given sets of primes II G´abor Rom´an Abstract.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
8
+ page_content=' We progress with the investigation started in article [7], namely the anal- ysis of the asymptotic behaviour of QP(x) for different sets P, where QP(x) is the element count of the set containing those positive square-free integers, which are smaller than-, or equal to x, and which are only divisible by the elements of P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
9
+ page_content=' We study how QP(x) behaves when we require that χ(p) = 1 must hold for every p ∈ P, where χ is a Dirichlet character.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
10
+ page_content=' 1 Introduction Let’s take a set of prime numbers P, and denote with QP(x) the element count of the set of all those positive square-free integers, which are smaller than-, or equal to x;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
11
+ page_content=' and which are only divisible by the elements of P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
12
+ page_content=' In article [7] we examined how QP(x) behaves asymptotically based on the structure of P in two scenarios.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
13
+ page_content=' During the first scenario, P contained those primes which are not greater than an x dependent bound λ(x), see [7, Prop.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
14
+ page_content=' 1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
15
+ page_content=' In the second scenario, P contained those primes which are not greater than an x dependent bound λ(x), and which fall into certain congruence classes modulo q, see [7, Prop.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
16
+ page_content=' 2].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
17
+ page_content=' In this article, we go further, and render the structure of P more complex.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
18
+ page_content=' We would want to restrict ourselves to primes p, for which a certain integer a is quadratic residue modulo p, but to generalise, we are going to require that χ(p) = 1 holds, where χ is a Dirichlet character.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
19
+ page_content=' This covers our goal, as the only real valued primitive non-principal χ(n) are given for positive n by the Kronecker symbol (D|n), where D is a fundamental discriminant, see [9].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
20
+ page_content=' MSC 2020: 11M06, 11M20, 11N36, 11N37, 11N69 Keywords: Square-free numbers, Combinatorial sieve, Dirichlet character, Square-free numbers in arithmetic progressions, L-functions, Euler product Affiliation: G´abor Rom´an – E¨otv¨os Lor´and University, Budapest, Hungary E-mail: rogpaai@inf.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
21
+ page_content='elte.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
22
+ page_content='hu arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
23
+ page_content='02377v1 [math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
24
+ page_content='NT] 6 Jan 2023 =P sciences2 G´abor Rom´an The results will be of course similar to the results in article [7], heavily depending on the conductor q(χ) of the Dirichlet character χ in context.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
25
+ page_content=' Proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
26
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
27
+ page_content=' Let χ be a real valued non-principal Dirichlet character.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
28
+ page_content=' Furthermore, let λ : R → [1, +∞) be a monotone increasing function which is in o(x1/2), and let P contain all the primes p which are not greater than λ(x), and for which χ(p) = 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
29
+ page_content=' Then for every ε > 0, there exit real constants a1 and a2 such that ea1 ln q(χ) ln λ(x) x ln x � ln λ(x) � q(χ)ε ≪ QP(x) ≪ ea2 ln q(χ) ln λ(x) x ln x � ln q(χ) � ln λ(x) (1) as x → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
30
+ page_content=' In addition, if χ is primitive, q(χ) is big enough, and L(s, χ) has no real zero in the interval (1 − cχ, 1), then there exists a real constant a3 such that ea3 ln q(χ) ln λ(x) x ln x � ln λ(x) � ln q(χ) ≪ QP(x) (2) where cχ ≤ 1 is a χ dependant positive constant.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
31
+ page_content=' We can see that for primitive χ, when q(χ) is big enough, and L(s, χ) doesn’t have a real zero close to 1, then we can cancel the ln q(χ) terms in the lower bounds by choosing a λ(x) which contains q(χ)ε, for ε > 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
32
+ page_content=' Assuming the Riemann hypothesis for L(s, χ) we can drop the primitiveness, and the bounds are perturbed only by expressions containing ln ln q(χ) instead of ln q(χ).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
33
+ page_content=' Proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
34
+ page_content='2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
35
+ page_content=' Let χ be a real valued non-principal Dirichlet character.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
36
+ page_content=' Then choose a function λ, and with it define the set P as in the case of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
37
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
38
+ page_content=' If the Riemann hypothesis holds for L(s, χ), then there exist real constants a4 and a5 such that ea4 ln ln q(χ) ln λ(x) x ln x � ln λ(x) � ln ln q(χ) ≪ QP(x) ≪ ea5 ln ln q(χ) ln λ(x) x ln x � ln ln q(χ) � ln λ(x) (3) as x → +∞ and q(χ) → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
39
+ page_content=' A natural extension of these results would be to only allow P to contain primes which are congruent to some mi modulo q, where q > 0 is an integer, and the m1, .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
40
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
41
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
42
+ page_content=' , mk naturals are pairwise distinct relative primes to q.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
43
+ page_content=' We are going to restrict ourselves to the case when q = 4;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
44
+ page_content=' and either m = 1, or m = 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
45
+ page_content=' Concerning the technicalities of this restriction see section 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
46
+ page_content=' Proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
47
+ page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
48
+ page_content=' Let χ be a real valued primitive non-principal Dirichlet character;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
49
+ page_content=' and either let m = 1, or m = 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
50
+ page_content=' Furthermore, let λ : R → [1, +∞) be a monotone increasing function which is in o(x1/2), and let P contain all those primes p which are not greater than λ(x), for which χ(p) = 1, and for which p ≡ m (mod 4) holds.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
51
+ page_content=' On square-free numbers generated from given sets of primes II 3 When m = 1, then for every ε > 0, there exit real constants b1 and b2 such that eb1 ln q(χ) ln λ(x) x ln x 4� ln λ(x) � q(χ)ε ≪ QP(x) ≪ eb2 ln q(χ) ln λ(x) x ln x � ln q(χ) 4� ln λ(x) (4) as x → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
52
+ page_content=' In addition, if q(χ) is big enough, and L(s, χ) has no real zero in the interval (1 − cχ, 1), then there exists a real constant b3 such that eb3 ln q(χ) ln λ(x) x ln x 4� ln λ(x) � ln q(χ) ≪ QP(x) (5) where cχ ≤ 1 is a χ dependant positive constant.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
53
+ page_content=' When m = 3, then for every ε > 0, there exit real constants b4 and b5 such that eb4 ln q(χ) ln λ(x) x ln x 4� ln λ(x) 4� q(χ)ε ln q(χ) ≪ QP(x) ≪ eb5 ln q(χ) ln λ(x) x ln x 4� q(χ)ε ln q(χ) 4� ln λ(x) (6) as x → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
54
+ page_content=' In addition, if q(χ) is big enough, L(s, χ) has no real zero in the interval (1 − cχ, 1), then there exists a real constant b6 such that QP(x) ≍ eb6 ln q(χ) ln λ(x) x ln x 4� ln λ(x) (7) where cχ ≤ 1 is a χ dependant positive constant.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
55
+ page_content=' As in the previous case, we can get much better results assuming that the Riemann hypothesis holds for L(s, χ).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
56
+ page_content=' Proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
57
+ page_content='4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
58
+ page_content=' Let χ be a real valued primitive non-principal Dirichlet character;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
59
+ page_content=' and either let m = 1, or m = 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
60
+ page_content=' Then choose a function λ, and with it define the set P as in the case of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
61
+ page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
62
+ page_content=' If the Riemann hypothesis holds for L(s, χ), and when m = 1, then there exist real constants b7, and b8 such that eb7 ln ln q(χ) ln λ(x) x ln x 4� ln λ(x) � ln ln q(χ) ≪ QP(x) ≪ eb8 ln ln q(χ) ln λ(x) x ln x � ln ln q(χ) 4� ln λ(x) (8) as x → +∞ and q(χ) → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
63
+ page_content=' In the same setting, but with m = 3, there exists a real constant b9 such that QP(x) ≍ eb9 ln ln q(χ) ln λ(x) x ln x 4� ln λ(x) (9) as x → +∞ and q(χ) → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
64
+ page_content=' 4 G´abor Rom´an 2 Proofs Throughout the proofs, when the index of a summation is p, or the index of a product is p, then p takes its values from the set of primes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
65
+ page_content=' Lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
66
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
67
+ page_content=' Let χ be a real valued non-principal Dirichlet character.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
68
+ page_content=' Then we have � p≤y χ(p) ln � 1 − 1 p � = − ln L(1, χ) + O � 1 ln y L′ L (1, χ) � + O(1) as y → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
69
+ page_content=' Proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
70
+ page_content=' Fix a real valued non-principal Dirichlet character χ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
71
+ page_content=' We begin the proof by showing that � p≤y χ(p) p = ln L(1, χ) + O � 1 ln y L′ L (1, χ) � + O(1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
72
+ page_content=' (10) holds as y → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
73
+ page_content=' First we show that the equality � p χ(p) p = ln L(1, χ) + O(1) (11) holds.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
74
+ page_content=' Based on [5, Sec.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
75
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
76
+ page_content='9], when χ is a non-principal (or in their case non-trivial) character, then L(s, χ) is entire.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
77
+ page_content=' By this, there is no pole at s = 1, so we have that the Euler product form L(1, χ) = � p � 1 − χ(p) p �−1 (12) see [5, Sec.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
78
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
79
+ page_content='1, Sec.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
80
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
81
+ page_content='9], or [2, Sec.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
82
+ page_content=' 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
83
+ page_content='5], converges.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
84
+ page_content=' For every non-principal Dirichlet character, L(1, χ) ̸= 0, see [2, Thm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
85
+ page_content=' 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
86
+ page_content='20, Lem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
87
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
88
+ page_content='7], and as χ is real valued, L(1, χ) is a positive real number, so we can take the logarithm of both sides of (12) to get ln L(1, χ) = − � p ln � 1 − χ(p) p � = � p χ(p) p + � p ∞ � k=2 χ(p)k kpk where we could use the Mercator series, see [1, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
89
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
90
+ page_content='24], as |χ(p)/p| < 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
91
+ page_content=' Based on the sum of the geometric series, see [1, 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
92
+ page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
93
+ page_content='10], we have � p ∞ � k=2 ���� χ(p)k kpk ���� ≤ � p ∞ � k=2 1 pk = � p 1 p(p − 1) < � p 1 p2 (13) which is finite.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
94
+ page_content=' On square-free numbers generated from given sets of primes II 5 Next we show that � y<p χ(p) p ≪ 1 ln y L′ L (1, χ) + O(1) (14) holds as y → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
95
+ page_content=' We show this by examining the expression � y<n≤z χ(n)Λ(n) n ln n (15) in two different ways.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
96
+ page_content=' Here Λ is the Mangoldt function, see [2, Sec.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
97
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
98
+ page_content='8].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
99
+ page_content=' – Relying on the definition of the Mangoldt function, there exists a natural m such that expression (15) is equal to � y<p≤z χ(p) p + m � k=2 � y<pk≤z χ(pk) kpk = � y<p≤z χ(p) p + O(1) (16) where we can apply similar reasoning as in expression (13).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
100
+ page_content=' Note that we can keep to positive infinity with z, the implied constant can be bounded.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
101
+ page_content=' – Define S(t) := � 1≤n≤t χ(n)Λ(n) n .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
102
+ page_content=' Using Abel’s identity, see [2, Thm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
103
+ page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
104
+ page_content='2], we get that expression (15) is equal to S(z) ln z − S(y) ln y + � z y S(t) t(ln t)2 dt.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
105
+ page_content=' (17) Based on [2, Sec.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
106
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
107
+ page_content='5] we have S(t) = � p≤t χ(p) ln(p) p + O(1) ≪ L′ L (1, χ) + O(1) as t → +∞, where the bound is due to [2, Lem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
108
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
109
+ page_content='5, Lem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
110
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
111
+ page_content='6].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
112
+ page_content=' Because of this, there exists a real threshold τ;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
113
+ page_content=' furthermore there exist real constants η1 and η2 such that for every t ≥ τ we have |S(t)| ≤ η1 L′ L (1, χ) + η2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
114
+ page_content=' Assuming that y ≥ τ holds, we can bound the integral in expression (17) by η1 �L′ L (1, χ) + η2 � � z y 1 t(ln t)2 dt = η1 �L′ L (1, χ) + η2 �� 1 ln t �z y (18) 6 G´abor Rom´an furthermore there exist real constants η3 and η4 such that the remaining terms in expression (17) can be bounded by η3 ln z L′ L (1, χ) + η4 ln y L′ L (1, χ) + O(1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
115
+ page_content=' (19) Keeping to positive infinity with z in expression (18), and expression (19), we get that expression (17) can be bounded by some real constant times 1 ln y L′ L (1, χ) + O(1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
116
+ page_content=' (20) Using the right hand side of equality (16) and expression (20) we get expression (14).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
117
+ page_content=' Combining (11) and (14) we get equality (10).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
118
+ page_content=' Because 0 < 1/p ≤ 1/2 for every prime p, we can use the Mercator series to write the sum in lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
119
+ page_content='1 as − � p≤y χ(p) ∞ � k=1 1 kpk = − � p≤y χ(p) p − � p≤y ∞ � k=2 χ(p) kpk .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
120
+ page_content=' We can use equality (10) on the first sum;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
121
+ page_content=' and relying on expression (13) the value of the double sum on the right hand side is in O(1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
122
+ page_content=' To prove proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
123
+ page_content='1 and proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
124
+ page_content='2 we introduce the following function.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
125
+ page_content=' α(y) := � p≤y χ(p)=1 � 1 − 1 p + 1 � Lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
126
+ page_content='2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
127
+ page_content=' Let χ be a real valued non-principal Dirichlet character.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
128
+ page_content=' Then we have α(y) ≍ 1 � L(1, χ) 1 √ln yeO � 1 ln y L′ L (1,χ) � as y → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
129
+ page_content=' Proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
130
+ page_content=' Fix a real valued non-principal Dirichlet character χ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
131
+ page_content=' Observe that we can rewrite α(y) as � p≤y χ(p)=1 � 1 − 1 p + 1 � = � p≤y χ(p)=1 � 1 − 1 p2 �−1 � p≤y χ(p)=1 � 1 − 1 p � where the first product on the right hand side can be bounded by a small positive constant.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
132
+ page_content=' Taking the logarithm of the second product on the right hand side we get � p≤y χ(p)=1 ln � 1 − 1 p � = 1 2 � p≤y (1 + χ(p)) ln � 1 − 1 p � On square-free numbers generated from given sets of primes II 7 where we can split the finite sum on the right hand side, and use lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
133
+ page_content='1 to get 1 2 � p≤y ln � 1 − 1 p � − ln � L(1, χ) + O � 1 ln y L′ L (1, χ) � + O(1) as y → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
134
+ page_content=' Via exponentiation, we get 1 � L(1, χ) 1 √ln yeO � 1 ln y L′ L (1,χ) � +O(1) where we have used [8, Thm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
135
+ page_content=' 7, Col.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
136
+ page_content='] stating that � p≤y � 1 − 1 p � ≍ 1 ln y for every y > 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
137
+ page_content=' Now we prove proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
138
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
139
+ page_content=' Proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
140
+ page_content=' Fix a real valued non-principal Dirichlet character χ, and select a function λ sat- isfying the requirements of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
141
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
142
+ page_content=' According article [7], we have to bound the product � p≤x1/2 � 1 − 1 p + 1 � α(λ(x))−1 ≍ 1 ln x � L(1, χ) � ln λ(x)eO � 1 ln λ(x) L′ L (1,χ) � (21) as x → +∞, where we have used [7, Lem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
143
+ page_content=' 2], and lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
144
+ page_content='2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
145
+ page_content=' When χ is a non-principal character, then based on article [3] we have cεq(χ)−ε < L(1, χ) < ln q(χ) (22) where ε is any positive number and cε is a positive number depending on ε.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
146
+ page_content=' Also, based on [5, Prop.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
147
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
148
+ page_content='7] we have L′ L (1, χ) ≪ ln q(χ) (23) where the implied constant being absolute.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
149
+ page_content=' Using these bounds in expression (21) and the method from article [7] we can get the bounds in expression (1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
150
+ page_content=' Assuming that χ is primitive, q(χ) is big enough, and that there exists a positive constant cχ ≤ 1 such that L(s, χ) has no real zero in the interval (1 − cχ, 1), we have 1 ln q(χ) ≪ L(1, χ) (24) where the implied constant being positive and absolute, see article [4].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
151
+ page_content=' Using this bound in expression (21) and the method from article [7] we get the bound in expression (2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
152
+ page_content=' 8 G´abor Rom´an The proof of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
153
+ page_content='2 is the following.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
154
+ page_content=' Proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
155
+ page_content=' Fix a real valued non-principal Dirichlet character χ, and select a function λ sat- isfying the requirements of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
156
+ page_content='2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
157
+ page_content=' We are going to bound expression (21), but with bounds based on the Riemann hypothesis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
158
+ page_content=' If χ is a real valued non-principal Dirichlet character, and we assume that the Riemann hypothesis holds for L(s, χ), then based on [6, Thm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
159
+ page_content=' 1] we have 1 + o(1) ε1 ln ln q(χ) < L(1, χ) < (1 + o(1))ε1 ln ln q(χ) (25) as q(χ) → ∞, where ε1 and ε2 are real constants.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
160
+ page_content=' (As a side note, infinitely many real primitive characters χ satisfy these inequalities without assuming that the Riemann hypothesis holds for L(s, χ), see article [3].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
161
+ page_content=') In the same setting, based on [5, Thm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
162
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
163
+ page_content='17] we have L′ L (1, χ) ≪ ln ln q(χ) (26) where the implied constant being absolute.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
164
+ page_content=' Using these bounds in expression (21) and the method from article [7] we get the bounds in expression (3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
165
+ page_content=' To prove proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
166
+ page_content='3 and proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
167
+ page_content='4 we introduce the following function.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
168
+ page_content=' βm(y) := � p≤y p≡m(4) χ(p)=1 � 1 − 1 p + 1 � Lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
169
+ page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
170
+ page_content=' Let χ be a real valued non-principal Dirichlet character.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
171
+ page_content=' Then we have β1(y) ≍ 1 4� L(1, χ)L(1, χχ4,3) 1 4√ln yeO � 1 ln y L′ L (1,χ) � +O � 1 ln y L′ L (1,χχ4,3) � (27) and β3(y) ≍ 4 � L(1, χχ4,3) L(1, χ) 1 4√ln yeO � 1 ln y L′ L (1,χ) � +O � 1 ln y L′ L (1,χχ4,3) � (28) as y → +∞, where χ4,3 is the non-principal Dirichlet character modulo 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
172
+ page_content=' Proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
173
+ page_content=' Fix a real valued non-principal Dirichlet character χ, and let either m = 1, or m = 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
174
+ page_content=' We can rewrite βm(y) as � p≤y p≡m(4) χ(p)=1 � 1 − 1 p + 1 � = � p≤y p≡m(4) χ(p)=1 � 1 − 1 p2 �−1 � p≤y p≡m(4) χ(p)=1 � 1 − 1 p � On square-free numbers generated from given sets of primes II 9 where the first product on the right hand side can be bounded by a small positive constant.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
175
+ page_content=' Taking the logarithm of the second product on the right hand side we get � p≤y p≡m(4) χ(p)=1 ln � 1 − 1 p � = 1 2 � p≤y p≡m(4) (1 + χ(p)) ln � 1 − 1 p � where we can split the finite sum on the right hand side as 1 2 � p≤y p≡m(4) ln � 1 − 1 p � + 1 2 � p≤y p≡m(4) χ(p) ln � 1 − 1 p � .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
176
+ page_content=' (29) Based on [2, Thm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
177
+ page_content=' 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
178
+ page_content='16] we can write the second sum in expression (29) as 1 2 � p≤y χ(p) ln � 1 − 1 p � 1 ϕ(4) � χ4 χ4(p)χ4(m) (30) where the internal sum iterates through the ϕ(4) Dirichlet characters modulo 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
179
+ page_content=' There are two Dirichlet characters modulo 4;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
180
+ page_content=' we are going to denote them as χ4,1 (the principal character), and as χ4,3 (the non-principal character).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
181
+ page_content=' Splitting the internal sum, we get χ4,1(m) 4 � p≤y χ(p)χ4,1(p) ln � 1 − 1 p � + χ4,3(m) 4 � p≤y χ(p)χ4,3(p) ln � 1 − 1 p � .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
182
+ page_content=' (31) Concerning the sum on the left hand side of expression (31), as χ4,1(m) = 1;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
183
+ page_content=' and as χ4,1(p) = 1 when (p, 4) = 1, otherwise χ4,1(p) = 0, we have 1 4 � p≤y (p,4)=1 χ(p) ln � 1 − 1 p � = 1 4 � p≤y χ(p) ln � 1 − 1 p � + O(1) where we can use lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
184
+ page_content='1 to get −1 4 ln L(1, χ) + O � 1 ln y L′ L (1, χ) � + O(1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
185
+ page_content=' Concerning the sum on the right hand side of expression (31), χχ4,3 is a real valued non- principal character, so we can use lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
186
+ page_content='1 again to get −χ4,3(m) 4 ln L(1, χχ4,3) + O � 1 ln y L′ L (1, χχ4,3) � + O(1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
187
+ page_content=' Substituting these result in expression (29), and exponentiating, in the case when m = 1, we get expression (27) as χ4,3(1) = 1, and because � p≤y p≡m(4) � 1 − 1 p � ≍ 1 √ln y 10 G´abor Rom´an based on the article of Williams [11].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
188
+ page_content=' Similarly in the case when m = 3 we get expression (28) as χ4,3(3) = −1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
189
+ page_content=' Now we proof proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
190
+ page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
191
+ page_content=' Proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
192
+ page_content=' Fix a real valued primitive non-principal Dirichlet character χ;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
193
+ page_content=' and either let m = 1, or m = 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
194
+ page_content=' Furthermore select a function λ satisfying the requirements of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
195
+ page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
196
+ page_content=' Based on article [7], we have to bound the product � p≤x1/2 � 1 − 1 p + 1 � βm(λ(x))−1 (32) when m = 1, and separately when m = 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
197
+ page_content=' When m = 1, then expression (32) is asymptotic to 1 ln x 4� L(1, χ)L(1, χχ4,3) 4� ln λ(x)eO � 1 ln y L′ L (1,χ) � +O � 1 ln y L′ L (1,χχ4,3) � (33) as x → +∞, where we have used [7, Lem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
198
+ page_content=' 2] and lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
199
+ page_content='3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
200
+ page_content=' Using the bounds from expression (22) and expression (23) we can bound expression (33) from below as 1 ln x 4� ln λ(x) 4� q(χ)εq(χχ4,3)εeO � 1 ln y (ln q(χ)+ln q(χχ4,3)) � and as 1 ln x 4� ln q(χ) 4� ln q(χχ4,3) 4� ln λ(x)eO � 1 ln y (ln q(χ)+ln q(χχ4,3)) � from above.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
201
+ page_content=' As χ and χ4,3 are both primitive, their product χχ4,3 is primitive too, see [10, Ch.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
202
+ page_content=' 3].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
203
+ page_content=' But then q(χχ4,3) ∈ O(q(χ)), see [5, Sec.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
204
+ page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
205
+ page_content='3].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
206
+ page_content=' Based on this and on the method in article [7] we get the bounds in expression (4).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
207
+ page_content=' Assuming that q(χ) is big enough, and that there exists a positive constant cχ ≤ 1 such that L(s, χ) has no real zero in the interval (1−cχ, 1), we can use bound (24) in expression (33) to get 1 ln x 4� ln λ(x) 4� ln q(χ) 4� ln q(χχ4,3) eO � 1 ln y (ln q(χ)+ln q(χχ4,3)) � from where we can get bound (5) based on the previous train of thoughts.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
208
+ page_content=' When m = 3, then expression (32) is asymptotic to 1 ln x 4 � L(1, χ) L(1, χχ4,3) 4� ln λ(x)eO � 1 ln y L′ L (1,χ) � +O � 1 ln y L′ L (1,χχ4,3) � (34) as x → +∞, where we have used [7, Lem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
209
+ page_content=' 2] and lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
210
+ page_content='3 again.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
211
+ page_content=' As in the previous case, we can bound expression (34) from below as 1 ln x 4 � q(χ)−ε ln q(χχ4,3) 4� ln λ(x)eO � 1 ln y (ln q(χ)+ln q(χχ4,3)) � On square-free numbers generated from given sets of primes II 11 and from above as 1 ln x 4 � ln q(χ) q(χχ4,3)−ε 4� ln λ(x)eO � 1 ln y (ln q(χ)+ln q(χχ4,3)) � from where we get the bounds in expression (6).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
212
+ page_content=' Assuming that q(χ) is big enough, and that there exists a positive constant cχ ≤ 1 such that L(s, χ) has no real zero in the interval (1−cχ, 1), we can use bound (24) in expression (34).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
213
+ page_content=' The logarithmic contribution of the terms L(1, χ) and L(1, χχ4,3) “cancel” each other, and we get asymptotic (7).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
214
+ page_content=' And finally, the proof of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
215
+ page_content='4 is the following.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
216
+ page_content=' Proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
217
+ page_content=' Fix a real valued primitive non-principal Dirichlet character χ;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
218
+ page_content=' and either let m = 1, or m = 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
219
+ page_content=' We use the same method as in the proof of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
220
+ page_content='3, but this time we assume that the Riemann hypothesis holds for L(s, χ).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
221
+ page_content=' When m = 1, then we can use the bounds from expression (25) and expression (26) to bound expression (33) from below as 1 ln x 4� ln λ(x) 4� ln ln q(χ) 4� ln ln q(χχ4,3) eO � 1 ln y (ln ln q(χ)+ln ln q(χχ4,3)) � and from above as 1 ln x 4� ln ln q(χ) 4� ln ln q(χχ4,3) 4� ln λ(x)eO � 1 ln y (ln ln q(χ)+ln ln q(χχ4,3)) � as x → +∞ and as q(χ) → +∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
222
+ page_content=' Using the same train of thought as in the proof of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
223
+ page_content='3, we get the bounds in expression (8).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
224
+ page_content=' When m = 3, then using the above applied bounds (25) and (26) we get the asymptotic (9) from asymptotic (34) via “cancellation” again.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
225
+ page_content=' 3 Remarks As we have already mentioned in section 1, before proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
226
+ page_content='3, a natural extension of proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
227
+ page_content='1 and proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
228
+ page_content='2 would be to only allow P to contain primes which are congruent to some mi modulo q, where q > 0 is an integer, and the m1, .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
229
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
230
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
231
+ page_content=' , mk naturals are pairwise distinct relative primes to q.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
232
+ page_content=' In the case of modulo 4, the results were already different for distinct m, so we can expect a similar outcome for larger moduli.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
233
+ page_content=' However a general strategy for the generalisation could go along the following train of thoughts.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
234
+ page_content=' We would have to supply an asymptotic for the product � p≤y p≡m(q) χ(p)=1 � 1 − 1 p + 1 � 12 G´abor Rom´an which could be done by the refinement of lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
235
+ page_content='3, and its proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
236
+ page_content=' If we follow this path, then expression (30) will turn into 1 2 � p≤y χ(p) ln � 1 − 1 p � 1 ϕ(q) � χq χq(p)χq(m) where we can separate the principal character from the non-principal ones as χq,1(m) ϕ(q) � p≤y χ(p)χq,1(p) ln � 1 − 1 p � and what remains is 1 ϕ(q) � χq̸=χq,1 χq(m) � p≤y χ(p)χq(p) ln � 1 − 1 p � .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
237
+ page_content=' Due to the fact that χq,1 is the principal character, the first sum can be handled with our already presented techniques.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
238
+ page_content=' The double sum is more problematic.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
239
+ page_content=' On the one hand, for the internal sum we would have to refine lemma 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
240
+ page_content='1 and its proof.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
241
+ page_content=' We would have to make sure that when χ is complex valued, then we can take the logarithm of L(1, χ) and its product form;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
242
+ page_content=' furthermore that the values of these two logarithms match.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
243
+ page_content=' On the other hand, we would have to obtain a good estimation for the external sum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
244
+ page_content=' References [1] Abramowitz M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
245
+ page_content=' and Stegun I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
246
+ page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
247
+ page_content=': Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
248
+ page_content=' Dover publications (1972).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
249
+ page_content=' [2] Apostol T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
250
+ page_content=' M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
251
+ page_content=': Introduction to Analytic Number Theory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
252
+ page_content=' Springer-Verlag (1976).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
253
+ page_content=' [3] Bateman P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
254
+ page_content=' T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
255
+ page_content=' and Chowla S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
256
+ page_content=' and Erd˝os P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
257
+ page_content=': Remarks on the size of L(1, χ).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
258
+ page_content=' Publ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
259
+ page_content=' Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
260
+ page_content=' Debrecen 1 (2-4) (1950) 165–182.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
261
+ page_content=' [4] Hoffstein J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
262
+ page_content=': On the Siegel–Tatuzawa theorem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
263
+ page_content=' Acta.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
264
+ page_content=' Arith.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
265
+ page_content=' 38 (2) (1980) 168–174.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
266
+ page_content=' [5] Iwaniec H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
267
+ page_content=' and Kowalski E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
268
+ page_content=': Analytic Number Theory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
269
+ page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
270
+ page_content='M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
271
+ page_content='S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
272
+ page_content=' Colloquium Publications (2004).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
273
+ page_content=' [6] Littlewood J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
274
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
275
+ page_content=': On the class-number of the corpus P( √ −k).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
276
+ page_content=' Proc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
277
+ page_content=' London Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
278
+ page_content=' Soc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
279
+ page_content=' 27 (1) (1928) 358–372.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
280
+ page_content=' [7] Rom´an G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
281
+ page_content=': On square-free numbers generated from given sets of primes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
282
+ page_content=' Comm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
283
+ page_content=' Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
284
+ page_content=' 30 (1) (2022) 229–237.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
285
+ page_content=' [8] Rosser J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
286
+ page_content=' B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
287
+ page_content=' and Schoenfeld L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
288
+ page_content=': Approximate formulas for some functions of prime numbers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
289
+ page_content=' Illinois Journal of Mathematics 6 (1) (1962) 64–94.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
290
+ page_content=' [9] Walfisz A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
291
+ page_content=': Zur additiven Zahlentheorie II.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
292
+ page_content='. Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
293
+ page_content=' Z.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
294
+ page_content=' 40 (1) (1936) 592–607.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
295
+ page_content=' [10] Washington L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
296
+ page_content=' C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
297
+ page_content=': Introduction to Cyclotomic Fields.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
298
+ page_content=' Springer-Verlag (1982).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
299
+ page_content=' [11] Williams K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
300
+ page_content='S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
301
+ page_content=' : Mertens’ Theorem for Arithmetic Progressions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
302
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
303
+ page_content=' Number Theory 6 (5) (1974) 353–359.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
304
+ page_content=' Received: Received date Accepted for publication: Accepted date Communicated by: Handling Editor' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1dE0T4oBgHgl3EQfdgBe/content/2301.02377v1.pdf'}
2NE0T4oBgHgl3EQfugGB/vector_store/index.faiss ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe99fa4fa0464d5ed2127f2f9e703e46a87ded1eccc35033da7458ffbeba8640
3
+ size 26148909
2dFIT4oBgHgl3EQf5Cs5/content/2301.11387v1.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3467a9d54f660f6ab44acdb114c640f1529407208ff19b6a85aea6925425c7b3
3
+ size 10337001
2tFQT4oBgHgl3EQfGDUx/vector_store/index.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2dd87bbd63e48c6f48e997cbb21999575bd19f433505dcec0ec0ffdcb0aec217
3
+ size 111645
6NFKT4oBgHgl3EQf_C4s/content/tmp_files/2301.11960v1.pdf.txt ADDED
@@ -0,0 +1,482 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ A Global Inventory of Feedback
2
+ Timothy M. Heckman
3
+ The William H. Miller III Department of Physics and Astronomy, The Johns Hopkins University,
4
+ Baltimore, MD 21218, USA
5
+
6
+ Philip N. Best
7
+ Institute for Astronomy, University of Edinburgh, Royal Observatory, Blackford Hill, Edinburgh,
8
+ EH9 3HJ, UK
9
+
10
+ Abstract
11
+ Feedback from both supermassive black holes and massive stars plays a fundamental role in the
12
+ evolution of galaxies and the inter-galactic medium. In this paper we use available data to
13
+ estimate the total amount of kinetic energy and momentum created per co-moving volume
14
+ element over the history of the universe from three sources: massive stars and supernovae,
15
+ radiation pressure and winds driven by supermassive black holes, and radio jets driven by
16
+ supermassive black holes. Kinetic energy and momentum injection from jets peaks at z ≈ 1,
17
+ while the other two sources peak at z ≈ 2. Massive stars are the dominant global source of
18
+ momentum injection. For supermassive black holes, we find that the amount of kinetic energy
19
+ from jets is about an order-of-magnitude larger than that from winds. We also find that amount
20
+ of kinetic energy created by massive stars is about 2.5 εstar times that carried by jets (where εstar
21
+ is the fraction of injected energy not lost to radiative cooling). We discuss the implications of
22
+ these results for the evolution of galaxies and the IGM. Because the ratio of black hole mass to
23
+ galaxy mass is a steeply increasing function of mass, we show that the relative importance of
24
+ black hole feedback to stellar feedback likewise increases with mass. We show that there is a
25
+ trend in the present-day universe which, in the simplest picture, is consistent with galaxies that
26
+ have been dominated by black hole feedback being generally quenched, while galaxies that
27
+ have been dominated by stellar feedback are star-forming. We also note that the amount of
28
+ kinetic energy carried by jets and winds appears sufficient to explain the properties of hot gas
29
+ in massive halos (> 1013 Mʘ).
30
+ 1. Introduction
31
+ The basic properties of galaxies, supermassive black holes, and the intra-group/intra-cluster
32
+ medium cannot be understood without considering the impact of the return of mass, metals,
33
+ energy, and momentum from both populations of massive stars (stellar winds and supernovae)
34
+ and supermassive black holes (winds and jets). Examples include the shape of the stellar mass
35
+ function, the quenching and subsequent suppression of star-formation in massive galaxies, the
36
+ mass-metallicity and mass-radius relations, the Kennicutt-Schmidt law of star-formation, and
37
+ the group/cluster X-ray luminosity-temperature relation (see reviews by Somerville & Davé
38
+ 2015, Naab & Ostriker 2017, Donahue & Voit 2022).
39
+
40
+ This input of energy and momentum from massive stars and black holes is generically referred
41
+ to as feedback. Even the highest resolution numerical simulations cannot fully include all the
42
+ relevant physics ab-initio, and must rely on “sub-grid physics” (essentially, recipes for processes
43
+ that cannot be spatially resolved). The same is true of semi-analytic models. This underscores
44
+ the importance of using observations to inform the choices that are made in simulations and
45
+ models. While there is now a considerable body of data on feedback from both massive stars
46
+ and supermassive back holes (e.g. Veilleux et al. 2020; McNamara & Nulsen 2007; Thompson &
47
+ Heckman 2023), we still have a very incomplete understanding of the impact of this feedback
48
+ on the surrounding gas.
49
+ In this paper, we will take a different approach from previous investigations of feedback, and
50
+ try to compile a global inventory (that is, integrated over cosmic time) of the amount of kinetic
51
+ energy and momentum per co-moving volume element injected by massive stars and
52
+ supermassive black holes. We will then compare the respective importance of these feedback
53
+ sources as a function of time and of galaxy and black hole mass.
54
+ 2. Methodology
55
+ 2.1 Massive Stars
56
+ To compute the total amount of kinetic energy injected by massive stars (stellar winds and
57
+ supernovae) per unit volume, we start with the present-day amount of stellar mass per unit
58
+ volume. We use the compilation in Madau & Dickinson (2016), adjusted to a standard Chabrier
59
+ Initial Mass Function (IMF - Chabrier, 2003). This value is 3.4 x 108 Mʘ Mpc-3. To compute the
60
+ corresponding amount of kinetic energy we need to first correct this to account for stars that
61
+ were formed but are no longer present. This requires multiplication by 1/(1-R), where R is the
62
+ so-called Returned Fraction, which is 0.3 for a Chabrier IMF. Thus, the total mass of stars
63
+ formed per unit volume is 4.9 x 108 Mʘ Mpc-3. Starburst99 (Leitherer et al. 1999) models for a
64
+ Chabrier IMF yield a total kinetic energy in stellar winds and supernova ejecta of 6.9 x 1015 erg
65
+ gm-1. This then gives a value for the kinetic energy density due to stars of Ustar = 6.8 x 1057 erg
66
+ Mpc-3.
67
+ How much of this kinetic energy is available to supply feedback? The stellar ejecta initially
68
+ carrying the energy collide and their kinetic energy is converted to thermal energy. This hot gas
69
+ can then expand and flow outward with the thermal energy being converted back into kinetic
70
+ energy (e.g. Chevalier & Clegg 1985). Some of the initial thermal energy can be lost through
71
+ radiative cooling, so that only a fraction εstar remains to provide feedback. Numerical
72
+ simulations that represent typical conditions in low-z star-forming galaxies yield εstar ≈ 0.1 (Kim
73
+ et al. 2020), with a value that increases with the star-formation rate per unit area (SFR/A). At
74
+ the much higher values of SFR/A seen in starbursts (e.g. Kennicutt & Evans 2012), simulations
75
+ and models predict far greater efficiency, with εstar ≈ 0.3 to 1.0 (Schneider et al. 2020; Fielding &
76
+ Bryan 2022). This is consistent with X-ray observations of the H-like and He-like Fe Kα emission-
77
+ lines in starburst galaxies from the very hot (108 K) gas created as the stellar ejecta are
78
+ thermalized through shocks (Thompson & Heckman 2023). These results imply that rather little
79
+
80
+ of the initial kinetic energy is lost through radiative cooling, and this is substantiated by
81
+ estimates of the rate of PΔV work done by the wind on the ambient gas (Thompson & Heckman
82
+ 2023). While there are no such constraints on galaxies at high (z > 1) redshift, we do know that
83
+ these galaxies have values of SFR/A similar to those seen in low-z starbursts (e.g. Forster-
84
+ Schreiber & Wuyts 2020), and that galactic winds driven by massive stars at this epoch are both
85
+ ubiquitous and very similar to those seen in low-z starburst galaxies (see Thompson & Heckman
86
+ 2023). Note that roughly 60% of the total present-day stellar mass was formed at z > 1, during
87
+ this “windy” epoch (Madau & Dickinson 2016).
88
+ The situation for momentum injection is less uncertain because momentum will be conserved
89
+ even in the face of significant radiative losses. We can simply use the methodology above but
90
+ use Starburst99 to compute the specific injection rate of momentum by massive stars
91
+ (supernovae, stellar winds, and radiation pressure). The value is 7.4 x 107 cm s-1, and for a total
92
+ stellar mass density of 4.8 x 108 Mʘ Mpc-3, this yields 7.1 x 1049 gm cm s-1 Mpc-3.
93
+ 2.2 Black-Hole Driven Winds and Radiation Pressure
94
+ Winds driven by supermassive black holes are multi-phase and have been measured in a
95
+ number of different ways. Molecular outflows have been detected in both emission and
96
+ absorption (see the review by Veilleux et al. 2020). Calculating kinetic energy outflow rates is
97
+ conceptually straightforward. The luminosity of a CO transition can be converted into a total
98
+ molecular gas mass, albeit with uncertainties (Tacconi et al. 2020). The measured outflow
99
+ velocity and the radius of the outflow then yields a kinetic energy flux given as ½ Mgas vout3 rout-1.
100
+ For absorption, the OH column density and outflow velocity yields an outflow rate (for an
101
+ assumed outflow size and OH/H2 conversion factor). The first compilation of molecular outflows
102
+ by Fiore et al. (2017) implied typical kinetic energy fluxes of dEwind/dt ≈ 3% LBol, however more
103
+ recent compilations of measurements (Lutz et al. 2020, Lamperti et al. 2022 and private
104
+ communication) have yielded much smaller values (median of 0.1%).
105
+ Similarly, the outflow rates of warm ionized gas can be measured using the Hα or [OIII]5007
106
+ luminosity and measured electron density to derive the total mass of ionized gas and then
107
+ measuring the outflow velocity and radius of the outflow to determine dEwind/dt. Different
108
+ recent measurements have come to drastically different results, with median values ranging
109
+ from as high as 1% of Lbol (Kakkad et al. 2022) to 0.3 % (Fiore et al. 2017), to 0.1% (Revalski et al.
110
+ 2021), to 0.01% (Dall’Agnol de Oliveira 2021), to 0.0003% (Trindade Falcao et al. 2021).
111
+ An independent measurement of the outflow rate in the warm ionized gas comes from
112
+ observations of BAL QSOs. Here, the absorption-lines can provide a column density and outflow
113
+ velocity. Direct measurements of the electron densities can be made using the ratio of column
114
+ densities in lines arising from an excited state vs. the ground state. Photoionization models
115
+ using the observed ionizing luminosity (Q) and the inferred value of the ionization parameter
116
+ (U) then yield a size for the outflow: rout = (Q/4π ne c U)1/2 (Miller et al. 2020). With a velocity,
117
+ radius, and column density, the kinetic energy flux can be estimated. The results span a huge
118
+
119
+ range, from 0.001% to 10% Lbol (median value of 0.3%). Highly ionized outflows are also
120
+ detected in about 40% of AGN (Tombesi et al. 2011) based on X-ray absorption-lines. However,
121
+ because the size scales of these outflows are so uncertain, the kinetic energy outflow rates are
122
+ also uncertain (by about two orders-of-magnitude, typically ranging between 0.01 and 1% of
123
+ LBol – Tombesi et al. 2012).
124
+ It is clear from the above that assigning a value for the ratio of dEwind/dt to Lbol is difficult. If we
125
+ take the median values of 0.1%, 0.3%, and 0.1% LBol for the molecular, warm-ionized, and
126
+ highly-ionized phases, we get a total value of 0.5% LBol. Multiplying this by the total bolometric
127
+ energy density per co-moving volume element volume produced by supermassive black holes
128
+ of Urad = 8.6 x 1058 erg Mpc-3 (Hopkins et al. 2007), yields Uwind = 4.3 x 1056 ergs Mpc-3. This is
129
+ 6% as large as the value derived for massive stars. Using the present-day mass per unit volume
130
+ in supermassive black holes of ρBH = 5 x 105 Mʘ Mpc-3 (Hopkins et al. 2007) this wind energy
131
+ density can also be expressed as Uwind = 5 x 10-4 ρBH c2.
132
+ We can also consider the amount of momentum provided by AGN. An initial estimate is
133
+ implied by the momentum carried by radiation (Urad/c) where Urad is the total amount of radiant
134
+ energy per unit volume produced over cosmic time by AGN. This yields an amount of
135
+ momentum per unit volume of 2.9 x 1048 gm cm s-1 Mpc-3 (about 4% of the value for massive
136
+ stars). Since the momentum flux (in the non-relativistic case) is just twice the kinetic energy flux
137
+ divided by the outflow velocity, we need only consider the momentum carried by the molecular
138
+ and warm ionized flows (since the BAL QSO and X-ray outflows are over an order-of-magnitude
139
+ faster, but carry similar kinetic energy fluxes).
140
+ For the molecular outflows, the data in Lutz (2020) and Lamperti et al. (2022 and private
141
+ communication) yield median values of dpwind/dt = 1.0 and 0.7 LBol/c respectively. The near
142
+ equality is consistent with the idea that the molecular outflows are driven by radiation
143
+ pressure. If so, then combining radiation pressure and the molecular outflows would be double-
144
+ counting in the inventory of momentum.
145
+ As noted above, there is a very wide range in the ratio between the kinetic energy flux in the
146
+ warm ionized gas and the AGN bolometric luminosity, and this translates directly into
147
+ uncertainties in the ratio of momentum flux and radiation pressure for this gas phase.
148
+ Estimated median values of this ratio range from ≈10 (Kakkad et al. 2022), to ≈1 (Fiore et al.
149
+ 2017; Revalski et al. 2021), to ≈0.1 (Dall’Agnol de Oliveira et al. 2021), to ≈0.01 (Trindade Falcao
150
+ et al. 2021). It appears that the momentum flux in the warm ionized outflows is not likely to be
151
+ significantly larger than those in the molecular gas or to that carried by radiation.
152
+ This represents a total injected momentum per unit volume of at most ≈1049 gm cm s-1 Mpc-3,
153
+ even if we simply add the three sources (radiation, molecular gas, ionized gas) together. This is
154
+ still an order of magnitude below the value for massive stars.
155
+
156
+
157
+ 2.3 Black Hole-Driven Jets
158
+ The earliest evidence for the outflow of kinetic energy driven by supermassive black holes came
159
+ from observations of “double lobes” of synchrotron radio emission that straddled massive
160
+ elliptical galaxies (Baade & Minkowski 1954). Subsequent radio observations at high angular
161
+ resolution showed narrow collimated features (“jets”) linking the two lobes to the galactic
162
+ nucleus (see Miley 1980).
163
+ It is now possible to quantify the amount of kinetic energy carried by jets as a function of the
164
+ luminosity of the radio source that they power. This can be done by joint observations of the
165
+ radio and X-ray emission. The expanding radio sources inflate lobes of relativistic plasma, which
166
+ in X-rays can be observed as cavities in the surrounding hot gas. Bırzan et al. (2004,2008), Dunn
167
+ et al. (2005), Rafferty et al. (2006), and Cavagnolo et al. (2010) derived the pΔV work (energy)
168
+ associated with the cavities in a sample of massive galaxies, groups, and clusters, and used the
169
+ buoyancy timescale (e.g. Churazov et al. 2001) to estimate their ages. They combined these
170
+ cavity powers with the monochromatic 1.4 GHz radio luminosities to show that the two were
171
+ well-correlated. The largest uncertainty in this method is the determination of the cavity energy
172
+ from the measured pressure and volume: Ecav = fcavpΔV . For the relativistic plasma of the radio
173
+ lobes the enthalpy of the cavity is 4pΔV. Taking fcav = 4, Heckman & Best (2014) derived the
174
+ following best-fit relation from the cavity data:
175
+ 1) dEjet/dt = 1.3 × 1038 (L1.4GHz/1026 W Hz−1)0.68 W
176
+ This empirical relation is very similar to predictions from theoretical models of radio jets.
177
+ Willett et al. (1999) used synchrotron properties to derive the relation:
178
+ 2) dEjet/dt = 2.8 × 1036 (fW)3/2 (L1.4GHz/1026 W Hz−1)0.84 W
179
+ Here fW is a dimensionless factor (in the range 1 to 20) accounting for the uncertainties in the
180
+ extrapolation from the population of relativistic electrons that produce the observed radio
181
+ synchrotron emission to the total energy. Agreement with the X-ray cavity data implies fW ≈10
182
+ to 20 (see Heckman & Best 2014).
183
+ We adopt the theoretical relation (equation 2), but calibrated by the cavity data (i.e. taking fW =
184
+ 15) and use this to convert the radio luminosity function of AGN between z = 0.1 and 3 (Yuan et
185
+ al. 2017) into a measure of the evolution in the rate of kinetic energy injection per unit volume
186
+ by radio jets.
187
+ The results are shown in Figure 1, and show that the peak rate of kinetic energy injection by jets
188
+ occurs at a significantly lower redshift (z ≈ 1) than the peak rate due to massive stars and black-
189
+ hole-driven winds (z ≈ 2, as also shown in Figure 1). We then integrate the energy injection rate
190
+ by interpolating the values at z = 0.1, 0.5, 1.0, 2.0, and 3.0 and extrapolating from z = 0.1 to 0
191
+ and from z = 3.0 to infinity (this extrapolation does not add significantly to the total - see Figure
192
+ 1). This then gives a value for the time-integrated total kinetic energy per unit volume due to
193
+
194
+ jets of Ujet = 2.6 x 1057 erg Mpc-3. These results are broadly in line with similar estimates derived
195
+ from low-frequency radio luminosity functions (Kondapally et al., private communication).
196
+ The time-integrated kinetic energy input from jets is ≈6 times larger than the value estimated
197
+ about for black-hole-driven winds, and 40% (400%) the total amount of kinetic energy
198
+ generated by massive stars for εstar = 1 (0.1). Alternatively, using the present-day mass per unit
199
+ volume in supermassive black holes of ρBH = 5 x 105 Mʘ (Hopkins et al. 2007) this jet energy
200
+ density can also be expressed as Ujet = 2.9 x 10-3 ρBH c2.
201
+ The kinetic energy carried by jets is in the form of relativistic bulk motion. In this case, the
202
+ momentum can be taken as p ≈ KE/c. The above value of Ujet then implies a momentum density
203
+ of 8.7 x 1046 gm cm s-1 Mpc-3. This is much less than the momentum carried by radiation and
204
+ winds from supermassive black holes, and the momentum produced by massive stars. Jets are
205
+ therefore far more important feedback sources in terms of kinetic energy than momentum.
206
+ 2.4 The Bottom Line
207
+ For total kinetic energy inventory, the largest single source is either massive stars (for εstar > 0.4)
208
+ or jets (for εstar < 0.4). AGN winds are only important at the <10% level. For the total
209
+ momentum inventory, massive stars dominate (AGN contribute at the ≈10% level). The peak
210
+ rate of kinetic energy injection by jets occurs at a substantially lower redshift than that from
211
+ stars or AGN winds (z ≈ 1 and 2, respectively). These results are summarized in Table 1 and
212
+ Figure 1.
213
+ ______________________________________________________________________________
214
+ Table 1 – Summary of Feedback Inventory
215
+ 1
216
+ 2
217
+ 3
218
+ 4
219
+ 5
220
+ 6
221
+ Sample
222
+ Log ρ
223
+ Log sKE
224
+ Log ρKE
225
+ Log sp
226
+ ρp
227
+ Massive Stars
228
+ 8.69
229
+ -5.11
230
+ 57.83
231
+ 7.87
232
+ 49.85
233
+ BH Winds
234
+ 5.70
235
+ -3.30
236
+ 56.63
237
+ 10.00
238
+ 49.00
239
+ BH Jets
240
+ 5.70
241
+ -2.54
242
+ 57.43
243
+ 7.94
244
+ 46.94
245
+ ______________________________________________________________________________
246
+ Notes:
247
+ Column 2 – The log of the present-day mass density of stars (row 3) and supermassive black holes (rows
248
+ 4 and 5) formed over cosmic time in units of Mʘ Mpc-3.
249
+ Column 3 – The log of the specific kinetic energy released: energy per unit mass in stars (row 3 and black
250
+ holes (rows 4 and 5). Given in units of c2, and assuming εstar = 1.0.
251
+ Column 4 – The log of the amount of kinetic energy created per unit volume (in ergs Mpc-3).
252
+ Column 5 – The specific momentum created (momentum per unit mass in stars (row 3) and black holes
253
+ (rows 4 and 5). In units of cm s-1.
254
+ Column 6 – The log of the amount of momentum created per unit volume (gm cm s-1 Mpc-3).
255
+ ______________________________________________________________________________
256
+
257
+
258
+ Figure 1 – A plot of the amount of kinetic energy injected per Gyr and co-moving cubic Mpc as a function
259
+ of lookback time for massive stars (supernovae and stellar winds; black) and black-hole-driven jets (blue)
260
+ and winds (red). For massive stars we show the cases in which 100% (solid line) and 10% (dashed line) of
261
+ the kinetic energy created is delivered to the surroundings (i.e. not lost to radiative cooling). Note that
262
+ for momentum injection, massive stars dominate at all epochs, with the same time dependence as for
263
+ kinetic energy injection (i.e. as given by the solid black line).
264
+ 3. Implications
265
+ 3.1 For Galaxies
266
+ To assess the implications of these results for galaxy evolution, it is essential to consider the
267
+ dependences of feedback on the masses of both galaxies and supermassive black holes. We can
268
+ go beyond these simple global values and examine the relative importance of feedback (both
269
+ kinetic energy and momentum) as a function of the ratio of supermassive black hole mass to
270
+ galaxy stellar mass. In Figure 2 we show a plot of black hole vs. galaxy mass that is similar to
271
+ that in Heckman & Best (2014) for the z ≈ 0.1 universe (based on SDSS). In this case, these are
272
+ present day stellar masses, and would need to be increased by a factor 1/(1-R) = 1.42 to
273
+ represent the total mass of stars ever formed. The masses for the black holes were estimated
274
+ from the M-σ relation from McConnell & Ma (2013). In figure 2, we have color-coded the plot
275
+ by the fraction of galaxies in which star-formation has been quenched, which we define to be
276
+
277
+ Redshift
278
+ 0
279
+ 0.5
280
+ 1
281
+ 2
282
+ 346
283
+ 57.5
284
+ 57.0
285
+ 56.5
286
+ 56.0
287
+ 55.5
288
+ 55.0
289
+ Supernovae (8star = 1.0)
290
+ - - Supernovae (star = 0.1)
291
+ 54.5
292
+ -AGNjets
293
+ AGNwinds
294
+ 54.0
295
+ 0
296
+ 2
297
+ 4
298
+ 6
299
+ 8
300
+ 10
301
+ 12
302
+ Lookback time / GyrSFR/Mstar < 10-11 yr-1. It is clear that the quenched fraction depends strongly on both the stellar
303
+ and black hole masses.
304
+ The mean relation between stellar and black hole mass in Figure 2 can be approximated as log
305
+ MBH = 2.0 log Mstar -14.0, implying MBH/Mstar α Mstar1.0 α MBH0.50. Thus, the relative importance of
306
+ feedback integrated over cosmic time from massive stars and black holes should be a strong
307
+ function of mass. Let us quantify this for kinetic energy and then for momentum. For kinetic
308
+ energy, in a given galaxy (and assuming that global averages can be applied to individual
309
+ galaxies; see below) the inventories above imply that the ratio KEBH/KEstar = 315 εstar-1 MBH/Mstar
310
+ (where Mstar is the present-day stellar mass). For momentum, the corresponding ratio is
311
+ pBH/pstar = 100 MBH/Mstar. We can then plot these relations in Figure 2 to see the regimes in
312
+ which feedback from supermassive black holes exceeds that from stars. For kinetic energy, we
313
+ show this separately for values of εstar = 0.1 and 1.0.
314
+
315
+ Figure 2 – A plot of the distribution of SDSS galaxies in the plane of galaxy stellar mass vs. supermassive
316
+ black hole mass. The latter were estimated using the MBH vs. σ relation in McConnell & Ma (2013). The
317
+ relative numbers of galaxies in each bin are indicated by the green contours (increasing by factors of 2)
318
+ and the color-coding represents the fraction of galaxies that are quenched (SFR/Mstar < 10-11 yr-1). The
319
+ dark blue dashed line indicates where the momentum injected by black holes equals that from massive
320
+ stars. The two light blue dashed lines indicate where the kinetic energy from black holes equals that from
321
+ massive stars for values of εstar = 0.1 and 1.0 (see text). The transition from predominantly star forming
322
+ to predominantly quenched galaxies occurs near the relationship for εstar = 0.1.
323
+
324
+ 10
325
+ 1.0
326
+ 0.8
327
+ Quenched fraction
328
+ 0.6
329
+ 8
330
+ 0.4
331
+ 0.1
332
+ KEBH
333
+ 0.2
334
+ 6
335
+ 0.0
336
+ 10.0
337
+ 10.5
338
+ 11.0
339
+ 11.5
340
+ 12.0
341
+ log1o(Stellar mass / Msun)In terms of momentum input, stars dominate over black holes in almost all cases. However,
342
+ considering kinetic energy, we find that the transition from galaxies that are mostly quenched
343
+ to those that are mostly star-forming occurs very near the dividing line between jet-dominated
344
+ feedback and stellar dominated feedback for a value of εstar ≈ 0.1. This is suggestive evidence
345
+ that quenching is driven by the feedback of kinetic energy from jets driven by supermassive
346
+ black holes. However, we caution that the transition from star forming to quiescent galaxies
347
+ also occurs at the transition from disk dominated to bulge dominated galaxies, so the causal
348
+ connections between galaxy structure, star formation, and black hole feedback are not entirely
349
+ clear.
350
+ We emphasize that the relations plotted in Figure 2 explicitly assume that the global relations
351
+ can be applied to individual galaxies, namely that the amount of feedback from massive stars in
352
+ a given galaxy is proportional to stellar mass and that the amount of feedback from jets is
353
+ proportional to the black hole mass. The former seems like a safe assumption, but the
354
+ dependence of the production of radio jets on black hole mass may be complex. We know that
355
+ there are essentially two populations of radio galaxies (e.g. Heckman & Best 2014). In one case
356
+ (“radiative mode”) the jets are launched by star-forming galaxies and are accompanied by
357
+ strong nuclear radiation (QSO-like). In the other class (“jet-mode”) the jets are launched by
358
+ quenched galaxies, with little accompanying nuclear radiation. The radiative mode becomes
359
+ more important at higher luminosities and at higher redshifts. For the jet-mode galaxies,
360
+ Sabater et al. (2019) find that the probability of producing a jet with a given luminosity depends
361
+ on both the stellar and black hole mass (and more strongly on the former).
362
+ The situation for radiative-mode radio galaxies is less clear, although the indications are that
363
+ any dependence of the ratio of KE/MBH on MBH or Mstar is weaker (e.g. Janssen et al. 2012,
364
+ Kondapally et al. 2022). In the context of Figure 2, it may be that the jet-mode is not the
365
+ dominant population in terms of actively quenching, since the jet-mode galaxies are already
366
+ quenched (instead, these may just ‘maintain’ a quenched state). If quenching is due to jets in
367
+ radiative-mode galaxies, the dividing line between quenched and star-forming galaxies in Figure
368
+ 2 would imply that time-integrated amount of jet energy contributed by a radiative mode
369
+ galaxy is proportional to its black hole mass (i.e. the integrated ratio of jet kinetic energy and
370
+ energy carried by radiation is independent of black hole mass in these galaxies).
371
+ Another way to consider this is to ask how the amount of energy supplied by stars and by black
372
+ holes scales with the binding energy of the galaxy. We take Ebind ≈ Mstar vc2 where vc is the galaxy
373
+ circular velocity. The Tully-Fischer relation for disk galaxies (McGaugh et al. 2000) and the
374
+ Faber-Jackson relation for ellipticals (Bernardi et al. 2003) both imply vc α Mstar1/4. Thus, we
375
+ have Ebind α Mstar3/2. Given that KEstar α Mstar and KEBH α MBH α Mstar2, this implies that KEstar/Ebind
376
+ α Mstar-1/2, while KEBH/Ebind α Mstar1/2 α MBH1/4. This again underscores the fundamental
377
+ difference in the mass-dependence of feedback from massive stars and supermassive black
378
+ holes: feedback from stars becomes increasingly impactful on the galaxy as the mass decreases,
379
+ while feedback from black holes has greater impact as the mass increases.
380
+
381
+ 3.2 For the Intra-Group and Intra-Cluster Media
382
+ It has long been known that the basic observed properties of the hot gas in groups and clusters
383
+ of galaxies (Mhalo > 1013 Mʘ) are not consistent with simple models of purely gravitational
384
+ processes operating during the formation of these systems (see Donahue & Voit 2022 and
385
+ references therein). A particularly simple example of this is the observed relationship between
386
+ the X-ray temperature (a proxy for halo mass) and X-ray luminosity. As the halo masses
387
+ decrease, the observed X-ray luminosities fall further below the relationship expected simply
388
+ from gravitational infall and heating. These lower luminosities arise because the hot gas in
389
+ these less-massive halos is more spatially-extended than the dark matter, with the resulting
390
+ drop in gas density leading to lower X-ray luminosities.
391
+ This could be due to the feedback of energy injected into the hot gas, which “lifts” the gas
392
+ outward. As described above, there is direct observational evidence in the local universe of
393
+ radio jets delivering energy to the hot gas in groups and clusters. As discussed in Donahue &
394
+ Voit (2022), for this to be responsible for lifting the hot gas, an amount of kinetic energy equal
395
+ to ≈0.5% MBH/c2 must be delivered. This is close to the value for jets and AGN winds that we
396
+ estimated above of ≈0.34%. Note that this could be supplemented by the kinetic energy from
397
+ massive stars (which would be 0.25 to 2.5 the value for jets for εstar = 0.1 and 1.0 respectively).
398
+ 4. Summary
399
+ Based on a global inventory of the amount of kinetic energy and momentum injected by
400
+ massive stars (stellar winds and supernovae), and by winds and jets driven by supermassive
401
+ black holes, we draw the following conclusions:
402
+ i)
403
+ The major sources of kinetic energy are massive stars and jets. Winds driven by
404
+ supermassive black holes provide <10% of the total. The global ratio of the kinetic
405
+ energy injected by massive stars to that injected by jets is 2.5 εstar (where εstar is the
406
+ fraction of injected energy from stars that is not lost to radiative cooling).
407
+ ii)
408
+ Massive stars are the dominant source of momentum injection (90% of the total).
409
+ AGN winds provide 10%, and radio jets are negligible.
410
+ iii)
411
+ The peak in the feedback from jets occurs at z ≈ 1, considerably later than the
412
+ contributions of AGN-winds and massive stars (peaking at z ≈ 2).
413
+ iv)
414
+ Since the ratio of the mass of the supermassive black hole to the galaxy stellar mass
415
+ increases steeply with mass, there will be a mass-dependence in the relative
416
+ importance of feedback from the two sources.
417
+ v)
418
+ For the assumptions that the total amount of kinetic energy from massive stars is
419
+ proportional to the galaxy’s stellar mass, and that the total amount of kinetic energy
420
+ from a supermassive black hole is proportional to its mass, we find that the
421
+ populations of quenched and star-forming galaxies occur in the regimes where
422
+ supermassive black hole feedback and massive star feedback dominate, respectively
423
+ (for a value of εstar ≈ 0.1).
424
+
425
+ vi)
426
+ By comparing the amount of kinetic energy injected as a function of the binding
427
+ energy of a galaxy, we show that feedback becomes more impactful as galaxy mass
428
+ decreases for massive stars, but more impactful as galaxy mass increases for black
429
+ holes.
430
+ vii)
431
+ The global amount of kinetic energy injected by radio jets and AGN winds per unit
432
+ volume, combined with the supermassive black hole mass function, yields an
433
+ efficiency for producing kinetic energy in jets of 0.34% c2. This is very close to the
434
+ amount of energy needed to explain X-ray luminosity-temperature relation in groups
435
+ and clusters (0.5% c2).
436
+
437
+ REFERENCES
438
+ Baade, W. & Minkowski, R. 1954, ApJ, 119, 206
439
+ Bernardi, M 2003, AJ, 125, 1849
440
+ Bırzan L., Rafferty D., McNamara B., Wise M., Nulsen P. 2004, ApJ, 607, 800
441
+ Bırzan L., McNamara B., Nulsen P., Carilli C., Wise, M. 2008. ApJ, 686, 859
442
+ Cavagnolo K., McNamara B., Nulsen P. et al. 2010, ApJ, 720, 1066
443
+ Chabrier, G. 2003, PASP, 115, 763
444
+ Chevalier, R., & Clegg, A. 1985, Nature, 317, 44
445
+ Churazov, E., Bruggen, M., Kaiser C., Bohringer H., Forman W. 2001, Ap J, 554,261
446
+ Dall'Agnol de Oliveira, B., Storchi-Bergmann, T., Kraemer, S. et al. 2021, MNRAS, 504, 3890
447
+ Donahue, M., & Voit, G.M, 2022, Physics Reports, 973, 1
448
+ Dunn, R., Fabian, A., Taylor, G. 2005, MNRAS, 364, 1343
449
+ Fielding, D. & Bryan, G. 2022, ApJ, 984, 82
450
+ Fiore, F., Feruglio, C., Shankar, F. et al. 2017, A&A, 601, 143
451
+ Heckman, T. & Best, P.N. 2014, ARA&A, 52, 589
452
+ Hopkins, P., Richards, G., & Hernquist, L. 2007, ApJ, 654, 731
453
+ Janssen, R., Best, P.N., Röttgering, H., Brinchmann, J. 2012, A&A, 541, 62
454
+ Kakkad, D.,Sani, E., Rojas, A. et al. 2022, MNRAS, 511, 2105
455
+ Kennicutt, R., & Evans, N. 2012, ARA&A, 50, 531
456
+ Kim, C.-G., Ostriker, E., Somerville, R. et al. 2020, ApJ, 900, 61
457
+ Kondapally, R., Best, P.N., Cochrane, R.K. et al. 2022, MNRAS, 513, 3742
458
+ Lamperti, I., Pereira-Santaella, M., Perna, M., et al. 2022, A&A, 668, 45
459
+ Leitherer, C., Schaerer, D., Goldader, J. et al. 1999, ApJS, 123, 3
460
+ Lutz, D., Sturm, E., Janssen, A., et al. 2020, A&A, 633, 134
461
+ Madau, P., & Dickinson, M. 2014, ARA&A, 52, 415
462
+ McConnell, M.J., Ma, C.-P., 2013, ApJ, 764, 184
463
+ McGaugh, S. et al. 2000, ApJ, 533, L99
464
+ McNamara, B., & Nulsen, P. 2007, ARA&A, 45, 117
465
+ Miley, G. 1980, ARA&A, 18, 165
466
+ Miller, T, Arav, N., Xu, X. & Kriss, G. 2020, MNRAS, 499, 1522
467
+ Naab, T. & Ostriker, J. 2017, ARA&A, 55, 59
468
+ Rafferty, D., McNamara, B., Nulsen, P., & Wise M. 2006, ApJ, 652, 216
469
+ Revalski, M., Meena, B., Martinez, F. et al. 2021, ApJ, 910, 139
470
+
471
+ Sabater, J., Best, P., Hardcastle, M. et al. 2019, A&A, 622, 17
472
+ Schneider, E., Ostriker, E., Robertson, B., & Thompson, T. 2020, ApJ, 895, 43
473
+ Somerville, R. & Davé, R. 2015, ARA&A,
474
+ Tacconi, L., Genzel, R., & Sternberg, A. 2020, ARA&A, 58, 157
475
+ Thompson, T, & Heckman, T. 2023, ARA&A, in press
476
+ Tombesi, F., Cappi, M., Reeves, J., et al. 2011, ApJ, 742, 44
477
+ Tombesi, F., Cappi, M., Reeves, J., & Braito, 2012, MNRAS, 422, 1
478
+ Trindade Falcao, A., Kraemer, S., Fischer, T., et al. 2021, MNRAS, 500, 1491
479
+ Veilleux, S., Maiolino, R., Bolatto, A. & Aalto, S. 2020, A&Arv, 28, 2
480
+ Willott, C., Rawlings, S., Blundell, K., & Lacy M. 1999, MNRAS, 309, 1017
481
+ Yuan, Z., Wang, J., Zhou, M. et al. 2017, 846, 78
482
+
6NFKT4oBgHgl3EQf_C4s/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
8NFLT4oBgHgl3EQfBC4o/content/tmp_files/2301.11968v1.pdf.txt ADDED
@@ -0,0 +1,792 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ arXiv:2301.11968v1 [hep-th] 27 Jan 2023
2
+ Strong Cosmic Censorship in light of Weak Gravity
3
+ Conjecture for Charged Black Holes
4
+ Jafar Sadeghi ⋆1,
5
+ Mohammad Reza Alipour ⋆2,
6
+ Saeed Noori Gashti⋆3
7
+ ⋆Department of Physics, Faculty of Basic Sciences,
8
+ University of Mazandaran P. O. Box 47416-95447, Babolsar, Iran
9
+ Abstract
10
+ In this paper, we investigate the strong cosmic censorship conjecture (SCC) for charged
11
+ black holes in the de Sitter space by considering the weak gravity conjecture (WGC). Us-
12
+ ing analytical methods, we find that the SCC is preserved for dS-charged black holes with
13
+ respect to some restriction qQ ≫ 1 and r+ ≥ Q with the help of the WGC condition
14
+ viz
15
+ q
16
+ m ≥ 1 for scalar fields. Where q, m are the charge and mass of the scalar field, and
17
+ r+, Q determine the radius of the outer event horizon and the charge of the black hole,
18
+ respectively. In that case, when the (WGC) is valid, SCC will definitely be satisfied for
19
+ the dS-charged black holes. On the other hand, the SCC is violated when the WGC is
20
+ not satisfied. Also, we examined the RN-dS charged black hole in the extremality state
21
+ and found that SCC can be violated with the condition Λr2
22
+ + = 1.
23
+ Keywords: Strong cosmic censorship conjecture; Weak gravity conjecture; RN-dS charged
24
+ black hole
25
+ Contents
26
+ 1
27
+ Introduction
28
+ 2
29
+ 2
30
+ Weak Gravity Conjecture
31
+ 4
32
+ 3
33
+ Charged Black Holes in dS Space
34
+ 6
35
+ 4
36
+ The Quasinormal Resonant Frequency Spectrum
37
+ 7
38
+ 5
39
+ Discussion and Result
40
+ 10
41
+ 1Email:
42
43
+ 2Email:
44
45
+ 3Email:
46
47
+ 1
48
+
49
+ 1
50
+ Introduction
51
+ One of the studies with a long history in general relativity is the study of the collapse of small
52
+ perturbations. We need more information on how these oscillations decay to understand better
53
+ the gravity concept, use gravitational wave data, and study and investigate the valuable features
54
+ of general relativity. One of the signs of the failure of determinism in general relativity can
55
+ be the emergence of an interesting phenomenon known as Cauchy horizons that appear in the
56
+ astrophysical solutions of Einstein’s equations. These horizons are such that it is impossible
57
+ to specify the history of the future of an observer that passes through such horizons using
58
+ Einstein’s equations and initial data. With these descriptions in the black holes’ space-time
59
+ background, it is an expected possibility that the perturbations of the outer region are infinitely
60
+ amplified by a mechanism known as the blue shift. They lead to a singularity boundary beyond
61
+ the Cauchy horizon in the interior of black holes, where field equations cease to make sense. The
62
+ Penrose strong cosmic censorship (SCC) confirms such an expectation. Of course, another point
63
+ is that astrophysical black holes are stable due to a special mechanism called the perturbation-
64
+ damping mechanism, which is applied in the outer region.
65
+ Therefore, considering whether
66
+ SCC retains real hinges or not depends on the very subtle competition between the collapse
67
+ of perturbations in the outer region and their amplification (blue shift) in the inner space-
68
+ time of black holes. In general, the fate of Cauchy horizons is related to the collapse of small
69
+ perturbations outside the event horizon. Hence, the validity of SCC is tied to the extent of
70
+ external damps fluctuation. In connection with various structures and conditions, SCC and
71
+ its satisfaction and violation have been investigated in various theories. The violation of this
72
+ conjecture near the extremal region studied in the investigation of higher curvature gravity [1].
73
+ Also, this conjecture has been challenged in investigating many charged black holes. In [2,3], this
74
+ conjecture was checked for a charged AdS black hole. It was shown that for a specific interval
75
+ for the parameter (β), this conjecture is satisfied and violated in other areas as well.
76
+ The
77
+ strong cosmic censorship conjecture has also been investigated in two dimensions. There have
78
+ been interesting outcomes regarding the violation of this conjecture near the extremal region
79
+ at specific points [4]. The study of this conjecture in the structure of three-dimensional black
80
+ strings has also carried interesting results, which you can see [5] for a deeper study. Studying
81
+ the validity and violation of this conjecture in many recent studies in different conditions and
82
+ frameworks has led to exciting results that you can see [6–9] for further study. Therefore, in
83
+ this article, we want to study a different structure of this conjecture. According to the above
84
+ explanation, we consider the general configuration of charged black holes. Then, using the
85
+ weak gravity conjecture, we will prove that SCC is valid for specific values for all charged black
86
+ holes. We will use the weak gravity conjecture to prove a general relation for all charged black
87
+ holes about SCC. In connection with SCC, we need to pay attention to more concepts, which
88
+ we will mention here for further study. The effectiveness of mass-inflation systems, which are
89
+ involved in the transformations of the inner Cauchy horizon associated with the space-time of
90
+ 2
91
+
92
+ black holes that are approximately flat, which is pathological in the estimation of SCC, into
93
+ a series of hypersurfaces which is singular non-extendable. Those that are in an indivisible
94
+ form are related to two different types of physical mechanisms [10–16]. First, the events in
95
+ the exterior space-time regions of dynamic black holes formed viz the collapse of the remnant
96
+ perturbation fields and second amplification of exponential blue shift related to the fields falling
97
+ into the inside of black holes. We can manage these two introduced different systems through
98
+ parameters such as (g) and (k−). It can be stated that the dimensionless physical ratio with
99
+ the help of these two parameters can determine the fate of the inner Cauchy horizons inside
100
+ such space-times of non-asymptotic flat black holes [8,17,18],
101
+ β ≡ g
102
+ k−
103
+ .
104
+ Of course, a certain range of parameters of black holes, such as mass and charge, etc., as
105
+ indicated in [8,17,18],
106
+ β > 1
107
+ 2.
108
+ So, space-time of the corresponding black holes can be physically expanded beyond their Cauchy
109
+ horizon which includes a pathological fact and a sign of algebraic failure or a violation of the
110
+ Penrose SCC in classical general relativity. For the dynamics of Einstein’s equations as well
111
+ as the destiny of the observers, the explosive structure of the curvature that is related to
112
+ (β < 1) does not have per se much physical significance: it indicates two theorems, not the
113
+ failure of the field equations mentioned in [19] and of course not the destruction of macroscopic
114
+ observers which is discussed in [13]. Therefore, the physical and mathematical formulation
115
+ of the conjecture of a SCC in such conditions leads to ignoring physical phenomena such as
116
+ impulsive gravitational waves or the formation of shocks in relativistic fluids.
117
+ Due to the
118
+ aforementioned reasons, the modern form of the CC conjecture was introduced that requires
119
+ a stronger constraint (β < 1
120
+ 2 ) and many works have been done to fit such constraints. For
121
+ example, by studying massless scalar fields in linear form and examining the entire parametric
122
+ space of a charged black hole, areas beyond mentioned range were obtained, which it seems
123
+ cannot be allowed. According to the above explanations, we organize the article in the following
124
+ form.
125
+ In section 2, we will give basic explanations about the weak gravity conjecture and also the
126
+ motivation to use it. In section 3, we will introduce charged black holes in dS space, and then
127
+ we will introduce the quasinormal resonant frequency spectrum in section 4. We will check the
128
+ conditions of compatibility and violation of (SCC) with respect to (WGC) for RNdS charged
129
+ black holes. Finally, we describe the results in section 5.
130
+ 3
131
+
132
+ 2
133
+ Weak Gravity Conjecture
134
+ As it is known in the literature, a new idea has been put forward as a swampland program to
135
+ check theories coupled to gravity, to check the consistency of quantum gravity, and finally, a
136
+ proof for string theory. Recently, ones have done lots of work on this field [20–30]. Due to the
137
+ special conditions of string theory and the fact that its testing and experimental investigations
138
+ seem a bit difficult, this idea has been proposed to test and investigate various concepts of
139
+ cosmology. The swampland program is challenged from two sides. From an up-bottom view
140
+ for introducing principles and limitations to introduce conjectures, as well as mathematical
141
+ formulations to examine cosmological concepts. A second look from the bottom-up in order
142
+ to test each of these conjectures with various concepts of cosmology including inflation and
143
+ matching with observable data, which is both a proof for this new idea and a proof for string
144
+ theory. So far, many conjectures have been proposed from this theory, and now, according to
145
+ the structure and further investigations, new conjectures will be added to this program. Some
146
+ of these conjectures face challenges and as a result, corrections are made to the conjectures.
147
+ We face some limitations in quantum gravity (QG). At the point when gravity is considered at
148
+ the quantum level, the hypothesis will be incompatible. Generally having a reliable quantum
149
+ hypothesis of gravity isn’t really straightforward and can in any case hold many surprises and
150
+ can be interesting for physical science at low energies. The objective of the swampland program
151
+ is to decide the limitations that an effective field theory(EFT) should fulfill to be viable with the
152
+ consideration of ultraviolet completion(UV) in QG. They are called swampland limitations, and
153
+ different suggestions are figured out as far as swampland conjectures(SC). The objective is to
154
+ recognize these limitations, accumulate proof to demonstrate or refute them inside the structure
155
+ of QG, give reasoning to make sense of them in a model-free manner, and comprehend their
156
+ phenomenological suggestions for low-energy EFTs. Albeit the swampland idea isn’t restricted
157
+ to string theory on a fundamental level, SC are frequently examined by string theory backdrops.
158
+ Without a doubt, the string theory gives an ideal structure to thorough quantitative testing
159
+ of conjectures and works on how we might interpret potential compressions of string theory.
160
+ Strangely, it has as of late been uncovered that a large number of these conjectures are to be
161
+ sure related, recommending that they may essentially be various countenances of some yet-to-
162
+ be-found crucial standard of QG. As far as possible have significant ramifications for cosmology
163
+ and particle physics. They can give new core values to building conjectures past the standard
164
+ models in high-energy physics. They may likewise prompt UV/IR blending, which breaks the
165
+ assumption for scale detachment and possibly gives new bits of knowledge into the regular issues
166
+ seen in our universe. Consequently, the presence of swampland is extraordinary information
167
+ for phenomenology. For a total rundown of references connected with swampland that might
168
+ be valuable, we allude in [20] the swampland program (SP) has likewise been surveyed and
169
+ presented. The shortfall of global symmetry (GS) and the completeness of charge spectra are
170
+ at the center of the SP. Nonetheless, they need phenomenological suggestions except if we can
171
+ 4
172
+
173
+ restrict the global symmetries [21,22] and whether there is any limitations point on the mass
174
+ of charged states. In any case, they just bound the complete hypothesis but not the low-energy
175
+ EFTs. Specifically, it is phenomenologically important whether all charged particles can be
176
+ really super heavy and even compare to black holes(BHs), or whether there is some thought of
177
+ completeness of the range that gets by at low energies. A large portion of the SCs examined
178
+ address exactly these inquiries. They want to profoundly explore these assertions and measure
179
+ them so we can draw nearer to the recuperation of a few global symmetries. For instance, we
180
+ can deduce recuperate a global symmetry (GS) U(1) by sending the gauge coupling(GC) to
181
+ nothing, which ought not to be permitted in QG. Attempting to comprehend string theory
182
+ for the study of this issue, it might turn out that if one somehow managed to try to do such
183
+ work, can give data about the imperatives that an EFT can fulfill to be viable with QG.
184
+ Likewise, WGC forbids this cycle by flagging the presence of new charged states that denies
185
+ the depiction of the EFTs. Thusly, it gives an upper bound on the mass of these charged states.
186
+ The WGC comprises of some parts: the electric and the magnetic electric-WGC: As indicated
187
+ by a quantum hypothesis, we have the following condition [20–30],
188
+ Q
189
+ m ≥ Q
190
+ M |ext = O(1),
191
+ (1)
192
+ and
193
+ Q = gq,
194
+ (2)
195
+ where, g and q are the gauge coupling and the quantized charge. The electric-WGC needs
196
+ the presence of an electrically charged condition of a higher charge-to-mass proportion than
197
+ extremal BH in that hypothesis, which is regularly a variable of the order one. One more
198
+ understanding of this conjecture is that the limitations region shows that scale force determines
199
+ stronger than the gravity on this mode — so subsequently is called WGC. This is an identical
200
+ equation since it expects that electromagnetic force is stronger gravitational force [20–30],
201
+ FGrav ≤ FEM
202
+ (3)
203
+ It implies that the charge is more prominent than the mass, so we get a similar condition as
204
+ above. This is as of now false within the sight of massless scalar fields. The motivations of
205
+ WGC are twofold. To begin with, it makes a QG boundary to reestablish the GS of U(1) by
206
+ sending g → 0. If a GC goes to zero as indicated by WGC, this conducts new light particles
207
+ and the cutoff the hypothesis arrives at nothing and nullifies the EFT. Because of the littleness
208
+ of the scale coupling, it relies upon how much energy the interaction with which you need to
209
+ portray the viable EFT. The smallness of the cycle energy leads to the smallness of the scale
210
+ coupling. On the other hand, if you need to keep the EFT substantial up to an extremely high
211
+ cut-off, the GC can’t be excessively small. This is an illustration of swampland limitations that
212
+ 5
213
+
214
+ becomes more grounded for higher energies. Obviously, a hypothesis with disappearing measure
215
+ coupling i.e., GS is incompatible because the cutoff of the viable EFT is likewise zero. One more
216
+ fundamental inspiration for WGC is that a kinematic prerequisite permits extremality BH to
217
+ have decomposed. Charged BHs should fulfill an extremality breaking point to stay away from
218
+ the presence of exposed singularities, as shown by the weak cosmic censorship (WCC). For a
219
+ given charge Q, this super bound shows that the this super bound shows that the mass M of
220
+ the BHs should be more noteworthy than the charge [20–30],
221
+ M ≥ Q
222
+ (4)
223
+ For the BHs to have a regular horizon.
224
+ Here, we set the extremal factor O(1) to one for
225
+ simplicity. The primary condition for starting the decay to the small black hole and particle
226
+ is the existence of the extremal BHs (M = Q). So, one can consider the decay of an extremal
227
+ BHs which one of the rot items has a charge more modest than its mass as far as possible, so
228
+ M1 ≥ Q1. Then the rot item can never again have a charge more modest than the mass, that
229
+ is m2 ≤ Q2. It is just a kinematic necessity. Since the second rot item violates the WCC, it
230
+ can’t be a BH, so it should be a particle. The above kinematic necessity can be acquired by
231
+ applying preservation of mass/energy and protection of charge as follows. The initial mass of
232
+ the BH should be more prominent than the amount of the mass of the rot items Mi and the
233
+ charge of the initial BH.
234
+ 3
235
+ Charged Black Holes in dS Space
236
+ The metric of charged black hole in spherical symmetric space is defined as follows,
237
+ dS2 = f(r)dt2 − f −1(r)dr2 − r2dΩ2,
238
+ dΩ2 = (dθ2 + sin2(θ)dϕ2).
239
+ (5)
240
+ Here, we consider f(r) = H(M, Q) − Λr2
241
+ 3
242
+ in general; where Q, M, Λ > 0 are electric charge, the
243
+ mass of the black hole and the cosmological constant respectively. In this case, we can obtain
244
+ its event horizons as follows,
245
+ f(r⋆) = 0
246
+
247
+ ⋆ ∈ (−, +, ..., c).
248
+ (6)
249
+ Considering the metric in general terms, we have different event horizons, where (r−) is the
250
+ Cauchy horizon, (r+) is the outer event horizon, and (rc) is the cosmological horizons. Using
251
+ Klein-Gordon’s differential equation, we can determine the dynamics of a massive charged
252
+ particle near a charged black hole [31–34],
253
+ 1
254
+ √−g∂µ(gµν√−g∂νΦ) − 2iqgµνAµ∂νΦ − q2gµνAµAνΦ − m2Φ = 0,
255
+ (7)
256
+ 6
257
+
258
+ where m and q are the mass and charge of the particle, respectively also, Aµ =
259
+ � Q
260
+ r , 0, 0, 0
261
+
262
+ . We
263
+ can define the scalar field Φ according to relation (7) as follows [36],
264
+ Φ(t, r, θ, φ) =
265
+
266
+ m
267
+
268
+
269
+ e−iωtYℓm(θ, ϕ)Φ(r).
270
+ (8)
271
+ The integer parameters ℓ and m are the spherical and the azimuthal harmonic indices of the
272
+ resonant eigenmodes which characterize the charged massive scalar fields in the charged black-
273
+ hole spacetime. By putting Eq.(8) in Eq.(7) and using dx =
274
+ dr
275
+ f(r), we get the Schr¨odinger-like
276
+ differential equation ,
277
+ d2φ(r)
278
+ dx2
279
+ + V (r)φ(r) = 0.
280
+ (9)
281
+ The effective radial potential due to a massive charged particle near a charged black hole is
282
+ defined as [8],
283
+ V (r) = qm
284
+ r2
285
+ � q
286
+ mα(r) − m
287
+ q β(r)
288
+
289
+ ,
290
+ (10)
291
+ where
292
+ α(r) = Q2
293
+
294
+ 1 − ωr
295
+ qQ
296
+ �2
297
+ ,
298
+ β(r) = r2f(r)H(r),
299
+ H(r) =
300
+ �ℓ(ℓ + 1)
301
+ m2r2
302
+ + f ′(r)
303
+ m2r + 1
304
+
305
+ .
306
+ (11)
307
+ Also, we can consider the boundary conditions for the special radial function near the outer
308
+ event horizon as an incoming wave and at the largest event horizon as an outgoing wave [34,35]:
309
+ φ(x) ∼
310
+
311
+ e
312
+ −i(ω− qQ
313
+ r+ )x,
314
+ for
315
+ r → r+ (x → −∞);
316
+ e−i(ω− qQ
317
+ rc )x,
318
+ for
319
+ r → rc (x → ∞).
320
+ (12)
321
+ According to the above boundary conditions, we can obtain the discrete spectrum of ω, defined
322
+ as the resonance frequency of the imaginary quasi-normal state.
323
+ 4
324
+ The Quasinormal Resonant Frequency Spectrum
325
+ In this section, we need to obtain the imaginary part of the resonance frequency to investigate
326
+ the linear dynamics of a massive charged particle near a general charged black hole. Also, we
327
+ need to do this in a dimensionless regime to do this analytically. Since, the q2
328
+ ¯h ≃
329
+ 1
330
+ 137 relationship
331
+ exists in our universe, we can consider it for black holes, even slightly charged, and get qQ ≫ 1.
332
+ In addition, the mechanism of the Schwinger-type pair-production in space-time of charged
333
+ black hole creates a limit to the black hole electric field with the
334
+ Q
335
+ r2
336
+ + ≪ m2
337
+ q relationship [37–40].
338
+ 7
339
+
340
+ Therefore, according to the above statement, we can consider SCC and define our constraint
341
+ regime following ansans,
342
+ m2r2
343
+ + ≫ ℓ(ℓ + 1)
344
+ and
345
+ m2r2
346
+ + ≫ 2k+r+,
347
+ (13)
348
+ where k+ = f ′(r+)/2 is the gravitational acceleration of the black hole at the outer event
349
+ horizon. In this area, we try to obtain the imaginary part of the resonance frequency in the
350
+ background of the general charged black hole near the event horizon.
351
+ Now, we use radial
352
+ potential (10) to determine the linear dynamics of the massive charged particle near the event
353
+ horizon of the black hole. We can consider this potential in region (13) as an effective potential
354
+ and obtain the quasinormal resonance modes analytically using standard WKB techniques
355
+ [41, 42]. In this region, we consider the maximum effective potential near the event horizon
356
+ of the black hole at point r = r0. In the following, we use the relationship (10), (11), and
357
+ V ′(r0) = 0 to obtain the point where the effective potential is maximum as follows,
358
+ r0 =
359
+ q2Q2
360
+ qQω − m2r2
361
+ +k+
362
+ (14)
363
+ According to the Schr¨odinger-like differential equation (9) and [41–43], we use the WKB
364
+ method to obtain the quasinormal mode frequencies through the following,
365
+ iK − (n + 1
366
+ 2) − Λ(n) = Ω(n)
367
+ (15)
368
+ where
369
+ K =
370
+ V0
371
+
372
+ 2V (2)
373
+ 0
374
+ Λ(n) =
375
+ 1
376
+
377
+ 2V (2)
378
+ 0
379
+
380
+
381
+
382
+ α2 + 1
383
+ 4
384
+
385
+ 8
386
+ V (4)
387
+ 0
388
+ V (2)
389
+ 0
390
+ − (60α2 + 7)
391
+ 288
392
+
393
+ V (3)
394
+ 0
395
+ V (2)
396
+ 0
397
+ �2
398
+
399
+ Ω(n) = n + 1
400
+ 2
401
+ 2V (2)
402
+ 0
403
+
404
+ 5 (188α2 + 77)
405
+ 6912
406
+
407
+ V (3)
408
+ 0
409
+ V (2)
410
+ 0
411
+ �4
412
+ − (100α2 + 51)
413
+ 384
414
+
415
+ V (3)
416
+ 0
417
+ �2
418
+ V (4)
419
+ 0
420
+
421
+ V (2)
422
+ 0
423
+ �3
424
+
425
+ 
426
+ + n + 1
427
+ 2
428
+ 2V (2)
429
+ 0
430
+
431
+ (68α2 + 67)
432
+ 2304
433
+
434
+ V (4)
435
+ 0
436
+ V (2)
437
+ 0
438
+ �2
439
+ + (28α2 + 19)
440
+ 288
441
+
442
+ V (3)
443
+ 0
444
+ V (5)
445
+ 0
446
+
447
+
448
+ V (2)
449
+ 0
450
+ �2
451
+ − (4α2 + 5)
452
+ 288
453
+ V (6)
454
+ 0
455
+ V (2)
456
+ 0
457
+
458
+ 
459
+ (16)
460
+ Here, V (k)
461
+ 0
462
+ ≡ dkV
463
+ dxk |r=r0 is the spatial derivative of the effective potential of equation (10), and
464
+ its scattering peak is evaluated at the point r = r0. Using relations (10), (11), (14) and (16),
465
+ 8
466
+
467
+ we will have the following relation in the region of (13),
468
+ K ≃
469
+ k2
470
+ +m4r4
471
+ +qQ
472
+ 2f0 (k+m2r2
473
+ + − qQω)2
474
+ Λ(n) ≃
475
+ k2
476
+ +m4 �
477
+ 17 − 60
478
+
479
+ n + 1
480
+ 2
481
+ �2�
482
+ r4
483
+ + + 2k+m2 �
484
+ 36
485
+
486
+ n + 1
487
+ 2
488
+ �2 − 7
489
+
490
+ qQr2
491
+
492
+ 16qQ (qQω − 3k+m2r2
493
+ +)2
494
+ × f0
495
+ A = 15k4
496
+ +m8
497
+
498
+ 148(n + 1
499
+ 2)2 − 41
500
+
501
+ r8
502
+ + + 12k3
503
+ +m6
504
+
505
+ 121 − 420(n + 1
506
+ 2)2
507
+
508
+ qQr6
509
+
510
+ B = 64q5Q5 �
511
+ k+m2r2
512
+ + − qQω
513
+ �4
514
+ Ω(n) ≃ −(n + 1
515
+ 2)q3Q3f 2
516
+ 0 × A
517
+ B
518
+ (17)
519
+ where f0 = f(r0). In the next step, to determine the study of SCC, we need to obtain the
520
+ minimum value of the fundamental imaginary resonance mode of the system. For this purpose,
521
+ using equations (15) and (17), we can calculate the Im(ω0),
522
+ ω ≃ qQ
523
+ r+
524
+ − 2k+m2r2
525
+ +
526
+ qQ
527
+
528
+ 1 − 14400
529
+ 11644
530
+ �(n + 1/2)f0
531
+ qQ
532
+ �4�
533
+ − i
534
+
535
+ 4f0k+(n + 1
536
+ 2)m2r2
537
+ +
538
+ q2Q2
539
+
540
+ 1 − 34qQf 4
541
+ 0
542
+ 11664
543
+
544
+ + O(f 2
545
+ 0)
546
+
547
+ (18)
548
+ Since we consider r0 near the event horizon (r+), we have f0 ≪ 1. For investigation the SCC,
549
+ it is necessary to find the minimum value of the resonance mode and evaluate its ratio to the
550
+ surface gravity of the event horizon,
551
+ β = −Im(ω0)
552
+ k+
553
+ ≃ 2f0
554
+ m2r2
555
+ +
556
+ q2Q2
557
+
558
+ 1 − 34qQf 4
559
+ 0
560
+ 11664
561
+
562
+ .
563
+ (19)
564
+ Since it is f0 ≪ 1, it is sufficient to have the conditions q2Q2 > m2r2
565
+ + in the relation above
566
+ concepts so that −Im(ω0)
567
+ k+
568
+ < 1
569
+ 2 is established. Therefore, we have the following condition for the
570
+ study of SCC,
571
+ q
572
+ m ≥ r+
573
+ Q .
574
+ (20)
575
+ from equation (20) determine that when r+ ≥ Q, we have the weak gravity conjecture condition.
576
+ We know that k− > k+, so the relationship of (19) and (20) is also established for β = −Im(ω0)
577
+ k−
578
+ <
579
+ 1
580
+ 2. Also, according to relation (19), when qQ < 2√f0mr+, SCC can be violated. Since qQ ≫ 1
581
+ and f0 ≪ 1, the mass of the scalar field and the radius of the event horizon must be very
582
+ massive and very large respectively. In the following, we obtain the extremality state of the
583
+ 9
584
+
585
+ RN-dS black hole. We will have the following relation for the RN-dS black hole with respect
586
+ to equation(5),
587
+ f(r) = 1 − 2M
588
+ r
589
+ + Q2
590
+ r2 − Λr2
591
+ 3 .
592
+ (21)
593
+ When k+ = k− = 0, we can obtain the black hole extremality state,
594
+ Qexe = r+
595
+
596
+ 1 − Λr2
597
+ +,
598
+ Mexe = r+(1 − 2
599
+ 3Λr2
600
+ +).
601
+ (22)
602
+ We substitute Eq.(22) in Eq.(19) to obtain β in the extremality state of the RN black hole,
603
+ β ≃ 2f0
604
+ m2
605
+ q2(1 − Λr2
606
+ +)
607
+
608
+ 1 − 34qf 4
609
+ 0
610
+
611
+ 1 − Λr2
612
+ +r+
613
+ 11664
614
+
615
+ (23)
616
+ According the above relationship, when the condition
617
+ q
618
+ m ≥
619
+ 1
620
+ 1−Λr2
621
+ + is satisfied, the SCC will
622
+ definitely be preserved, and since Λr2
623
+ + < 1, the weak gravity conjecture will also be satisfied.
624
+ On the other hand, when Λr2
625
+ + ≪ 1, we will have the SCC condition in light of the WGC,
626
+ q
627
+ m ≥ 1 + Λr2
628
+ +,
629
+ (24)
630
+ from the above relation WGC is clearly obtained. In relation (23) when Λr2
631
+ + = 1, we have
632
+ β → ∞ and the SCC is violated. Also, these result and conditions are completely compatible
633
+ with [44,45].
634
+ 5
635
+ Discussion and Result
636
+ One of the indications of the failure of determinism GR can be the rise of a fascinating pecu-
637
+ liarity known as the Cauchy horizon that shows up in the astrophysical solutions of Einstein’s
638
+ equations. These horizons are such that it is difficult to indicate the history of the future of
639
+ an observer that passes come of such horizons utilizing Einstein’s conditions and initial infor-
640
+ mation. With these descriptions in the black holes’ background space-time, it is a predicted
641
+ possibility that the perturbations of the external area are infinitely enhanced by a system
642
+ known as the blue shift. They lead to a singularity beyond the Cauchy horizon the inside
643
+ of BHs, where field conditions fail to seem good. The Penrose cosmic censorship conjecture
644
+ (SCC) affirms such an assumption. Obviously, another point is that astrophysical BHs are
645
+ stable because of an exceptional component called the perturbation-damping system, which is
646
+ applied in the outer region. Also, the SCC resolves the issue of the idea of the singularities
647
+ tracked down in many answers to Einstein’s gravitational field equations: Are such singular-
648
+ ities conventionally described by unbounded curvature? Is the presence of a Cauchy horizon,
649
+ 10
650
+
651
+ an unsteady characteristic element of answers of Einstein’s equations? Recently researchers,
652
+ remarking on the historical backdrop of the SCC conjecture, overview a portion of the headway
653
+ made in research coordinated either toward satisfying SCC or toward revealing a portion of its
654
+ shortcomings. They specifically around model adaptations of SCC which have been demon-
655
+ strated for constrained groups of spacetimes viz the Gowdy spacetimes and the role played by
656
+ the conventional presence of Asymptotically speed term dominated conduct in these answers.
657
+ Also additionally note some work on spacetimes containing weak null singularities, and their
658
+ importance for the SCC [44, 45, 47]. SCC conjecture has been one of the main acts of pure
659
+ confidence with regard to GR, confirming the deterministic idea of the related field relations.
660
+ However, it holds well for asymptotically level spacetimes, an expected disappointment of the
661
+ SCC conjecture could emerge for spacetimes acquiring Cauchy horizon alongside a positive
662
+ cosmological constant viz its potential failure about this issue. Researchers have unequivocally
663
+ exhibited that infringement of the restriction SCC turns out as expected within the sight of
664
+ a Maxwell field even with the presence of higher spacetime aspects. Specifically, for higher
665
+ dimensions of the RN black holes, the infringement of SCC is at a bigger scope compared with
666
+ the 4D case, for specific of the cosmological constant. Then again, for a brane world BH, the
667
+ impact of an additional dimension is to make the infringement of cosmic censorship weaker.
668
+ For rotating BHs, intriguingly, the SCC is constantly holding even in the presence of higher
669
+ dimensions. A comparable situation is likewise noticed for rotating BHs on the brane [47].
670
+ In this paper, we investigated dynamically formed charged black holes. Also, to satisfy the
671
+ SCC, the inner Cauchy horizons of the black hole must be unstable. Here, to check the SCC,
672
+ it is necessary to get two −Im(ω0) and k− parameters to demonstrate the decay rate of the
673
+ remaining perturbation fields in the outer regions of the black hole and the blue-shift growth
674
+ rate of the in-falling fields of the black hole, respectively. Therefore, if β = −Im(ω0)
675
+ k−
676
+ < 1/2, SCC
677
+ will be maintained. We found that for the dS charged black hole with respect to r+ ≥ Q in
678
+ light of the WGC, viz q/m ≥ 1, SCC will definitely be satisfied. We also found that there will
679
+ be a possibility of violation of SCC for the massive scalar field as well as when the radius of the
680
+ event horizon of the charged black hole is very large. We also found SCC will be violated in
681
+ the extremality state for the charged RN-dS black hole when Λr2
682
+ + = 1 which is also mentioned
683
+ in [44,45]. Also, these results and conditions are completely compatible with [44,45]. On the
684
+ other hand, in [8,46], when the scalar field is uncharged, the SCC is violated, which is consistent
685
+ with (19) in this paper. Because can be obtained β > 1/2 if assume the charge of the scalar
686
+ field is zero viz q = 0. The above study also raises some questions as follows.
687
+ Is the relationship researched in this article also valid for black holes in higher dimensions? Do
688
+ other black holes in different frames satisfy the SCC and WGC simultaneously? Is it possible to
689
+ consider the SCC relation with WGC monitoring for all black holes? Is it may such a structure
690
+ also be established for black holes on the brane? We leave these questions for future work.
691
+ 11
692
+
693
+ References
694
+ [1] A. K. Mishra and S. Chakraborty, strong cosmic censorship conjecture in higher curvature
695
+ gravity, Phys. Rev. D 101, 064041 (2020).
696
+ [2] C. Singha, S. Chakraborty and N. Dadhich, Strong cosmic censorship conjecture for a
697
+ charged BTZ black hole, J. High Energ. Phys. 2022, 28 (2022).
698
+ [3] C. Singha and N. Dadhich, Strong cosmic censorship conjecture for a charged AdS black
699
+ hole, arXiv:2209.10153 (2022).
700
+ [4] U. Moitra, Strong cosmic censorship in two dimensions, Phys. Rev. D 103, L081502 (2021).
701
+ [5] J. Ho, W. Kim, W. Kim and B. H. Lee, Investigations of strong cosmic censorship in
702
+ 3-dimensional black strings, J. High Energ. Phys. 2022, 18 (2022).
703
+ [6] C. Y. Shaoa, L. J. Xina, W. Zhangb, C. G. Shaoa, Strong cosmic censorship for a charged
704
+ black hole surrounded by quintessence, Physics Letters B 835, 137512 (2022).
705
+ [7] M. Casals and C. I. S. Marinho, Glimpses of violation of strong cosmic censorship in
706
+ rotating black holes, Phys. Rev. D 106, 044060 (2022).
707
+ [8] V. Cardoso, J. L. Costa, K. Destounis, P. Hintz and A. Jansen, Quasinormal Modes and
708
+ Strong Cosmic Censorship, Phys. Rev. Lett. 120, 031103 (2018).
709
+ [9] R. A. Konoplya and A. Zhidenko, How general is the strong cosmic censorship bound for
710
+ quasinormal modes?, Journal of Cosmology and Astroparticle Physics 2022, (2022).
711
+ [10] P.R. Brady, I.G. Moss and R.C. Myers, Cosmic Censorship: As Strong As Ever, Phys.
712
+ Rev. Lett. 80, 3432 (1998).
713
+ [11] W.A. Hiscock, Evolution of the interior of a charged black hole, Phys. Lett. A 83, 110
714
+ (1981).
715
+ [12] E. Poisson and W. Israel, Internal structure of black holes, Phys. Rev. D 41, 1796 (1990).
716
+ [13] A. Ori, Inner structure of a charged black hole: An exact mass-inflation solution, Phys.
717
+ Rev. Lett. 67, 789 (1991).
718
+ [14] M. L. Gnedin and N. Y. Gnedin, Destruction of the Cauchy horizon in the Reissner-
719
+ Nordstrom black hole, Class. Quantum Gravity 10, 1083 (1993).
720
+ [15] P.R. Brady and J.D. Smith, Black Hole Singularities: A Numerical Approach, Phys. Rev.
721
+ Lett. 75, 1256 (1995).
722
+ 12
723
+
724
+ [16] S. Hod and T. Piran, Mass Inflation in Dynamical Gravitational Collapse of a Charged
725
+ Scalar Field, Phys. Rev. Lett. 81, 1554 (1998).
726
+ [17] M. Dafermos, Black Holes Without Spacelike Singularities, Commun. Math. Phys. 332,
727
+ 729 (2014).
728
+ [18] P. Hintz and A. Vasy, Analysis of linear waves near the Cauchy horizon of cosmological
729
+ black holes, J. Math. Phys. 58, 081509 (2017).
730
+ [19] S. Klainerman, I. Rodnianski, and J. Szeftel, The bounded L2 curvature conjecture, Invent.
731
+ Math. 202, 91 (2015).
732
+ [20] E. Palti, The Swampland: Introduction and Review, Fortsch. Phys., 67, 1900037 (2019).
733
+ [21] S. Fichet and P. Saraswat, Approximate Symmetries and Gravity, JHEP. 01, 088 (2020).
734
+ [22] T. Daus, A. Hebecker, S. Leonhardt and J. March-Russell, Towards a Swampland Global
735
+ Symmetry conjcture using weak gravity, Nucl. Phys. B. 960, 115167 (2020).
736
+ [23] M. van Beest, J. Calder´on-Infante, D. Mirfendereski and I. Valenzuela, Lectures on the
737
+ swampland program in string compactifications. Physics Reports 989, 1-50 (2022).
738
+ [24] E. Palti, A Brief Introduction to the Weak Gravity Conjecture, LHEP 2020, 176 (2020).
739
+ [25] C. Vafa, The String landscape and the swampland, hep-th/0509212.
740
+ [26] N. Arkani-Hamed, L. Motl, A. Nicolis and C. Vafa, The String landscape, black holes and
741
+ gravity as the weakest force, JHEP 06, 060 (2007).
742
+ [27] J. Sadeghi, M. Shokri, S. N. Gashti and M. R. Alipour, RPS thermodynamics of Taub–NUT
743
+ AdS black holes in the presence of central charge and the weak gravity conjecture. Gen
744
+ Relativ Gravit 54, 129 (2022).
745
+ [28] J. Sadeghi, M. Shokri, M. R. Alipour and S. N. Gashti, Weak gravity conjecture from
746
+ conformal field theory: a challenge from hyperscaling violating and Kerr-Newman-AdS
747
+ black holes, Chinese Physics C 47, 015103 (2023).
748
+ [29] J. Sadeghi, S. N. Gashti and E N. Mezerji, The investigation of universal relation between
749
+ corrections to entropy and extremality bounds with verification WGC, Physics of the Dark
750
+ Universe 30, 100626 (2020).
751
+ [30] J. Sadeghi, B. Pourhassan, S. N. Gashti and S. Upadhyay, Weak Gravity Conjecture,
752
+ Black Branes and Violations of Universal Thermodynamic Relation, Annals of Physics
753
+ 447, 169168 (2022).
754
+ 13
755
+
756
+ [31] R. A. Konoplya and A. Zhidenko, Massive charged scalar field in the Kerr-Newman back-
757
+ ground: Quasinormal modes, late-time tails and stability, Phys. Rev. D 88, 024054 (2013).
758
+ [32] S. Hod, Mode-Coupling in Rotating Gravitational Collapse of a Scalar Field, Phys. Rev.
759
+ D 61, 024033 (2000).
760
+ [33] E. N. Mezerji, J. Sadeghi, The correlation of WGC and hydrodynamics bound with R4
761
+ correction in the charged AdSd+ 2 black brane, Nuclear Physics B 981, 115858 (2022).
762
+ [34] S. Hod, Strong cosmic censorship in charged black-hole spacetimes: As strong as ever,
763
+ Nuclear PhysicsB 941, 636 (2019).
764
+ [35] R.A. Konoplya, A. Zhidenko, Charged scalar field instability between the event and cos-
765
+ mological horizons, Phys. Rev. D 90, 064048 (2014).
766
+ [36] R. A. Konoplya and A. Zhidenko, Quasinormal modes of black holes: From astrophysics
767
+ to string theory, Rev. Mod. Phys. 83, 793 (2011).
768
+ [37] J. Schwinger, On Gauge Invariance and Vacuum Polarization, Phys. Rev. 82, 664 (1951).
769
+ [38] B. Carter, Charge and Particle Conservation in Black-Hole Decay, Phys. Rev. Lett. 33,
770
+ 558 (1974).
771
+ [39] S. Hod, Best approximation to a reversible process in black-hole physics and the area
772
+ spectrum of spherical black holes, Phys. Rev. D 59, 024014 (1999).
773
+ [40] W. T. Zaumen, Upper bound on the electric charge of a black hole, Nature 247, 531 (1974).
774
+ [41] S. Iyer and C.M. Will, Black-hole normal modes: A WKB approach. I. Foundations and
775
+ application of a higher-order WKB analysis of potential-barrier scattering, Phys. Rev. D
776
+ 35, 3621 (1987).
777
+ [42] S. Iyer, Black-hole normal modes: A WKB approach. II. Schwarzschild black holes, Phys.
778
+ Rev. D 35, 3632 (1987).
779
+ [43] B. F. Schutz and C. M. Will, Black hole normal modes - A semianalytic approach, Astro-
780
+ phys. J. 291, L33 (1985).
781
+ [44] O. J. C. Dias, H. S. Reall, and J. E. Santos, Strong cosmic censorship for charged de Sitter
782
+ black holes with a charged scalar field, Class. Quantum Grav. 36, 045005 (2019).
783
+ [45] O. J. C. Dias, H. S. Reall, and J. E. Santos, Strong cosmic censorship: taking the rough
784
+ with the smooth, J. High Energ. Phys. 2018, 1 (2018).
785
+ 14
786
+
787
+ [46] V. Cardoso, J. L. Costa, K. Destounis, P. Hintz and A. Jansen, Strong cosmic censorship
788
+ in charged black-hole spacetimes: Still subtle, Phys. Rev. D 98, 104007 (2018).
789
+ [47] M. Rahman, S. Chakraborty, S. S. Guptab and A. A. Sen, Fate of strong cosmic censorship
790
+ conjecture in presence of higher spacetime dimensions, JHEP 03, 178 (2019).
791
+ 15
792
+
8NFLT4oBgHgl3EQfBC4o/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
9dE3T4oBgHgl3EQfrApq/content/tmp_files/2301.04656v1.pdf.txt ADDED
The diff for this file is too large to render. See raw diff
 
9dE3T4oBgHgl3EQfrApq/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
9tFQT4oBgHgl3EQfJjU4/content/2301.13256v1.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:914f886ee0a2c70d77202ab44913996ef5770141a6ae4bafa7afbd53808cf018
3
+ size 12367653
C9AzT4oBgHgl3EQfwf5z/content/tmp_files/2301.01723v1.pdf.txt ADDED
@@ -0,0 +1,846 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Instability of the cosmological DBI-Galileon in the
2
+ non-relativistic limit
3
+ C. Leloup1,2, L. Heitz3 and J. Neveu3,4
4
+ 1 Universit´e Paris-Cit´e, CNRS, Astroparticule et Cosmologie, 75013 Paris, France
5
+ 2 Kavli Institute for the Physics and Mathematics of the Universe (Kavli IPMU,
6
+ WPI), UTIAS, The University of Tokyo, Kashiwa, Chiba 277-8583, Japan
7
+ 3 Universit´e Paris-Saclay, CNRS, IJCLab, 91405, Orsay, France
8
+ 4 Sorbonne Universit´e, CNRS, Universit´e de Paris, LPNHE, 75252 Paris Cedex 05,
9
+ France
10
+ Abstract.
11
+ The DBI-Galileon model is a tensor-scalar theory of gravity which finds its
12
+ foundation as the most general theory of the dynamics of a 4D brane embedded in
13
+ a 5D bulk.
14
+ It is of particular interest as it provides a few free parameters with a
15
+ physical meaning, such as the cosmological constant which is there related to the
16
+ brane tension. Most studies of this model have been performed assuming a maximally
17
+ symmetric geometry for the 5D bulk, in which it has been shown that the theory
18
+ reduces to various types of Galileon. In contrast, the general case for the geometry of
19
+ the bulk provides a different covariantization of the Galileon model than the covariant
20
+ Galileon: the DBI-Galileon. From the tight constraints on the gravitational waves
21
+ speed, we are naturally led to consider the non-relativistic limit of the model where
22
+ the kinetic energy of the brane is small compared to its tension, that we study in
23
+ the context of late-time cosmology.
24
+ The DBI-Galileon in the non-relativistic limit
25
+ is simply an expansion around General Relativity (GR) which can be expressed as
26
+ a shift-symmetric Horndeski theory. We developed the description of this theory at
27
+ the background and perturbation level. However, by studying the scalar and tensor
28
+ perturbations around a flat FLRW background, we found that they contain a ghost
29
+ degree of freedom leading to fatal instability of the vacuum for every combination of the
30
+ free parameters. As a lesson, we emphasized which of the Horndeski terms competes
31
+ to avoid this instability in more general cases.
32
+ 1. Introduction
33
+ Dark energy has been modelled by a large variety of theories since decades. Among
34
+ these, many rely on the introduction of additional scalar fields whose dynamics, at the
35
+ origin of the late-time acceleration of the expansion of the Universe, is determined by
36
+ arbitrary parametric functions, potentials and/or coupling (see e.g. [1]). These are the
37
+ so-called scalar-tensor theories of modified gravity. In particular, the class of Horndeski
38
+ theories is of great interest as it contains all models of modified gravity with a single
39
+ arXiv:2301.01723v1 [hep-th] 4 Jan 2023
40
+
41
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
42
+ 2
43
+ additional scalar field leading to second-order equations of motion [2, 3]. Extensions of
44
+ Horndeski theories to scalar-tensor theories of one scalar field with equations of motion of
45
+ higher orders have also been explored [4, 5]. Particular Horndeski theories are described
46
+ by the specification of four arbitrary functions of the scalar field and its kinetic energy,
47
+ leading to a huge variety of models and phenomenological behaviours.
48
+ Among these wide classes of models, some can be built from first physical principles
49
+ or arguments of symmetry.
50
+ For instance, the Galileon model [6] and its covariant
51
+ extension [7] was built by imposing a galilean symmetry for the scalar field, leaving
52
+ only five free numerical parameters. We can also cite, among many others, the pure
53
+ kinetic gravity theory [8], massive gravity in the non-relativistic limit [9, 10] and the
54
+ DBI-Galileon [11] which is the main object of this paper.
55
+ The DBI-Galileon is a model that falls into the class of Brane-world scenarios of
56
+ extra-dimension theories, where the matter fields are confined on a 4D brane while
57
+ gravity can propagate into the additional spatial dimensions. Of most interest for the
58
+ DBI-Galileon is the case of a single extra-dimension as it has been shown that theories
59
+ with more co-dimensions exhibit ghosts either in the flat or self-accelerating de Sitter
60
+ solution [12]. The action include a volume term for the 4D brane in the 5D bulk which
61
+ leads to the well-known Dirac-Born-Infeld (DBI) action.
62
+ This action, and DBI-like
63
+ extensions, can lead to a self-accelerating solution and has been thoroughly studied as
64
+ a candidate model in the early Universe cosmic inflation paradigm [13, 14]. In addition,
65
+ the DBI-Galileon model exhibits the Galileon Lagrangians in the non-relativistic limit
66
+ [11] but giving a physical meaning to their free parameters: the Planck mass in the
67
+ brane, the Planck mass in the bulk, etc. In particular, the brane tension here plays the
68
+ role of the cosmological constant which brings a possible interpretation of its nature.
69
+ The original probe brane construction has been revisited in [15] where the matter
70
+ metric is disformally related to a standard gravitational metric, or in [16] in the
71
+ framework of spontaneous symmetry breaking for the 5D space-time symmetries broken
72
+ by the presence of the brane, bridging the gap with Brane-world scenarios developed
73
+ in the context of quantum field theory and an interpretation of the scalar field as a
74
+ Nambu-Goldstone boson [17, 18]. The DBI-Galileon model has been studied extensively
75
+ in special cases of the maximally symmetric bulk geometry [19, 20]. However, to our
76
+ knowledge, no study of the DBI-Galileon in the late-time cosmology setting as a potential
77
+ candidate for Dark Energy has been performed so far.
78
+ In this paper we develop the DBI-Galileon theory in the non relativistic limit
79
+ (Section 2) and study its dynamics in the Friedmann-Lemaˆıtre-Robertson-Walker flat
80
+ metric (Section 3). The perturbation stability is explored in Section 4 and then discussed
81
+ in Section 5.
82
+ 2. DBI-Galileon in the non-relativistic limit
83
+ DBI-Galileon
84
+ We are interested in the description of a four dimensional brane universe embedded in
85
+
86
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
87
+ 3
88
+ a five dimensional bulk from the cosmological perspective. In this context, it has been
89
+ shown in [11] that the most general action on the brane is given by the 4D Lovelock
90
+ terms [21] inside the brane and the boundary terms associated to the 5D Lovelock terms
91
+ in the bulk:
92
+ S =
93
+
94
+ dx4√−g
95
+
96
+ −Λ − M 3
97
+ 5K + M 2
98
+ P
99
+ 2 R − β M 3
100
+ 5
101
+ m2 KGB + Lm (˜qµν, ψm)
102
+
103
+ ,
104
+ (1)
105
+ where g is the induced metric on the brane, Λ the brane cosmological constant, K is the
106
+ extrinsic curvature of the brane, R is the Ricci scalar on the brane, KGB is the boundary
107
+ term on the brane of the Gauss-Bonnet scalar in the bulk and Lm is the Lagrangian
108
+ density of matter that lives confined in the brane. In the above action has been defined
109
+ the 4D Planck mass MP, its 5D counterpart M5 and their ratio m = M 3
110
+ 5/M 2
111
+ P, and β is
112
+ an arbitrary constant parameter.
113
+ Because the action is defined on the 4-dimensional brane, the above quantities are
114
+ expressed in terms of the induced metric g, as opposed to the 5-dimensional bulk metric
115
+ G. Assuming we have a coordinate system (xµ, y) in the bulk, Greek letters being defined
116
+ to span from 0 to 3, such that the length element in this frame is ds2 = qµνdxµdxν+(dy)2.
117
+ The brane position in the bulk is defined by y = π (xµ) such that we can express the
118
+ induced metric from the bulk metric:
119
+ gµν = qµν + ∂µπ∂νπ
120
+ and
121
+ gµν = qµν − γ2∂µπ∂νπ
122
+ (2)
123
+ with the Lorentz factor γ =
124
+
125
+ 1 + (∂π)2�−1/2. From this expression of the induced metric,
126
+ we can explicitly write the action (1) using the bulk metric q and the scalar field π. As
127
+ an illustration, see how the cosmological constant part of the action on the brane can
128
+ be expressed:
129
+ SΛ = −Λ
130
+
131
+ d4x√−g = −Λ
132
+
133
+ d4x√−q
134
+
135
+ 1 + (∂π)2
136
+ (3)
137
+ As was pointed out in the original paper by de Rham and Tolley [11], we recover a DBI
138
+ term in the action. This leads us to associate the cosmological constant Λ with a brane
139
+ tension f, following Λ = f 4 for unit convenience.
140
+ Non-relativistic limit
141
+ If the derivatives of the field vanish, ∂µπ = 0, then we recover standard GR. Because
142
+ we are interested in the cosmological setting, where predictions from the ΛCDM model
143
+ based on GR match precisely a large range of observations, we consider only small
144
+ corrections to GR. Therefore, we consider the DBI-Galileon in the so-called non-
145
+ relativistic limit where (∂π)2 ≪ 1. In particular, the DBI part of the action in the
146
+ non-relativistic limit becomes:
147
+ Sf = −f 4
148
+
149
+ d4x√−q
150
+
151
+ 1 + ∇µπ∇µπ
152
+ 2
153
+ − (∇µπ∇µπ)2
154
+ 8
155
+ + . . .
156
+
157
+ (4)
158
+
159
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
160
+ 4
161
+ We see that, to have a canonically normalized field, we can proceed to the field
162
+ redefinition π → ϕ = f 2π. Up to degree 5 in ∂ϕ/f 2, we find the following Lagrangian
163
+ operators for the DBI-Galileon in the non-relativistic limit:
164
+ Lf = 1 + L2
165
+ 2f 4 − X2
166
+ 2 + . . .
167
+ (5)
168
+ LK = − L3
169
+ 2f 6 − 2X
170
+ f 6 [ψ] + . . .
171
+ (6)
172
+ LR = ¯R + 1
173
+ f 4
174
+
175
+ [Φ]2 −
176
+
177
+ Φ2�
178
+ − f 4X
179
+ 2
180
+ ¯R − 2 ¯Rµν∇µϕ∇νϕ
181
+
182
+ + L4
183
+ 4f 8 + . . .
184
+ (7)
185
+ LKGB = 2
186
+ f 6
187
+ ��
188
+ − ¯Rµν [Φ] + 2 ¯RµρΦρ
189
+ ν + ¯RµρνλΦρλ�
190
+ ∇µϕ∇νϕ + 1
191
+ 3 [Φ]3 − [Φ]
192
+
193
+ Φ2�
194
+ + 2
195
+ 3
196
+
197
+ Φ3�
198
+ −1
199
+ 2
200
+ ¯R [ψ]
201
+
202
+ + L5
203
+ 3f 10 + . . .
204
+ (8)
205
+ We defined on the above expressions the tensor Φµν = ∇µ∇νϕ, and the three scalars
206
+ [Φ] = Φµ
207
+ µ, [ψ] = ∂µϕ·Φµν·∂νϕ and X = − (∂ϕ)2 /2f 4. Furthermore, the L2...5 Lagrangian
208
+ operators are the covariant Galileon model Lagrangian operators as defined in [6, 7].
209
+ Therefore, the theory described here is a generalization of the Galileon model. We can
210
+ note that the additional terms in LR and LKGB compared to the Galileon Lagrangian
211
+ operators vanish in the particular case of a flat geometry. We conclude that the DBI-
212
+ Galileon in the non-relativistic limit is a different coviariantization of the flat Galileon
213
+ than the covariant Galileon [7] or dRGT massive gravity [9, 10], with an additional
214
+ terms in Lf and another one in LK. That point noted, we will stay at leading order
215
+ in X in the following. Being a theory of a scalar field interacting with a metric, with
216
+ equations of motion of at most second order, it can be described as a Horndeski theory
217
+ [2, 3] with the following Horndeski functions:
218
+ G2 = A (ϕ) − f 4 (1 − X + . . .)
219
+ (9)
220
+ G3 = M 3
221
+ 5
222
+ f 2 (X + . . .)
223
+ (10)
224
+ G4 = M 2
225
+ P
226
+ 2
227
+
228
+ 1 − X − X2/2 . . .
229
+
230
+ (11)
231
+ G5 = −2β M 3
232
+ 5
233
+ m2f 2
234
+
235
+ X + X2 + . . .
236
+
237
+ (12)
238
+ 3. Cosmological background evolution
239
+ We expand the dynamics of the fields around a flat FLRW background on the 4D slice
240
+ of the bulk along xµ where the properties of the brane, gµν and ϕ, are defined. On this
241
+ slice, the length element is:
242
+ ds2 = qµνdxµdxν = −dt2 + a2 (t) δijdxidxj
243
+ (13)
244
+ It might appear more natural to expand around a flat FLRW background on the
245
+ brane with its induced metric gµν. However, the two metrics qµν and gµν are related
246
+
247
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
248
+ 5
249
+ by a disformal transformation involving the scalar field which, at the background level,
250
+ depends only on the physical time:
251
+ ¯gµν = ¯qµν + ( ˙¯ϕ (t))
252
+ 2
253
+ f 4
254
+ δµ0δν0
255
+ (14)
256
+ where barred quantities are taken at the background level. Therefore, the geometry on
257
+ the brane is also of the FLRW type, but with a different definition for the physical time,
258
+ leading to a different scale factor and expansion history. Because in our case ˙¯ϕ ≪ f 2,
259
+ the expansion history on the 4D slice and on the brane will, then, be approximately the
260
+ same. In addition, equations (9)-(12) determine a self-consistent Horndeski theory of
261
+ gravity with the metric qµν and the scalar field ϕ without referring to the induced metric
262
+ gµν. Thus, in the following we apply the well-known techniques used in the context of
263
+ Horndeski theories.
264
+ In the non-relativistic limit, action (1) reads:
265
+ S =
266
+
267
+ dx4√−q (Lf + LK + LR + LKGB + Lm (qµν, ψm)) .
268
+ (15)
269
+ We define Ω0
270
+ m and Ω0
271
+ r the standard present energy density parameters for pressureless
272
+ matter and radiation respectively, and ¯H the normalized Hubble rate H/H0 with H0
273
+ the present Hubble constant. Prime symbol denotes the derivative with respect to ln a.
274
+ We set:
275
+ ˜x = ϕ′H0
276
+ f 2 ,
277
+ Ω0
278
+ Λ =
279
+ Λ
280
+ 3H2
281
+ 0M 2
282
+ P
283
+ =
284
+ f 4
285
+ 3H2
286
+ 0M 2
287
+ P
288
+ ,
289
+ η =
290
+ M 3
291
+ 5
292
+ M 2
293
+ PH0
294
+ ,
295
+ ξ = β
296
+ η ,
297
+ κ = MPH0
298
+ f 2
299
+ .
300
+ (16)
301
+ Then, the two Friedmann equations derived from action (15) are:
302
+ ¯H2 = −3
303
+ 2
304
+ ¯H4˜x2 − 15
305
+ 8
306
+ ¯H6˜x4 + η ¯H4˜x3 − 10
307
+ 3 ξ ¯H6˜x3 − 14
308
+ 3 ξ ¯H8˜x5
309
+ + Ω0
310
+ m
311
+ a3 + Ω0
312
+ r
313
+ a4 + Ω0
314
+ Λ
315
+
316
+ 1 + 1
317
+ 2
318
+ ¯H2˜x2
319
+
320
+ (17)
321
+ ¯H2 + 2
322
+ 3
323
+ ¯H ¯H′ = −2
324
+
325
+
326
+ 2 ¯H6˜x3 + 5 ¯H5˜x3 ¯H′ + 3 ¯H6˜x2˜x′ + 2 ¯H8˜x5 + 5 ¯H8˜x4˜x′ + 7 ¯H7˜x5 ¯H′�
327
+ − 1
328
+ 2
329
+ ¯H4˜x2 − ¯H3˜x2 ¯H′ − 2
330
+ 3
331
+ ¯H4˜x˜x′ + 1
332
+ 3η ¯H3˜x2( ¯H˜x)′ − 3
333
+ 8
334
+ ¯H6˜x4 − 5
335
+ 4
336
+ ¯H5˜x4 ¯H′ − ¯H6˜x3˜x′
337
+ − Ω0
338
+ r
339
+ 3a4 + Ω0
340
+ Λ
341
+
342
+ 1 − 1
343
+ 2
344
+ ¯H2˜x2
345
+
346
+ (18)
347
+ We see that we recover the ΛCDM equations when setting ˜x to zero, but with a physical
348
+ interpretation of the cosmological constant as the brane tension.
349
+ The DBI model
350
+ proposed here is then an extension of the standard model of cosmology, and as such
351
+ follows a late accelerated expansion but with a physical interpretation of the origin of
352
+ Λ as a brane tension. Using the same methodology and similar notations as in [22], we
353
+ derive the field ϕ equation of motion from action (15). This leads to a system of two
354
+
355
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
356
+ 6
357
+ coupled equations
358
+ ˜x′
359
+ = −˜x + αλ − σγ
360
+ σβ − αω
361
+ ¯H′
362
+ = ωγ − βλ
363
+ σβ − αω
364
+ (19)
365
+ with
366
+ β = 2 ¯H4 + 9
367
+ 2
368
+ ¯H6˜x2 − Ω0
369
+ Λ ¯H2 − 2η ¯H4˜x + 4ξ ¯H6˜x + 40
370
+ 3 ξ ¯H8˜x3
371
+ α = 6 ¯H3˜x − Ω0
372
+ Λ ¯H˜x + 15
373
+ 2
374
+ ¯H5˜x3 − 3η ¯H3˜x2 + 10ξ ¯H5˜x2 + 70
375
+ 3 ξ ¯H7˜x4
376
+ γ = 4 ¯H4˜x − 2Ω0
377
+ Λ ¯H2˜x − η ¯H4˜x2 + 2ξ ¯H6˜x2 − 10
378
+ 3 ξ ¯H8˜x4
379
+ ω = 4
380
+ 3
381
+ ¯H4˜x + ¯H6˜x3 − 1
382
+ 3η ¯H4x2 + 2ξ ¯H6˜x2 + 10
383
+ 3 ξ ¯H8˜x4
384
+ σ = 2
385
+ 3
386
+ ¯H + 2 ¯H3˜x2 + 15
387
+ 12
388
+ ¯H5˜x4 − 1
389
+ 3η ¯H3˜x3 + 10
390
+ 3 ξ ¯H5˜x3 + 14
391
+ 3 ξ ¯H7˜x5
392
+ λ = ¯H2 + Ω0
393
+ r
394
+ 3a4 − Ω0
395
+ Λ + 1
396
+ 2Ω0
397
+ Λ ¯H2˜x2 − 1
398
+ 3
399
+ ¯H4˜x2 − 5
400
+ 8
401
+ ¯H6˜x4 + 1
402
+ 3η ¯H4˜x3 − 2
403
+ 3ξ ¯H6˜x3 − 2ξ ¯H8˜x5
404
+ Given values for the parameters Ω0
405
+ m, η, ξ and κ, and initial conditions ˜x0 and H0,
406
+ this system can be integrated to compute background cosmology observables like the
407
+ distance moduli of type Ia supernovae. In Figure 1 we illustrate this with a Hubble
408
+ diagram prediction compared with recent type Ia supernova data [23]. As an initial
409
+ condition for ˜x, we chose to set ˜x0 = 6 × 10−8 today.
410
+ This is the maximum value
411
+ allowed by the constraint on gravitational wave speed (see Section 4) coming from the
412
+ quasi simultaneous observation of photons and gravitational waves after neutron star
413
+ merger event GW170817A [24]. Nevertheless, before discussing more the cosmological
414
+ scenarios proposed by the DBI-Galileon model, stability conditions much be computed
415
+ first to assess the viability of the models for any set of parameters at the perturbation
416
+ level.
417
+ 4. Stability conditions
418
+ To be viable as a description of our Universe, the model has to fulfill stability
419
+ conditions.
420
+ These requirements apply to degrees of freedom propagating around
421
+ the fixed background, i.e.
422
+ to the cosmological perturbations, that are capable of
423
+ undermining the stability of the Universe.
424
+ In the determination of the stability
425
+ conditions for the DBI-Galileon model in the non-relativistic limit, we use the formalism
426
+ described in [25] for Horndeski theories. In our case, these stability conditions are defined
427
+
428
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
429
+ 7
430
+ 14
431
+ 16
432
+ 18
433
+ 20
434
+ 22
435
+ 24
436
+ 26
437
+ Distance moduli [mag]
438
+ 0.00
439
+ 0.25
440
+ 0.50
441
+ 0.75
442
+ 1.00
443
+ 1.25
444
+ 1.50
445
+ 1.75
446
+ Redshift z
447
+ 0.2
448
+ 0.0
449
+ Residuals [mag]
450
+ Figure 1.
451
+ Hubble diagram prediction for the non-relativistic DBI-Galileon model for Ω0
452
+ Λ = 0.7,
453
+ ˜x0 = 6 × 10−8, η = ξ = 0 (blue) compared with binned Pantheon data (black points) [23]. We used
454
+ M = 23.81 for the offset magnitude of the diagram. Residuals to the fit are presented in the bottom
455
+ panel.
456
+ from the following quantities derived from the particular Horndeski functions (9) to (12):
457
+ ω1 = 1 + 1
458
+ 2
459
+ ¯H2˜x2 + 3
460
+ 8
461
+ ¯H4˜x4 + 2ξ ¯H4˜x3 + 2ξ ¯H6˜x5
462
+ (20)
463
+ ω2 = 2 ¯H + 3 ¯H3˜x2 + 15
464
+ 4
465
+ ¯H5˜x4 − η ¯H3˜x3 + 10ξ ¯H5˜x3 + 14ξ ¯H7˜x5
466
+ (21)
467
+ ω3 = 9
468
+
469
+ − ¯H2 + 1
470
+ 2Ω0
471
+ Λ ¯H2˜x2 − 3 ¯H4˜x2 + 2η ¯H4˜x3 − 10ξ ¯H6˜x3 − 45
472
+ 8
473
+ ¯H6˜x4 − 56
474
+ 3 ξ ¯H8˜x5
475
+
476
+ (22)
477
+ ω4 = 1 − 1
478
+ 2
479
+ ¯H2˜x2 − 1
480
+ 8
481
+ ¯H4˜x4 + 2ξ ¯H3˜x2( ¯H˜x)′ + 2ξ ¯H5˜x4( ¯H˜x)′
482
+ (23)
483
+ Tensorial stability conditions
484
+ In order to avoid ghosts and Laplacian instabilities, we impose the following constraints
485
+ on the sign of the kinetic term and on the sign of the gravitational waves speed squared:
486
+ Qt ≡ ω1
487
+ 4 = 1
488
+ 4 + 1
489
+ 8
490
+ ¯H2˜x2 + 3
491
+ 32
492
+ ¯H4˜x4 + 1
493
+ 2ξ ¯H4˜x3 + 1
494
+ 2ξ ¯H6˜x5 > 0
495
+ (24)
496
+ c2
497
+ t ≡ ω4
498
+ ω1
499
+ = 1 − 1
500
+ 2 ¯H2˜x2 − 1
501
+ 8 ¯H4˜x4 + 2ξ ¯H3˜x2( ¯H˜x)′ + 2ξ ¯H5˜x4( ¯H˜x)′
502
+ 1 + 1
503
+ 2 ¯H2˜x2 + 3
504
+ 8 ¯H4˜x4 + 2ξ ¯H4˜x3 + 2ξ ¯H6˜x5
505
+ ≥ 0
506
+ (25)
507
+ In particular, we see that the gravitational wave speed depends on ˜x, and tends to
508
+ 1 when ˜x → 0 :
509
+ 4Qt ≃ 1 + 2( ¯H˜x)2
510
+ (26)
511
+ ct ≃ 1 − 1
512
+ 2( ¯H˜x)2
513
+ (27)
514
+
515
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
516
+ 8
517
+ Given the very tight constraint on the speed of gravitational waves, equal to the speed
518
+ of light up to a ∼ 10−15 difference [24, 26, 27], this justifies a posteriori the relevance
519
+ of the non-relativistic limit where ˜x ≪ 1. Moreover, we see that tensorial perturbations
520
+ are stable in this limit since Qt > 0.
521
+ Scalar stability conditions
522
+ Similar stability conditions apply to the scalar degrees of freedom, here including the
523
+ scalar perturbations of matter components:
524
+ Qs ≡ ω1 (4ω1ω3 + 9ω2
525
+ 2)
526
+ 3ω2
527
+ 2
528
+ > 0
529
+ (28)
530
+ c2
531
+ s ≡ 3 (2ω2
532
+ 1ω2H − ω2
533
+ 2ω4 + 4ω1ω2 ˙ω1 − 2ω2
534
+ 1 ˙ω2) − 6ω2
535
+ 1
536
+ � (1 + wi) ρi
537
+ ω1 (4ω1ω3 + 9ω2
538
+ 2)
539
+ ≥ 0
540
+ (29)
541
+ where wi and ρi are respectively the equation of state parameter and the energy density
542
+ of the fluid i, and the sum runs over all the components of the Universe (here only
543
+ pressureless matter and radiation). At the lowest order in ˜x, we get:
544
+ Qs ≃ 3
545
+ 2(Ω0
546
+ Λ − ¯H2)˜x2
547
+ (30)
548
+ c2
549
+ s ≃ 1 + 2
550
+
551
+ η ¯H2˜x′ − ¯H ¯H′ − 2ξ ¯H4˜x′�
552
+ 3
553
+
554
+ Ω0
555
+ Λ − ¯H2�
556
+ (31)
557
+ With ˜x ≪ 1, a fit of the DBI-Galileon model to data leads to cosmological
558
+ parameters close to the standard model ones: Ω0
559
+ m ≈ 0.3 and Ω0
560
+ Λ ≈ 0.7 [28]. Therefore,
561
+ from the first Friedmann equation, we get Ω0
562
+ Λ < ¯H2 for all relevant models in agreement
563
+ with cosmological observations. As Qs ≤ 0, the DBI-Galileon model contains scalar
564
+ instabilities unless it reduces to GR. One way to avoid this would be to add a spatial
565
+ curvature to the metric, but with a strong energy density (at least ∼ 0.3) which is also
566
+ excluded by observations [28].
567
+ 5. Discussion
568
+ Physical interpretation
569
+ From the definition (28), we see that the dominant terms come G2 (giving the Ω0
570
+ Λ term)
571
+ and G4 (giving the ¯H2 term). The competition between the two terms leads to the
572
+ ghost-like behaviour in a cosmological setting: Qs ≤ 0. In other words, it is the result
573
+ of the competition between the DBI and the Einstein-Hilbert terms. The DBI action
574
+ will have the effect of stretching the brane towards an extremal surface, whereas the
575
+ Einstein-Hilbert term on the brane will tend to make the brane contract on itself from
576
+ the effect of curvature. However, in the non-relativistic limit of the DBI-Galileon, the
577
+ Einstein-Hilbert term destabilizes the scalar field perturbations and the stretching effect
578
+ from the cosmological constant is not strong enough to counterbalance, leading to an
579
+ instantaneous decay of the vacuum state.
580
+
581
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
582
+ 9
583
+ Because the DBI-Galileon action is the most general one can find of a 4D probe
584
+ brane in a 5D bulk, we expect this statement to be quite general for all such theories
585
+ studied in the current context. Indeed, this is true in a standard cosmological setting
586
+ which is realised with an FLRW slicing of the bulk space-time (equivalent to an FLRW
587
+ background on the brane). Therefore, the only way to evade this ghostly behaviour in
588
+ cosmology is to include the full relativistic dynamics of the theory (˜x ∼ 1). We have
589
+ seen that, in this case, we expect significant deviations of the speed of gravitational
590
+ waves ct from c. This is not a definitive impossibility though if the full DBI-Galileon
591
+ is viewed as an effective theory valid only at cosmological scales for which the speed
592
+ of gravitational waves has not been probed [29, 30].
593
+ Indeed, the constraint on the
594
+ gravitational speed from the observation of GW170817 in coincidence with GRB170817A
595
+ [24] is only valid on small scales probed by LIGO and Virgo. A modification of the
596
+ dispersion relation of gravitational waves at small scales from operators present in the
597
+ UV complete theory could allow ct ̸= c on cosmological scales while being compatible
598
+ with current astrophysical observations. Waiting for the next generation of gravitational
599
+ wave interferometers, in particular LISA, which will be able to probe this relation at
600
+ larger scales [31], this possibility remains open.
601
+ Direct coupling to matter
602
+ In the context of cosmology, where standard model matter is present, there might be
603
+ direct coupling to the scalar field. In that case, the metric ˜q to which matter is sensitive
604
+ is different than the space-time metric q:
605
+ S =
606
+
607
+ dx4√−q (Lf + LK + LR + LKGB) +
608
+
609
+ dx4�
610
+ −˜qLm (˜qµν, ψm) .
611
+ (32)
612
+ It has been shown in [32] that the two metrics are related by a disformal transformation
613
+ of the following form:
614
+ qµν = A
615
+
616
+ ϕ, ˜X
617
+
618
+ ˜qµν + B
619
+
620
+ ϕ, ˜X
621
+ � ∂µϕ∂νϕ
622
+ f 4
623
+ (33)
624
+ where A and B are arbitrary functions of the scalar field and ˜X = −˜qµν∂µϕ∂νϕ/2f 4.
625
+ For simplicity and following the treatment of the covariant Galileon [22], we assume
626
+ that A and B are constant parameters. This can be further justified by the fact that,
627
+ a dependency on X would introduce, in general, higher order terms which would go
628
+ beyond the framework of Horndeski theories [33], and a dependency on ϕ would, in
629
+ general, break the shift symmetry followed by the scalar field ϕ in the probe brane
630
+ context. Note that, when A = −B, matter is coupled to the induced metric on the
631
+ brane.
632
+ Contrary to the covariant Galileon, the DBI-Galileon action is not invariant by
633
+ such a change of reference frame. However, new terms that can not be absorbed into
634
+ a redefinition of the parameters arise only at higher order in ˜X. Therefore, the non-
635
+ relativistic dynamics is not change by the introduction of a direct coupling between the
636
+
637
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
638
+ 10
639
+ scalar field and matter of the form (33) with constant parameters. In particular, this
640
+ does not prevent the perturbations around the FLRW background from showing ghost
641
+ instabilities.
642
+ Generalization
643
+ The DBI-Galileon is a particular example of the more general class of shift-symmetric
644
+ Horndeski theories. These are subclass of Horndeski theories which are invariant under
645
+ a shift symmetry of the scalar field ϕ → ϕ + c [34, 35]. In these theories, the arbitrary
646
+ Horndeski functions are restricted to be functions of X alone. In order to make the
647
+ non-relativistic limit apparent, we Taylor expand these arbitrary functions around GR:
648
+ G2 ≡ Λ +
649
+ +∞
650
+
651
+ n=1
652
+ g(n)
653
+ 2 Xn
654
+ (34)
655
+ G3 ≡
656
+ +∞
657
+
658
+ n=1
659
+ g(n)
660
+ 3 Xn
661
+ (35)
662
+ G4 ≡ M 2
663
+ P
664
+ 2
665
+ +
666
+ +∞
667
+
668
+ n=1
669
+ g(n)
670
+ 4 Xn
671
+ (36)
672
+ G5 ≡
673
+ +∞
674
+
675
+ n=1
676
+ g(n)
677
+ 5 Xn
678
+ (37)
679
+ The constant terms in G3 and G5 do not appear in the expansion as they lead
680
+ to total derivative terms. Because the Horndeski functions depend only on X, the ω
681
+ functions that determine the stability conditions reduce to:
682
+ ω1 ≡ 2G4 − 2X
683
+
684
+ 2G4,X + ˙φHG5,X
685
+
686
+ (38)
687
+ ω2 ≡ 4HG4 − 2X
688
+
689
+ ˙φG3,X + 8HG4,X + 5 ˙φH2G5,X
690
+
691
+ − 4X2H
692
+
693
+ 4G4,XX + ˙φHG5,XX
694
+
695
+ (39)
696
+ ω3 ≡ − 18H2G4 + 3X
697
+
698
+ G2,X + 12 ˙φHG3,X + 42H2G4,X + 30 ˙φH3G5,X
699
+
700
+ + 6X2 �
701
+ G2,XX + 3 ˙φHG3,XX + 48H2G4,XX + 13H3 ˙φG5,XX
702
+
703
+ + 12X3H2 �
704
+ 6G4,XXX + H ˙φG5,XXX
705
+
706
+ (40)
707
+ ω4 ≡ 2G4 − 2X ¨φG5,X
708
+ (41)
709
+ From these, we can compute the quantity Qs up to first order in X:
710
+ Qs = X
711
+ H2
712
+
713
+ g(1)
714
+ 2
715
+ + 6H2g(1)
716
+ 4
717
+
718
+ + O
719
+
720
+ X
721
+ 3
722
+ 2
723
+
724
+ (42)
725
+ This leads to a very simple formulation of the no-ghost condition, independent of
726
+ X, in the context of Shift-Symmetric Horndeski theories in the non-relativistic limit:
727
+ g(1)
728
+ 2
729
+ + 6H2g(1)
730
+ 4
731
+ > 0
732
+ (43)
733
+
734
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
735
+ 11
736
+ In the context of the brane galileon, where g(1)
737
+ 4
738
+ = −M 2
739
+ P/2 and g(1)
740
+ 2
741
+ = Λ, this is
742
+ equivalent to the inequality which is never fulfilled in flat space:
743
+ Λ − 3M 2
744
+ PH2 > 0
745
+
746
+ Ω0
747
+ Λ > ¯H2
748
+ (44)
749
+ Other stability conditions are given by:
750
+ c2
751
+ s ≃ 1 + 2¨φg(1)
752
+ 3
753
+ + 4 ˙Hg(1)
754
+ 4
755
+ + 2¨φH2g(1)
756
+ 5
757
+ g(1)
758
+ 2
759
+ + 6H2g(1)
760
+ 4
761
+ > 0
762
+ (45)
763
+ Qt ≃ M 2
764
+ P
765
+ 4
766
+ (46)
767
+ c2
768
+ t ≃ 1
769
+ (47)
770
+ where we expressed these quantities at the lowest order. The two tensorial conditions
771
+ are, thus, automatically satisfied in this context.
772
+ On the other hand, the stability
773
+ conditions for scalar perturbations at the lowest order give a simple inequality involving
774
+ the parameters of the Taylor expansion, that can be easily checked at the background
775
+ level.
776
+ 6. Conclusion
777
+ We described the DBI-Galileon theory of a four-dimensional brane evolving in a 5D bulk
778
+ space-time in the non-relativistic limit where its local kinetic energy is small compared
779
+ to its tension. This model belongs to the class of shift-symmetric Horndeski theories,
780
+ themselves being a subclass of the more general family of Horndeski theories. From
781
+ the construction of the DBI-Galileon model, the free parameters of the model acquire a
782
+ physical meaning. In particular, the interpretation of the cosmological constant is linked
783
+ to the brane tension energy density. We derived the equations driving the evolution
784
+ of the late-time Universe around a spatially flat FLRW cosmological background and
785
+ studied the stability of scalar and tensorial perturbations.
786
+ This model reduces to
787
+ an expansion around standard GR, and therefore around standard ΛCDM in the
788
+ cosmological context. As such, it is naturally compatible with data in the non-relativistic
789
+ limit provided the effect of the scalar field is small enough, even considering the speed
790
+ of the gravitational waves.
791
+ However, it revealed fatal ghostly behaviour for scalar
792
+ perturbations around the FLRW background. From there, we derived the corresponding
793
+ stability conditions for shift-symmetric Horndeski theories in the non-relativistic limit
794
+ in the cosmological context and found very simple formulations for these conditions.
795
+ Acknowledgements
796
+ We would like to thank Marc Besan¸con, Arnaud de Mattia and Vanina Ruhlmann-
797
+ Kleider for their comments on the present paper. We also want to thank David Langlois
798
+ for useful and interesting comments and suggestions.
799
+
800
+ Instability of the cosmological DBI-Galileon in the non-relativistic limit
801
+ 12
802
+ References
803
+ [1] Brax P 2017 Reports on Progress in Physics 81 016902 URL https://dx.doi.org/10.1088/
804
+ 1361-6633/aa8e64
805
+ [2] Horndeski G W 1974 Int. J. Theor. Phys. 10 363–384
806
+ [3] Deffayet C, Gao X, Steer D A and Zahariade G 2011 Phys. Rev. D 84 064039 (Preprint 1103.3260)
807
+ [4] Langlois D and Noui K 2016 JCAP 07 016 (Preprint 1512.06820)
808
+ [5] Langlois D 2019 Int. J. Mod. Phys. D 28 1942006 (Preprint 1811.06271)
809
+ [6] Nicolis A, Rattazzi R and Trincherini E 2009 Phys. Rev. D 79 064036 (Preprint 0811.2197)
810
+ [7] Deffayet C, Esposito-Farese G and Vikman A 2009 Phys. Rev. D 79 084003 (Preprint 0901.1314)
811
+ [8] Gubitosi G and Linder E V 2011 Physics Letters B 703 113–118 ISSN 0370-2693 URL https:
812
+ //www.sciencedirect.com/science/article/pii/S0370269311008707
813
+ [9] de Rham C and Gabadadze G 2010 Phys. Rev. D 82 044020 (Preprint 1007.0443)
814
+ [10] de Rham C, Gabadadze G and Tolley A J 2011 Phys. Rev. Lett. 106 231101 (Preprint 1011.1232)
815
+ [11] de Rham C and Tolley A J 2010 JCAP 05 015 (Preprint 1003.5917)
816
+ [12] Hinterbichler K, Trodden M and Wesley D 2010 Phys. Rev. D 82 124018 (Preprint 1008.1305)
817
+ [13] Silverstein E and Tong D 2004 Phys. Rev. D 70 103505 (Preprint hep-th/0310221)
818
+ [14] Langlois D, Renaux-Petel S, Steer D A and Tanaka T 2008 Phys. Rev. D 78 063523 (Preprint
819
+ 0806.0336)
820
+ [15] Zumalacarregui M, Koivisto T S and Mota D F 2013 Phys. Rev. D 87 083010 (Preprint 1210.8016)
821
+ [16] Goon G, Hinterbichler K, Joyce A and Trodden M 2012 JHEP 06 004 (Preprint 1203.3191)
822
+ [17] Dobado A and Maroto A L 2001 Nucl. Phys. B 592 203–218 (Preprint hep-ph/0007100)
823
+ [18] Cembranos J A R, Dobado A and Maroto A L 2001 Goldstone bosons and solitons on the brane
824
+ 20th International Symposium on Lepton and Photon Interactions at High Energies (LP 01)
825
+ (Preprint hep-ph/0107155)
826
+ [19] Burrage C, de Rham C and Heisenberg L 2011 JCAP 05 025 (Preprint 1104.0155)
827
+ [20] Goon G, Hinterbichler K and Trodden M 2011 JCAP 07 017 (Preprint 1103.5745)
828
+ [21] Lovelock D 1971 J. Math. Phys. 12 498–501
829
+ [22] Appleby S and Linder E V 2012 Journal of Cosmology and Astroparticle Physics 2012 043 URL
830
+ https://dx.doi.org/10.1088/1475-7516/2012/03/043
831
+ [23] Scolnic D M et al. 2018 The Astrophysical Journal 859 101 ISSN 15384357 (Preprint 1710.00845)
832
+ URL http://arxiv.org/abs/1710.00845http://dx.doi.org/10.17909/T95Q4X
833
+ [24] Abbott B P et al. (LIGO Scientific, Virgo, Fermi-GBM, INTEGRAL) 2017 Astrophys. J. Lett.
834
+ 848 L13 (Preprint 1710.05834)
835
+ [25] De Felice A and Tsujikawa S 2012 JCAP 02 007 (Preprint 1110.3878)
836
+ [26] Creminelli P and Vernizzi F 2017 Phys. Rev. Lett. 119 251302 (Preprint 1710.05877)
837
+ [27] Ezquiaga J M and Zumalac´arregui M 2017 Phys. Rev. Lett. 119 251304 (Preprint 1710.05901)
838
+ [28] Planck Collaboration 2020 A&A 641 A6 URL https://doi.org/10.1051/0004-6361/201833910
839
+ [29] de Rham C and Melville S 2018 Phys. Rev. Lett. 121 221101 (Preprint 1806.09417)
840
+ [30] Ezquiaga J M and Zumalac´arregui M 2018 Front. Astron. Space Sci. 5 44 (Preprint 1807.09241)
841
+ [31] Barausse E et al. 2020 Gen. Rel. Grav. 52 81 (Preprint 2001.09793)
842
+ [32] Bekenstein J D 1993 Phys. Rev. D 48 3641–3647 (Preprint gr-qc/9211017)
843
+ [33] Bettoni D and Liberati S 2013 Phys. Rev. D 88 084020 (Preprint 1306.6724)
844
+ [34] Sotiriou T P and Zhou S Y 2014 Phys. Rev. Lett. 112 251102 (Preprint 1312.3622)
845
+ [35] Sotiriou T P and Zhou S Y 2014 Phys. Rev. D 90 124063 (Preprint 1408.1698)
846
+
C9AzT4oBgHgl3EQfwf5z/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,457 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf,len=456
2
+ page_content='Instability of the cosmological DBI-Galileon in the non-relativistic limit C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
3
+ page_content=' Leloup1,2, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
4
+ page_content=' Heitz3 and J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
5
+ page_content=' Neveu3,4 1 Universit´e Paris-Cit´e, CNRS, Astroparticule et Cosmologie, 75013 Paris, France 2 Kavli Institute for the Physics and Mathematics of the Universe (Kavli IPMU, WPI), UTIAS, The University of Tokyo, Kashiwa, Chiba 277-8583, Japan 3 Universit´e Paris-Saclay, CNRS, IJCLab, 91405, Orsay, France 4 Sorbonne Universit´e, CNRS, Universit´e de Paris, LPNHE, 75252 Paris Cedex 05, France Abstract.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
6
+ page_content=' The DBI-Galileon model is a tensor-scalar theory of gravity which finds its foundation as the most general theory of the dynamics of a 4D brane embedded in a 5D bulk.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
7
+ page_content=' It is of particular interest as it provides a few free parameters with a physical meaning, such as the cosmological constant which is there related to the brane tension.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
8
+ page_content=' Most studies of this model have been performed assuming a maximally symmetric geometry for the 5D bulk, in which it has been shown that the theory reduces to various types of Galileon.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
9
+ page_content=' In contrast, the general case for the geometry of the bulk provides a different covariantization of the Galileon model than the covariant Galileon: the DBI-Galileon.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
10
+ page_content=' From the tight constraints on the gravitational waves speed, we are naturally led to consider the non-relativistic limit of the model where the kinetic energy of the brane is small compared to its tension, that we study in the context of late-time cosmology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
11
+ page_content=' The DBI-Galileon in the non-relativistic limit is simply an expansion around General Relativity (GR) which can be expressed as a shift-symmetric Horndeski theory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
12
+ page_content=' We developed the description of this theory at the background and perturbation level.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
13
+ page_content=' However, by studying the scalar and tensor perturbations around a flat FLRW background, we found that they contain a ghost degree of freedom leading to fatal instability of the vacuum for every combination of the free parameters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
14
+ page_content=' As a lesson, we emphasized which of the Horndeski terms competes to avoid this instability in more general cases.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
15
+ page_content=' 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
16
+ page_content=' Introduction Dark energy has been modelled by a large variety of theories since decades.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
17
+ page_content=' Among these, many rely on the introduction of additional scalar fields whose dynamics, at the origin of the late-time acceleration of the expansion of the Universe, is determined by arbitrary parametric functions, potentials and/or coupling (see e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
18
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
19
+ page_content=' [1]).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
20
+ page_content=' These are the so-called scalar-tensor theories of modified gravity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
21
+ page_content=' In particular, the class of Horndeski theories is of great interest as it contains all models of modified gravity with a single arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
22
+ page_content='01723v1 [hep-th] 4 Jan 2023 Instability of the cosmological DBI-Galileon in the non-relativistic limit 2 additional scalar field leading to second-order equations of motion [2, 3].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
23
+ page_content=' Extensions of Horndeski theories to scalar-tensor theories of one scalar field with equations of motion of higher orders have also been explored [4, 5].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
24
+ page_content=' Particular Horndeski theories are described by the specification of four arbitrary functions of the scalar field and its kinetic energy, leading to a huge variety of models and phenomenological behaviours.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
25
+ page_content=' Among these wide classes of models, some can be built from first physical principles or arguments of symmetry.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
26
+ page_content=' For instance, the Galileon model [6] and its covariant extension [7] was built by imposing a galilean symmetry for the scalar field, leaving only five free numerical parameters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
27
+ page_content=' We can also cite, among many others, the pure kinetic gravity theory [8], massive gravity in the non-relativistic limit [9, 10] and the DBI-Galileon [11] which is the main object of this paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
28
+ page_content=' The DBI-Galileon is a model that falls into the class of Brane-world scenarios of extra-dimension theories, where the matter fields are confined on a 4D brane while gravity can propagate into the additional spatial dimensions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
29
+ page_content=' Of most interest for the DBI-Galileon is the case of a single extra-dimension as it has been shown that theories with more co-dimensions exhibit ghosts either in the flat or self-accelerating de Sitter solution [12].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
30
+ page_content=' The action include a volume term for the 4D brane in the 5D bulk which leads to the well-known Dirac-Born-Infeld (DBI) action.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
31
+ page_content=' This action, and DBI-like extensions, can lead to a self-accelerating solution and has been thoroughly studied as a candidate model in the early Universe cosmic inflation paradigm [13, 14].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
32
+ page_content=' In addition, the DBI-Galileon model exhibits the Galileon Lagrangians in the non-relativistic limit [11] but giving a physical meaning to their free parameters: the Planck mass in the brane, the Planck mass in the bulk, etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
33
+ page_content=' In particular, the brane tension here plays the role of the cosmological constant which brings a possible interpretation of its nature.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
34
+ page_content=' The original probe brane construction has been revisited in [15] where the matter metric is disformally related to a standard gravitational metric, or in [16] in the framework of spontaneous symmetry breaking for the 5D space-time symmetries broken by the presence of the brane, bridging the gap with Brane-world scenarios developed in the context of quantum field theory and an interpretation of the scalar field as a Nambu-Goldstone boson [17, 18].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
35
+ page_content=' The DBI-Galileon model has been studied extensively in special cases of the maximally symmetric bulk geometry [19, 20].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
36
+ page_content=' However, to our knowledge, no study of the DBI-Galileon in the late-time cosmology setting as a potential candidate for Dark Energy has been performed so far.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
37
+ page_content=' In this paper we develop the DBI-Galileon theory in the non relativistic limit (Section 2) and study its dynamics in the Friedmann-Lemaˆıtre-Robertson-Walker flat metric (Section 3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
38
+ page_content=' The perturbation stability is explored in Section 4 and then discussed in Section 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
39
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
40
+ page_content=' DBI-Galileon in the non-relativistic limit DBI-Galileon We are interested in the description of a four dimensional brane universe embedded in Instability of the cosmological DBI-Galileon in the non-relativistic limit 3 a five dimensional bulk from the cosmological perspective.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
41
+ page_content=' In this context,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
42
+ page_content=' it has been shown in [11] that the most general action on the brane is given by the 4D Lovelock terms [21] inside the brane and the boundary terms associated to the 5D Lovelock terms in the bulk: S = � dx4√−g � −Λ − M 3 5K + M 2 P 2 R − β M 3 5 m2 KGB + Lm (˜qµν,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
43
+ page_content=' ψm) � ,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
44
+ page_content=' (1) where g is the induced metric on the brane,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
45
+ page_content=' Λ the brane cosmological constant,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
46
+ page_content=' K is the extrinsic curvature of the brane,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
47
+ page_content=' R is the Ricci scalar on the brane,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
48
+ page_content=' KGB is the boundary term on the brane of the Gauss-Bonnet scalar in the bulk and Lm is the Lagrangian density of matter that lives confined in the brane.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
49
+ page_content=' In the above action has been defined the 4D Planck mass MP, its 5D counterpart M5 and their ratio m = M 3 5/M 2 P, and β is an arbitrary constant parameter.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
50
+ page_content=' Because the action is defined on the 4-dimensional brane, the above quantities are expressed in terms of the induced metric g, as opposed to the 5-dimensional bulk metric G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
51
+ page_content=' Assuming we have a coordinate system (xµ, y) in the bulk, Greek letters being defined to span from 0 to 3, such that the length element in this frame is ds2 = qµνdxµdxν+(dy)2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
52
+ page_content=' The brane position in the bulk is defined by y = π (xµ) such that we can express the induced metric from the bulk metric: gµν = qµν + ∂µπ∂νπ and gµν = qµν − γ2∂µπ∂νπ (2) with the Lorentz factor γ = � 1 + (∂π)2�−1/2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
53
+ page_content=' From this expression of the induced metric, we can explicitly write the action (1) using the bulk metric q and the scalar field π.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
54
+ page_content=' As an illustration, see how the cosmological constant part of the action on the brane can be expressed: SΛ = −Λ � d4x√−g = −Λ � d4x√−q � 1 + (∂π)2 (3) As was pointed out in the original paper by de Rham and Tolley [11], we recover a DBI term in the action.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
55
+ page_content=' This leads us to associate the cosmological constant Λ with a brane tension f, following Λ = f 4 for unit convenience.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
56
+ page_content=' Non-relativistic limit If the derivatives of the field vanish, ∂µπ = 0, then we recover standard GR.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
57
+ page_content=' Because we are interested in the cosmological setting, where predictions from the ΛCDM model based on GR match precisely a large range of observations, we consider only small corrections to GR.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
58
+ page_content=' Therefore, we consider the DBI-Galileon in the so-called non- relativistic limit where (∂π)2 ≪ 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
59
+ page_content=' In particular, the DBI part of the action in the non-relativistic limit becomes: Sf = −f 4 � d4x√−q � 1 + ∇µπ∇µπ 2 − (∇µπ∇µπ)2 8 + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
60
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
61
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
62
+ page_content=' � (4) Instability of the cosmological DBI-Galileon in the non-relativistic limit 4 We see that, to have a canonically normalized field, we can proceed to the field redefinition π → ϕ = f 2π.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
63
+ page_content=' Up to degree 5 in ∂ϕ/f 2, we find the following Lagrangian operators for the DBI-Galileon in the non-relativistic limit: Lf = 1 + L2 2f 4 − X2 2 + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
64
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
65
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
66
+ page_content=' (5) LK = − L3 2f 6 − 2X f 6 [ψ] + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
67
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
68
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
69
+ page_content=' (6) LR = ¯R + 1 f 4 � [Φ]2 − � Φ2� − f 4X 2 ¯R − 2 ¯Rµν∇µϕ∇νϕ � + L4 4f 8 + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
70
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
71
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
72
+ page_content=' (7) LKGB = 2 f 6 �� − ¯Rµν [Φ] + 2 ¯RµρΦρ ν + ¯RµρνλΦρλ� ∇µϕ∇νϕ + 1 3 [Φ]3 − [Φ] � Φ2� + 2 3 � Φ3� −1 2 ¯R [ψ] � + L5 3f 10 + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
73
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
74
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
75
+ page_content=' (8) We defined on the above expressions the tensor Φµν = ∇µ∇νϕ, and the three scalars [Φ] = Φµ µ, [ψ] = ∂µϕ·Φµν·∂νϕ and X = − (∂ϕ)2 /2f 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
76
+ page_content=' Furthermore, the L2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
77
+ page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
78
+ page_content='5 Lagrangian operators are the covariant Galileon model Lagrangian operators as defined in [6, 7].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
79
+ page_content=' Therefore, the theory described here is a generalization of the Galileon model.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
80
+ page_content=' We can note that the additional terms in LR and LKGB compared to the Galileon Lagrangian operators vanish in the particular case of a flat geometry.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
81
+ page_content=' We conclude that the DBI- Galileon in the non-relativistic limit is a different coviariantization of the flat Galileon than the covariant Galileon [7] or dRGT massive gravity [9, 10], with an additional terms in Lf and another one in LK.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
82
+ page_content=' That point noted, we will stay at leading order in X in the following.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
83
+ page_content=' Being a theory of a scalar field interacting with a metric, with equations of motion of at most second order, it can be described as a Horndeski theory [2, 3] with the following Horndeski functions: G2 = A (ϕ) − f 4 (1 − X + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
84
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
85
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
86
+ page_content=') (9) G3 = M 3 5 f 2 (X + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
87
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
88
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
89
+ page_content=') (10) G4 = M 2 P 2 � 1 − X − X2/2 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
90
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
91
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
92
+ page_content=' � (11) G5 = −2β M 3 5 m2f 2 � X + X2 + .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
93
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
94
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
95
+ page_content=' � (12) 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
96
+ page_content=' Cosmological background evolution We expand the dynamics of the fields around a flat FLRW background on the 4D slice of the bulk along xµ where the properties of the brane, gµν and ϕ, are defined.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
97
+ page_content=' On this slice, the length element is: ds2 = qµνdxµdxν = −dt2 + a2 (t) δijdxidxj (13) It might appear more natural to expand around a flat FLRW background on the brane with its induced metric gµν.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
98
+ page_content=' However, the two metrics qµν and gµν are related Instability of the cosmological DBI-Galileon in the non-relativistic limit 5 by a disformal transformation involving the scalar field which, at the background level, depends only on the physical time: ¯gµν = ¯qµν + ( ˙¯ϕ (t)) 2 f 4 δµ0δν0 (14) where barred quantities are taken at the background level.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
99
+ page_content=' Therefore, the geometry on the brane is also of the FLRW type, but with a different definition for the physical time, leading to a different scale factor and expansion history.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
100
+ page_content=' Because in our case ˙¯ϕ ≪ f 2, the expansion history on the 4D slice and on the brane will, then, be approximately the same.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
101
+ page_content=' In addition, equations (9)-(12) determine a self-consistent Horndeski theory of gravity with the metric qµν and the scalar field ϕ without referring to the induced metric gµν.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
102
+ page_content=' Thus, in the following we apply the well-known techniques used in the context of Horndeski theories.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
103
+ page_content=' In the non-relativistic limit, action (1) reads: S = � dx4√−q (Lf + LK + LR + LKGB + Lm (qµν, ψm)) .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
104
+ page_content=' (15) We define Ω0 m and Ω0 r the standard present energy density parameters for pressureless matter and radiation respectively, and ¯H the normalized Hubble rate H/H0 with H0 the present Hubble constant.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
105
+ page_content=' Prime symbol denotes the derivative with respect to ln a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
106
+ page_content=' We set: ˜x = ϕ′H0 f 2 , Ω0 Λ = Λ 3H2 0M 2 P = f 4 3H2 0M 2 P , η = M 3 5 M 2 PH0 , ξ = β η , κ = MPH0 f 2 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
107
+ page_content=' (16) Then,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
108
+ page_content=' the two Friedmann equations derived from action (15) are: ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
109
+ page_content='¯H2 = −3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
110
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
111
+ page_content='¯H4˜x2 − 15 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
112
+ page_content='8 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
113
+ page_content='¯H6˜x4 + η ¯H4˜x3 − 10 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
114
+ page_content='3 ξ ¯H6˜x3 − 14 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
115
+ page_content='3 ξ ¯H8˜x5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
116
+ page_content='+ Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
117
+ page_content='m ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
118
+ page_content='a3 + Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
119
+ page_content='r ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
120
+ page_content='a4 + Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
121
+ page_content='Λ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
122
+ page_content='� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
123
+ page_content='1 + 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
124
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
125
+ page_content='¯H2˜x2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
126
+ page_content='� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
127
+ page_content='(17) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
128
+ page_content='¯H2 + 2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
129
+ page_content='3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
130
+ page_content='¯H ¯H′ = −2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
131
+ page_content='3ξ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
132
+ page_content='� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
133
+ page_content='2 ¯H6˜x3 + 5 ¯H5˜x3 ¯H′ + 3 ¯H6˜x2˜x′ + 2 ¯H8˜x5 + 5 ¯H8˜x4˜x′ + 7 ¯H7˜x5 ¯H′� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
134
+ page_content='− 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
135
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
136
+ page_content='¯H4˜x2 − ¯H3˜x2 ¯H′ − 2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
137
+ page_content='3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
138
+ page_content='¯H4˜x˜x′ + 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
139
+ page_content='3η ¯H3˜x2( ¯H˜x)′ − 3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
140
+ page_content='8 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
141
+ page_content='¯H6˜x4 − 5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
142
+ page_content='4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
143
+ page_content='¯H5˜x4 ¯H′ − ¯H6˜x3˜x′ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
144
+ page_content='− Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
145
+ page_content='r ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
146
+ page_content='3a4 + Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
147
+ page_content='Λ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
148
+ page_content='� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
149
+ page_content='1 − 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
150
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
151
+ page_content='¯H2˜x2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
152
+ page_content='� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
153
+ page_content='(18) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
154
+ page_content='We see that we recover the ΛCDM equations when setting ˜x to zero,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
155
+ page_content=' but with a physical interpretation of the cosmological constant as the brane tension.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
156
+ page_content=' The DBI model proposed here is then an extension of the standard model of cosmology, and as such follows a late accelerated expansion but with a physical interpretation of the origin of Λ as a brane tension.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
157
+ page_content=' Using the same methodology and similar notations as in [22], we derive the field ϕ equation of motion from action (15).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
158
+ page_content=' This leads to a system of two ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
159
+ page_content='Instability of the cosmological DBI-Galileon in the non-relativistic limit ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
160
+ page_content='6 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
161
+ page_content='coupled equations ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
162
+ page_content='˜x′ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
163
+ page_content='= −˜x + αλ − σγ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
164
+ page_content='σβ − αω ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
165
+ page_content='¯H′ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
166
+ page_content='= ωγ − βλ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
167
+ page_content='σβ − αω ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
168
+ page_content='(19) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
169
+ page_content='with ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
170
+ page_content='β = 2 ¯H4 + 9 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
171
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
172
+ page_content='¯H6˜x2 − Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
173
+ page_content='Λ ¯H2 − 2η ¯H4˜x + 4ξ ¯H6˜x + 40 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
174
+ page_content='3 ξ ¯H8˜x3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
175
+ page_content='α = 6 ¯H3˜x − Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
176
+ page_content='Λ ¯H˜x + 15 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
177
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
178
+ page_content='¯H5˜x3 − 3η ¯H3˜x2 + 10ξ ¯H5˜x2 + 70 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
179
+ page_content='3 ξ ¯H7˜x4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
180
+ page_content='γ = 4 ¯H4˜x − 2Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
181
+ page_content='Λ ¯H2˜x − η ¯H4˜x2 + 2ξ ¯H6˜x2 − 10 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
182
+ page_content='3 ξ ¯H8˜x4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
183
+ page_content='ω = 4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
184
+ page_content='3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
185
+ page_content='¯H4˜x + ¯H6˜x3 − 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
186
+ page_content='3η ¯H4x2 + 2ξ ¯H6˜x2 + 10 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
187
+ page_content='3 ξ ¯H8˜x4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
188
+ page_content='σ = 2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
189
+ page_content='3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
190
+ page_content='¯H + 2 ¯H3˜x2 + 15 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
191
+ page_content='12 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
192
+ page_content='¯H5˜x4 − 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
193
+ page_content='3η ¯H3˜x3 + 10 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
194
+ page_content='3 ξ ¯H5˜x3 + 14 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
195
+ page_content='3 ξ ¯H7˜x5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
196
+ page_content='λ = ¯H2 + Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
197
+ page_content='r ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
198
+ page_content='3a4 − Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
199
+ page_content='Λ + 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
200
+ page_content='2Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
201
+ page_content='Λ ¯H2˜x2 − 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
202
+ page_content='3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
203
+ page_content='¯H4˜x2 − 5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
204
+ page_content='8 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
205
+ page_content='¯H6˜x4 + 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
206
+ page_content='3η ¯H4˜x3 − 2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
207
+ page_content='3ξ ¯H6˜x3 − 2ξ ¯H8˜x5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
208
+ page_content='Given values for the parameters Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
209
+ page_content='m,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
210
+ page_content=' η,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
211
+ page_content=' ξ and κ,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
212
+ page_content=' and initial conditions ˜x0 and H0,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
213
+ page_content=' this system can be integrated to compute background cosmology observables like the distance moduli of type Ia supernovae.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
214
+ page_content=' In Figure 1 we illustrate this with a Hubble diagram prediction compared with recent type Ia supernova data [23].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
215
+ page_content=' As an initial condition for ˜x, we chose to set ˜x0 = 6 × 10−8 today.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
216
+ page_content=' This is the maximum value allowed by the constraint on gravitational wave speed (see Section 4) coming from the quasi simultaneous observation of photons and gravitational waves after neutron star merger event GW170817A [24].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
217
+ page_content=' Nevertheless, before discussing more the cosmological scenarios proposed by the DBI-Galileon model, stability conditions much be computed first to assess the viability of the models for any set of parameters at the perturbation level.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
218
+ page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
219
+ page_content=' Stability conditions To be viable as a description of our Universe, the model has to fulfill stability conditions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
220
+ page_content=' These requirements apply to degrees of freedom propagating around the fixed background, i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
221
+ page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
222
+ page_content=' to the cosmological perturbations, that are capable of undermining the stability of the Universe.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
223
+ page_content=' In the determination of the stability conditions for the DBI-Galileon model in the non-relativistic limit, we use the formalism described in [25] for Horndeski theories.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
224
+ page_content=' In our case, these stability conditions are defined Instability of the cosmological DBI-Galileon in the non-relativistic limit 7 14 16 18 20 22 24 26 Distance moduli [mag] 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
225
+ page_content='00 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
226
+ page_content='25 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
227
+ page_content='50 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
228
+ page_content='75 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
229
+ page_content='00 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
230
+ page_content='25 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
231
+ page_content='50 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
232
+ page_content='75 Redshift z 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
233
+ page_content='2 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
234
+ page_content='0 Residuals [mag] Figure 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
235
+ page_content=' Hubble diagram prediction for the non-relativistic DBI-Galileon model for Ω0 Λ = 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
236
+ page_content='7, ˜x0 = 6 × 10−8, η = ξ = 0 (blue) compared with binned Pantheon data (black points) [23].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
237
+ page_content=' We used M = 23.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
238
+ page_content='81 for the offset magnitude of the diagram.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
239
+ page_content=' Residuals to the fit are presented in the bottom panel.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
240
+ page_content=' ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
241
+ page_content='from the following quantities derived from the particular Horndeski functions (9) to (12): ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
242
+ page_content='ω1 = 1 + 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
243
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
244
+ page_content='¯H2˜x2 + 3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
245
+ page_content='8 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
246
+ page_content='¯H4˜x4 + 2ξ ¯H4˜x3 + 2ξ ¯H6˜x5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
247
+ page_content='(20) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
248
+ page_content='ω2 = 2 ¯H + 3 ¯H3˜x2 + 15 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
249
+ page_content='4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
250
+ page_content='¯H5˜x4 − η ¯H3˜x3 + 10ξ ¯H5˜x3 + 14ξ ¯H7˜x5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
251
+ page_content='(21) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
252
+ page_content='ω3 = 9 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
253
+ page_content='� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
254
+ page_content='− ¯H2 + 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
255
+ page_content='2Ω0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
256
+ page_content='Λ ¯H2˜x2 − 3 ¯H4˜x2 + 2η ¯H4˜x3 − 10ξ ¯H6˜x3 − 45 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
257
+ page_content='8 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
258
+ page_content='¯H6˜x4 − 56 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
259
+ page_content='3 ξ ¯H8˜x5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
260
+ page_content='� ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
261
+ page_content='(22) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
262
+ page_content='ω4 = 1 − 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
263
+ page_content='2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
264
+ page_content='¯H2˜x2 − 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
265
+ page_content='8 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
266
+ page_content='¯H4˜x4 + 2ξ ¯H3˜x2( ¯H˜x)′ + 2ξ ¯H5˜x4( ¯H˜x)′ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
267
+ page_content='(23) ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
268
+ page_content='Tensorial stability conditions ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
269
+ page_content='In order to avoid ghosts and Laplacian instabilities,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
270
+ page_content=' we impose the following constraints on the sign of the kinetic term and on the sign of the gravitational waves speed squared: Qt ≡ ω1 4 = 1 4 + 1 8 ¯H2˜x2 + 3 32 ¯H4˜x4 + 1 2ξ ¯H4˜x3 + 1 2ξ ¯H6˜x5 > 0 (24) c2 t ≡ ω4 ω1 = 1 − 1 2 ¯H2˜x2 − 1 8 ¯H4˜x4 + 2ξ ¯H3˜x2( ¯H˜x)′ + 2ξ ¯H5˜x4( ¯H˜x)′ 1 + 1 2 ¯H2˜x2 + 3 8 ¯H4˜x4 + 2ξ ¯H4˜x3 + 2ξ ¯H6˜x5 ≥ 0 (25) In particular,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
271
+ page_content=' we see that the gravitational wave speed depends on ˜x,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
272
+ page_content=' and tends to 1 when ˜x → 0 : 4Qt ≃ 1 + 2( ¯H˜x)2 (26) ct ≃ 1 − 1 2( ¯H˜x)2 (27) Instability of the cosmological DBI-Galileon in the non-relativistic limit 8 Given the very tight constraint on the speed of gravitational waves,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
273
+ page_content=' equal to the speed of light up to a ∼ 10−15 difference [24,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
274
+ page_content=' 26,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
275
+ page_content=' 27],' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
276
+ page_content=' this justifies a posteriori the relevance of the non-relativistic limit where ˜x ≪ 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
277
+ page_content=' Moreover, we see that tensorial perturbations are stable in this limit since Qt > 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
278
+ page_content=' Scalar stability conditions Similar stability conditions apply to the scalar degrees of freedom,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
279
+ page_content=' here including the scalar perturbations of matter components: Qs ≡ ω1 (4ω1ω3 + 9ω2 2) 3ω2 2 > 0 (28) c2 s ≡ 3 (2ω2 1ω2H − ω2 2ω4 + 4ω1ω2 ˙ω1 − 2ω2 1 ˙ω2) − 6ω2 1 � (1 + wi) ρi ω1 (4ω1ω3 + 9ω2 2) ≥ 0 (29) where wi and ρi are respectively the equation of state parameter and the energy density of the fluid i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
280
+ page_content=' and the sum runs over all the components of the Universe (here only pressureless matter and radiation).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
281
+ page_content=' At the lowest order in ˜x, we get: Qs ≃ 3 2(Ω0 Λ − ¯H2)˜x2 (30) c2 s ≃ 1 + 2 � η ¯H2˜x′ − ¯H ¯H′ − 2ξ ¯H4˜x′� 3 � Ω0 Λ − ¯H2� (31) With ˜x ≪ 1, a fit of the DBI-Galileon model to data leads to cosmological parameters close to the standard model ones: Ω0 m ≈ 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
282
+ page_content='3 and Ω0 Λ ≈ 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
283
+ page_content='7 [28].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
284
+ page_content=' Therefore, from the first Friedmann equation, we get Ω0 Λ < ¯H2 for all relevant models in agreement with cosmological observations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
285
+ page_content=' As Qs ≤ 0, the DBI-Galileon model contains scalar instabilities unless it reduces to GR.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
286
+ page_content=' One way to avoid this would be to add a spatial curvature to the metric, but with a strong energy density (at least ∼ 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
287
+ page_content='3) which is also excluded by observations [28].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
288
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
289
+ page_content=' Discussion Physical interpretation From the definition (28), we see that the dominant terms come G2 (giving the Ω0 Λ term) and G4 (giving the ¯H2 term).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
290
+ page_content=' The competition between the two terms leads to the ghost-like behaviour in a cosmological setting: Qs ≤ 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
291
+ page_content=' In other words, it is the result of the competition between the DBI and the Einstein-Hilbert terms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
292
+ page_content=' The DBI action will have the effect of stretching the brane towards an extremal surface, whereas the Einstein-Hilbert term on the brane will tend to make the brane contract on itself from the effect of curvature.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
293
+ page_content=' However, in the non-relativistic limit of the DBI-Galileon, the Einstein-Hilbert term destabilizes the scalar field perturbations and the stretching effect from the cosmological constant is not strong enough to counterbalance, leading to an instantaneous decay of the vacuum state.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
294
+ page_content=' Instability of the cosmological DBI-Galileon in the non-relativistic limit 9 Because the DBI-Galileon action is the most general one can find of a 4D probe brane in a 5D bulk, we expect this statement to be quite general for all such theories studied in the current context.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
295
+ page_content=' Indeed, this is true in a standard cosmological setting which is realised with an FLRW slicing of the bulk space-time (equivalent to an FLRW background on the brane).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
296
+ page_content=' Therefore, the only way to evade this ghostly behaviour in cosmology is to include the full relativistic dynamics of the theory (˜x ∼ 1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
297
+ page_content=' We have seen that, in this case, we expect significant deviations of the speed of gravitational waves ct from c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
298
+ page_content=' This is not a definitive impossibility though if the full DBI-Galileon is viewed as an effective theory valid only at cosmological scales for which the speed of gravitational waves has not been probed [29, 30].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
299
+ page_content=' Indeed, the constraint on the gravitational speed from the observation of GW170817 in coincidence with GRB170817A [24] is only valid on small scales probed by LIGO and Virgo.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
300
+ page_content=' A modification of the dispersion relation of gravitational waves at small scales from operators present in the UV complete theory could allow ct ̸= c on cosmological scales while being compatible with current astrophysical observations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
301
+ page_content=' Waiting for the next generation of gravitational wave interferometers, in particular LISA, which will be able to probe this relation at larger scales [31], this possibility remains open.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
302
+ page_content=' Direct coupling to matter In the context of cosmology, where standard model matter is present, there might be direct coupling to the scalar field.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
303
+ page_content=' In that case, the metric ˜q to which matter is sensitive is different than the space-time metric q: S = � dx4√−q (Lf + LK + LR + LKGB) + � dx4� −˜qLm (˜qµν, ψm) .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
304
+ page_content=' (32) It has been shown in [32] that the two metrics are related by a disformal transformation of the following form: qµν = A � ϕ, ˜X � ˜qµν + B � ϕ, ˜X � ∂µϕ∂νϕ f 4 (33) where A and B are arbitrary functions of the scalar field and ˜X = −˜qµν∂µϕ∂νϕ/2f 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
305
+ page_content=' For simplicity and following the treatment of the covariant Galileon [22], we assume that A and B are constant parameters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
306
+ page_content=' This can be further justified by the fact that, a dependency on X would introduce, in general, higher order terms which would go beyond the framework of Horndeski theories [33], and a dependency on ϕ would, in general, break the shift symmetry followed by the scalar field ϕ in the probe brane context.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
307
+ page_content=' Note that, when A = −B, matter is coupled to the induced metric on the brane.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
308
+ page_content=' Contrary to the covariant Galileon, the DBI-Galileon action is not invariant by such a change of reference frame.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
309
+ page_content=' However, new terms that can not be absorbed into a redefinition of the parameters arise only at higher order in ˜X.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
310
+ page_content=' Therefore, the non- relativistic dynamics is not change by the introduction of a direct coupling between the Instability of the cosmological DBI-Galileon in the non-relativistic limit 10 scalar field and matter of the form (33) with constant parameters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
311
+ page_content=' In particular, this does not prevent the perturbations around the FLRW background from showing ghost instabilities.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
312
+ page_content=' Generalization The DBI-Galileon is a particular example of the more general class of shift-symmetric Horndeski theories.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
313
+ page_content=' These are subclass of Horndeski theories which are invariant under a shift symmetry of the scalar field ϕ → ϕ + c [34, 35].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
314
+ page_content=' In these theories, the arbitrary Horndeski functions are restricted to be functions of X alone.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
315
+ page_content=' In order to make the non-relativistic limit apparent, we Taylor expand these arbitrary functions around GR: G2 ≡ Λ + +∞ � n=1 g(n) 2 Xn (34) G3 ≡ +∞ � n=1 g(n) 3 Xn (35) G4 ≡ M 2 P 2 + +∞ � n=1 g(n) 4 Xn (36) G5 ≡ +∞ � n=1 g(n) 5 Xn (37) The constant terms in G3 and G5 do not appear in the expansion as they lead to total derivative terms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
316
+ page_content=' Because the Horndeski functions depend only on X,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
317
+ page_content=' the ω functions that determine the stability conditions reduce to: ω1 ≡ 2G4 − 2X � 2G4,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
318
+ page_content='X + ˙φHG5,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
319
+ page_content='X � (38) ω2 ≡ 4HG4 − 2X � ˙φG3,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
320
+ page_content='X + 8HG4,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
321
+ page_content='X + 5 ˙φH2G5,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
322
+ page_content='X � − 4X2H � 4G4,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
323
+ page_content='XX + ˙φHG5,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
324
+ page_content='XX � (39) ω3 ≡ − 18H2G4 + 3X � G2,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
325
+ page_content='X + 12 ˙φHG3,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
326
+ page_content='X + 42H2G4,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
327
+ page_content='X + 30 ˙φH3G5,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
328
+ page_content='X � + 6X2 � G2,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
329
+ page_content='XX + 3 ˙φHG3,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
330
+ page_content='XX + 48H2G4,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
331
+ page_content='XX + 13H3 ˙φG5,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
332
+ page_content='XX � + 12X3H2 � 6G4,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
333
+ page_content='XXX + H ˙φG5,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
334
+ page_content='XXX � (40) ω4 ≡ 2G4 − 2X ¨φG5,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
335
+ page_content='X (41) From these,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
336
+ page_content=' we can compute the quantity Qs up to first order in X: Qs = X H2 � g(1) 2 + 6H2g(1) 4 � + O � X 3 2 � (42) This leads to a very simple formulation of the no-ghost condition,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
337
+ page_content=' independent of X,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
338
+ page_content=' in the context of Shift-Symmetric Horndeski theories in the non-relativistic limit: g(1) 2 + 6H2g(1) 4 > 0 (43) Instability of the cosmological DBI-Galileon in the non-relativistic limit 11 In the context of the brane galileon,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
339
+ page_content=' where g(1) 4 = −M 2 P/2 and g(1) 2 = Λ,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
340
+ page_content=' this is equivalent to the inequality which is never fulfilled in flat space: Λ − 3M 2 PH2 > 0 ⇔ Ω0 Λ > ¯H2 (44) Other stability conditions are given by: c2 s ≃ 1 + 2¨φg(1) 3 + 4 ˙Hg(1) 4 + 2¨φH2g(1) 5 g(1) 2 + 6H2g(1) 4 > 0 (45) Qt ≃ M 2 P 4 (46) c2 t ≃ 1 (47) where we expressed these quantities at the lowest order.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
341
+ page_content=' The two tensorial conditions are, thus, automatically satisfied in this context.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
342
+ page_content=' On the other hand, the stability conditions for scalar perturbations at the lowest order give a simple inequality involving the parameters of the Taylor expansion, that can be easily checked at the background level.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
343
+ page_content=' 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
344
+ page_content=' Conclusion We described the DBI-Galileon theory of a four-dimensional brane evolving in a 5D bulk space-time in the non-relativistic limit where its local kinetic energy is small compared to its tension.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
345
+ page_content=' This model belongs to the class of shift-symmetric Horndeski theories, themselves being a subclass of the more general family of Horndeski theories.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
346
+ page_content=' From the construction of the DBI-Galileon model, the free parameters of the model acquire a physical meaning.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
347
+ page_content=' In particular, the interpretation of the cosmological constant is linked to the brane tension energy density.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
348
+ page_content=' We derived the equations driving the evolution of the late-time Universe around a spatially flat FLRW cosmological background and studied the stability of scalar and tensorial perturbations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
349
+ page_content=' This model reduces to an expansion around standard GR, and therefore around standard ΛCDM in the cosmological context.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
350
+ page_content=' As such, it is naturally compatible with data in the non-relativistic limit provided the effect of the scalar field is small enough, even considering the speed of the gravitational waves.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
351
+ page_content=' However, it revealed fatal ghostly behaviour for scalar perturbations around the FLRW background.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
352
+ page_content=' From there, we derived the corresponding stability conditions for shift-symmetric Horndeski theories in the non-relativistic limit in the cosmological context and found very simple formulations for these conditions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
353
+ page_content=' Acknowledgements We would like to thank Marc Besan¸con, Arnaud de Mattia and Vanina Ruhlmann- Kleider for their comments on the present paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
354
+ page_content=' We also want to thank David Langlois for useful and interesting comments and suggestions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
355
+ page_content=' Instability of the cosmological DBI-Galileon in the non-relativistic limit 12 References [1] Brax P 2017 Reports on Progress in Physics 81 016902 URL https://dx.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
356
+ page_content='doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
357
+ page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
358
+ page_content='1088/ 1361-6633/aa8e64 [2] Horndeski G W 1974 Int.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
359
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
360
+ page_content=' Theor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
361
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
362
+ page_content=' 10 363–384 [3] Deffayet C, Gao X, Steer D A and Zahariade G 2011 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
363
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
364
+ page_content=' D 84 064039 (Preprint 1103.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
365
+ page_content='3260) [4] Langlois D and Noui K 2016 JCAP 07 016 (Preprint 1512.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
366
+ page_content='06820) [5] Langlois D 2019 Int.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
367
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
368
+ page_content=' Mod.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
369
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
370
+ page_content=' D 28 1942006 (Preprint 1811.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
371
+ page_content='06271) [6] Nicolis A, Rattazzi R and Trincherini E 2009 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
372
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
373
+ page_content=' D 79 064036 (Preprint 0811.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
374
+ page_content='2197) [7] Deffayet C, Esposito-Farese G and Vikman A 2009 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
375
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
376
+ page_content=' D 79 084003 (Preprint 0901.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
377
+ page_content='1314) [8] Gubitosi G and Linder E V 2011 Physics Letters B 703 113–118 ISSN 0370-2693 URL https: //www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
378
+ page_content='sciencedirect.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
379
+ page_content='com/science/article/pii/S0370269311008707 [9] de Rham C and Gabadadze G 2010 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
380
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
381
+ page_content=' D 82 044020 (Preprint 1007.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
382
+ page_content='0443) [10] de Rham C, Gabadadze G and Tolley A J 2011 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
383
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
384
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
385
+ page_content=' 106 231101 (Preprint 1011.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
386
+ page_content='1232) [11] de Rham C and Tolley A J 2010 JCAP 05 015 (Preprint 1003.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
387
+ page_content='5917) [12] Hinterbichler K, Trodden M and Wesley D 2010 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
388
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
389
+ page_content=' D 82 124018 (Preprint 1008.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
390
+ page_content='1305) [13] Silverstein E and Tong D 2004 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
391
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
392
+ page_content=' D 70 103505 (Preprint hep-th/0310221) [14] Langlois D, Renaux-Petel S, Steer D A and Tanaka T 2008 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
393
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
394
+ page_content=' D 78 063523 (Preprint 0806.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
395
+ page_content='0336) [15] Zumalacarregui M, Koivisto T S and Mota D F 2013 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
396
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
397
+ page_content=' D 87 083010 (Preprint 1210.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
398
+ page_content='8016) [16] Goon G, Hinterbichler K, Joyce A and Trodden M 2012 JHEP 06 004 (Preprint 1203.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
399
+ page_content='3191) [17] Dobado A and Maroto A L 2001 Nucl.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
400
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
401
+ page_content=' B 592 203–218 (Preprint hep-ph/0007100) [18] Cembranos J A R, Dobado A and Maroto A L 2001 Goldstone bosons and solitons on the brane 20th International Symposium on Lepton and Photon Interactions at High Energies (LP 01) (Preprint hep-ph/0107155) [19] Burrage C, de Rham C and Heisenberg L 2011 JCAP 05 025 (Preprint 1104.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
402
+ page_content='0155) [20] Goon G, Hinterbichler K and Trodden M 2011 JCAP 07 017 (Preprint 1103.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
403
+ page_content='5745) [21] Lovelock D 1971 J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
404
+ page_content=' Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
405
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
406
+ page_content=' 12 498–501 [22] Appleby S and Linder E V 2012 Journal of Cosmology and Astroparticle Physics 2012 043 URL https://dx.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
407
+ page_content='doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
408
+ page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
409
+ page_content='1088/1475-7516/2012/03/043 [23] Scolnic D M et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
410
+ page_content=' 2018 The Astrophysical Journal 859 101 ISSN 15384357 (Preprint 1710.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
411
+ page_content='00845) URL http://arxiv.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
412
+ page_content='org/abs/1710.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
413
+ page_content='00845http://dx.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
414
+ page_content='doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
415
+ page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
416
+ page_content='17909/T95Q4X [24] Abbott B P et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
417
+ page_content=' (LIGO Scientific, Virgo, Fermi-GBM, INTEGRAL) 2017 Astrophys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
418
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
419
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
420
+ page_content=' 848 L13 (Preprint 1710.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
421
+ page_content='05834) [25] De Felice A and Tsujikawa S 2012 JCAP 02 007 (Preprint 1110.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
422
+ page_content='3878) [26] Creminelli P and Vernizzi F 2017 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
423
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
424
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
425
+ page_content=' 119 251302 (Preprint 1710.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
426
+ page_content='05877) [27] Ezquiaga J M and Zumalac´arregui M 2017 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
427
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
428
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
429
+ page_content=' 119 251304 (Preprint 1710.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
430
+ page_content='05901) [28] Planck Collaboration 2020 A&A 641 A6 URL https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
431
+ page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
432
+ page_content='1051/0004-6361/201833910 [29] de Rham C and Melville S 2018 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
433
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
434
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
435
+ page_content=' 121 221101 (Preprint 1806.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
436
+ page_content='09417) [30] Ezquiaga J M and Zumalac´arregui M 2018 Front.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
437
+ page_content=' Astron.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
438
+ page_content=' Space Sci.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
439
+ page_content=' 5 44 (Preprint 1807.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
440
+ page_content='09241) [31] Barausse E et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
441
+ page_content=' 2020 Gen.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
442
+ page_content=' Rel.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
443
+ page_content=' Grav.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
444
+ page_content=' 52 81 (Preprint 2001.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
445
+ page_content='09793) [32] Bekenstein J D 1993 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
446
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
447
+ page_content=' D 48 3641–3647 (Preprint gr-qc/9211017) [33] Bettoni D and Liberati S 2013 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
448
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
449
+ page_content=' D 88 084020 (Preprint 1306.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
450
+ page_content='6724) [34] Sotiriou T P and Zhou S Y 2014 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
451
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
452
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
453
+ page_content=' 112 251102 (Preprint 1312.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
454
+ page_content='3622) [35] Sotiriou T P and Zhou S Y 2014 Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
455
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
456
+ page_content=' D 90 124063 (Preprint 1408.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
457
+ page_content='1698)' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/C9AzT4oBgHgl3EQfwf5z/content/2301.01723v1.pdf'}
CdE2T4oBgHgl3EQf9AmF/vector_store/index.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1615a59309dd5ac5f8ba4473c2a7efd6e34590dcaf0d4319b61f1166c84d59b
3
+ size 195964
DNFQT4oBgHgl3EQfPTY7/content/tmp_files/2301.13278v1.pdf.txt ADDED
@@ -0,0 +1,1485 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ The younger flagellum coordinates the beating in
2
+ C. reinhardtii
3
+ Da Wei1,3,Greta Quaranta2, Marie-Eve Aubin-Tam1†, Daniel S.W. Tam2∗
4
+ 1Department of Bionanoscience, Delft University of Technology,
5
+ 2628CJ Delft, Netherlands.
6
+ 2Laboratory for Aero and Hydrodynamics, Delft University of Technology,
7
+ 2628CD Delft, Netherlands.
8
+ 3Beijing National Laboratory for Condensed Matter Physics, Institute of Physics,
9
+ Chinese Academy of Sciences; Beijing 100190, China.
10
+ †Corresponding author. Email: [email protected];
11
+ ∗Corresponding author. Email: [email protected].
12
+ 1
13
+ arXiv:2301.13278v1 [physics.bio-ph] 30 Jan 2023
14
+
15
+ Abstract
16
+ Eukaryotes swim with coordinated flagellar (ciliary) beating and steer by fine-tuning
17
+ the coordination. The model organism for studying flagellate motility, C. reinhardtii (CR),
18
+ employs synchronous, breast-stroke-like flagellar beating to swim, and it modulates the
19
+ beating amplitudes differentially to steer. This strategy hinges on both inherent flagellar
20
+ asymmetries (e.g. different response to chemical messengers) and such asymmetries be-
21
+ ing effectively coordinated in the synchronous beating. In CR, the synchrony of beating is
22
+ known to be supported by a mechanical connection between flagella, however, how flagellar
23
+ asymmetries persist in the synchrony remains elusive. For example, it has been speculated
24
+ for decades that one flagellum leads the beating, as its dynamic properties (i.e. frequency,
25
+ waveform, etc.) appear to be copied by the other one. In this study, we combine experi-
26
+ ments, computations, and modeling efforts to elucidate the roles played by each flagellum
27
+ in synchronous beating. With a non-invasive technique to selectively load each flagellum,
28
+ we show that the coordinated beating essentially responds to only load exerted on the cis
29
+ flagellum; and that such asymmetry in response derives from a unilateral coupling between
30
+ the two flagella. Our results highlight a distinct role for each flagellum in coordination and
31
+ have implication for biflagellates’ tactic behaviors.
32
+ One-Sentence Summary: The younger flagellum of C. reinhardtii coordinates the synchronous
33
+ beating and couples to external forces.
34
+ 2
35
+
36
+ Introduction
37
+ The ability to swim towards desirable environments and away from hazardous ones is funda-
38
+ mental to the survival of many microorganisms. These so-called tactic behaviors are exhib-
39
+ ited by many motile microorganisms ranging from bacteria (1, 2) to larger flagellates and cili-
40
+ ates (3–5). Different microorganisms have developed specific strategies for steering, depending
41
+ on the tactic behavior and on their specific sensory and motility repertoire. For example, bacte-
42
+ ria modulate the tumbling rate (1) while flagellates and ciliates modulate the waveform (6–9),
43
+ amplitude (10, 11) and frequency of their flagellar/ciliary (4, 12) beating. The goal of these
44
+ active modulations of the motility is to achieve a spatially asymmetric generation of propulsive
45
+ force to steer the organism.
46
+ C. reinhardtii (CR), the model organism for studies of flagellar motility, achieves tactic nav-
47
+ igation by a fine-tuned differential modulation on its two flagella. Studying this organism offers
48
+ great opportunities to look into how flagella coordinate with each other and how such coordi-
49
+ nation helps facilitate targeted steering. CR has a symmetric cell body and two near-identical
50
+ flagella inherited from the common ancestors of land plants and animals (13). It swims by
51
+ beating its two flagella synchronously and is capable of photo- and chemotaxis (10, 14). For
52
+ this biflagellated organism, effective steering hinges on both flagellar asymmetry and flagellar
53
+ coordination. On the one hand, the two flagella must be asymmetric to respond differentially
54
+ to stimuli (10,15); on the other hand, the differential responses must be coordinated by the cell
55
+ such that the beating would remain synchronized to guarantee effective swimming. Understand-
56
+ ing this remarkable feat requires knowledge about both flagellar asymmetry and coordination.
57
+ The two flagella are known to be asymmetric in several, possibly associated, aspects. First
58
+ of all, they differ in developmental age (16, 17). The flagellum closer to the eyespot, the cis(-
59
+ eyespot) flagellum, is always younger than the other one, the trans(-eyespot) flagellum. This
60
+ is because the cis is organized by a basal body (BB) that develops from a pre-matured one in
61
+ the mother cell; and this younger BB also organizes the flagellar root (D4 rootlet) that dictates
62
+ the eyespot formation (18). Second, the two flagella have asymmetric protein composition (19–
63
+ 21). For example, the trans flagellum is richer in CAH6, a protein possibly involved in CO2
64
+ sensing (14,20). Finally, the flagella have different dynamic properties (22–24). Their beating is
65
+ modulated differentially by second messengers such as calcium (22,23) and cAMP (25). When
66
+ beating alone, the trans beats at a frequency 30%-40% higher than the cis (23,26–28); the trans
67
+ also displays an attenuated waveform (29) and a much stronger noise (29,30).
68
+ 3
69
+
70
+ Remarkably, despite these inherent asymmetries, CR cells establish robust synchronization
71
+ between the flagella. Such coordination enables efficient swimming and steering of the cells
72
+ and takes basis on the fibrous connections between flagellar bases (31,32). Intriguingly, in the
73
+ coordinated beating, both flagella display dynamic properties, i.e., flagellar waveform, beating
74
+ frequency (∼50 Hz), and frequency fluctuation, that are more similar to those of the cis flag-
75
+ ellum (26, 28–30, 33). This has led to a long-standing hypothesis that ”the cis somehow tunes
76
+ the trans flagellum” (26). This implies that the symmetric flagellar beating (”breast-stroke”)
77
+ observed is the result of interactions between two flagella playing differential roles in coordi-
78
+ nation. How does the basal coupling make this possible? Recent theoretical efforts show that
79
+ the basal coupling can give rise to different synchronization modes (34–36); and that flagellar
80
+ dynamics, such as beating frequency, may simply emerge from the interplay between mechan-
81
+ ics of basal coupling and bio-activity (36). Yet, most theoretical efforts examining flagellar
82
+ synchronization have assumed two identical flagella, limiting the results’ implication for the
83
+ realistic case. Moreover, little experiments directly probe the flagella’s differential roles during
84
+ synchronous beating (37). Therefore, flagellar coordination in this model organism remains un-
85
+ clear. To clarify the picture experimentally, one needs to selectively force each flagellum, and
86
+ characterize the dynamics of the flagellar response.
87
+ In this study, we address this challenge and devise a non-invasive approach to apply external
88
+ forces selectively on the cis- or the trans-flagella. Oscillatory background flows are imposed
89
+ along an angle with respect to the cell’s symmetry axis. Such flows result in controlled hydro-
90
+ dynamic forces, which are markedly different on the two flagella. With experiments, hydrody-
91
+ namic computations, and modeling, we show definitively that the two flagella are unilaterally
92
+ coupled, such that the younger flagellum (cis) coordinates the beating, whereas the elder one
93
+ simply copies the dynamic properties of the younger. This also means that only external forces
94
+ on the cis may mechanically fine-tune the coordination. We also study the effect of calcium
95
+ in the cis’ leading role as calcium is deeply involved in flagellar asymmetry and hence photo-
96
+ tactic steering. In addition, a well-known mutant that lacks flagellar dominance (ptx1) (23,38)
97
+ is examined. Results show that the coordinating role of cis does not need environmental free
98
+ calcium, whereas it does require the genes lost in ptx1. Our results discern the differential roles
99
+ of CR’s flagella, highlight an advanced function of the inter-flagellar mechanical coupling, and
100
+ have implications for biflagellates’ tactic motility.
101
+ 4
102
+
103
+ Experimental scheme for selective loading
104
+ We set out to establish a non-invasive experimental technique that exerts differential loads on the
105
+ flagella of CR. Following the study by Quaranta et al. (31), we induce oscillatory background
106
+ flows to exert hydrodynamic forcing to flagella of captured cells. With programmed oscillations
107
+ of the piezoelectric stage, the amplitude, frequency, and direction of the background flows are
108
+ all controlled, enabling selective loading.
109
+ To quantitatively estimate the selectivity of the flows along different angles (θ), we compute
110
+ the flagellar loads under the flows along θ = −45◦, 0◦, and 45◦, see Fig. 1A. Computations
111
+ based on boundary element methods (BEM) and slender-body theory (SBT) give the real-time
112
+ drag force F on each flagellum and the power P exerted by the viscous forces on each flagellum.
113
+ For given realistic flagellar shapes, we compare computed loads with and without external flows.
114
+ From these we isolate the loads from the induced flows FFlow and PFlow (Methods).
115
+ Loads on each flagellum under flows of θ = 0◦, −45◦, 45◦ are presented in Fig. 2. Upper
116
+ panels display the magnitude of the drag force FFlow = |FFlow|; while lower panels show viscous
117
+ power PFlow. Force magnitudes are scaled by F0 = 6πµRU0 = 9.9 pN; while the powers by
118
+ P0 = F0U0 = 1.1 fW. F0 is the Stokes drag on a typical free-swimming cell (radius R = 5 µm,
119
+ speed U0 = 110 µm/s, water viscosity µ = 0.95 mPa·s).
120
+ Evidently, along θ = 0◦, flows load the flagella equally (Fig. 2A). However, at θ = −45◦,
121
+ flows load the cis flagellum ∼ 2 times larger than the trans (Fig. 2B, F c
122
+ Flow ≈ 2F t
123
+ Flow); whereas
124
+ flows at θ = 45◦ do the opposite (Fig. 2C). The selectivity also manifests in (the absolute values
125
+ of) PFlow. We do notice that flows along θ = +45◦ are able to synchronize the flagella with
126
+ PFlow < 0, meaning that the flagella are working against the flows, and this shall be discussed
127
+ in later sections.
128
+ Hereon forward, we refer to θc-flows, flows for which θ = −45◦ and the cis-flagellum is
129
+ selectively loaded. Likewise, θt-flows denote flows on θ = +45◦ that selectively load the trans.
130
+ θa-flows denote the axial flow along θ = 0◦. We next introduce how we quantify the flows’
131
+ effective forcing strength (ε) on the cell.
132
+ Phase dynamics of flagellar beating is extracted from videography (31,39,40). Recordings
133
+ are masked and thresholded to highlight the flagella (Fig. 1B-C). Then the mean pixel values
134
+ over time within two sampling windows (Fig. 1D) are converted to observable-invariant flagellar
135
+ phases (41), Fig. 1E. Throughout this study, as cis and trans always beat synchronously (Fig. 1E
136
+ inset), their phases ϕc,t are used interchangeably as the flagellar phase ϕ. The flagellar phase
137
+ 5
138
+
139
+ dynamics under external periodic forcing is described by Adler equation (42–44):
140
+ d∆ϕ
141
+ dt
142
+ = −2πν − 2πε sin(∆ϕ) + ζ(t).
143
+ (1)
144
+ ∆ϕ = ϕ − 2πfft is the phase difference between the beating and the forcing, with ff the
145
+ forcing frequency, and ε the forcing strength. The detuning ν = ff − f0 is the frequency
146
+ mismatch between the beating (f0) and forcing. ζ(t) represents a white noise that satisfies
147
+ ⟨ζ(τ + t)ζ(τ)⟩ = 2Teffδ(t), with Teff an effective temperature and δ(t) the Dirac delta function.
148
+ When the forcing strength outweighs the detuning (ε > |ν|), synchronization with the flow
149
+ (d∆ϕ/dt = 0) emerges, see the plateaus marked black in Fig. 1F. We characterize synchro-
150
+ nization with τ = tsync/ttot, where tsync is the total time of flow synchronization and ttot the
151
+ flow duration. Fig. 1F presents the phase dynamics which are representative and range from:
152
+ no synchronization (τ=0, i), unstable synchronization (0 < τ < 1, ii-iii), and stable synchro-
153
+ nization (τ=1, iv). In this study, the frequency range in ν for which τ ≥ 0.5 is used to measure
154
+ ε (see Fig. 1F inset). This method is equivalent to previous fitting-based methods (28, 31), see
155
+ SM. Sec.S1.
156
+ Asymmetric susceptibility to flow synchronization
157
+ Now we examine cell responses to flows of various amplitudes and along different directions.
158
+ First we explore flow synchronization over a broad range of amplitudes and frequencies. θa-
159
+ flows with frequencies ff ∈ [40, 75] Hz and amplitudes U ∈ [390, 2340] µm/s are imposed. The
160
+ scanned range covers reported intrinsic frequencies of both the cis and trans flagellum (22,24,
161
+ 26, 27); while the amplitude reaches the maximum instantaneous speed of a beating flagellum
162
+ (∼ 2000 µm/s). Fig. 3A displays the resultant flow-synchronized time fractions τ. Up until the
163
+ strongest flow amplitude, the large forces cannot disrupt the synchronized flagellar beating. In
164
+ addition, synchronization is never established around frequencies other than f0. This shows that
165
+ the inter-flagella coupling is much stronger than the maximum amplitude of forcing.
166
+ Next we examine the synchronization with the θc-flows and θt-flows. Flows of a fixed
167
+ amplitude (∼ 7U0) but varying frequencies around f0 are applied to each captured cell (see
168
+ Methods). With these, the flow-synchronized time fraction τ as a function of the detuning (ν)
169
+ and flow direction (θc,a,t) is recorded and helps quantify the flows’ effective forcing ε(θ).
170
+ Comparing τ(ν; θc) to τ(ν; θt), with τ(ν; θa) as reference, we find that θc-flows are the most
171
+ effective in synchronizing the beating (Fig. 3B). We illustrate this point with the profiles of an
172
+ 6
173
+
174
+ exemplary cell (Fig. 3B inset). First, although both the θc-flow (red) and the θt-flow (blue)
175
+ can synchronize the cell at small detunings (|ν| <0.5Hz), the θc-flow maintains the synchro-
176
+ nization for the whole time ( τ(θc) =1), while the θt-flow for a slightly smaller time fraction
177
+ ( τ(θc) ≈0.85). This is due to phase-slips (step-like changes in ∆ϕ(t) in Fig. 1F) between flag-
178
+ ella and the flow, and means that the θt-flow synchronization is less stable. Additionally, for
179
+ intermediate detuning (0.5 Hz< |ν| <4 Hz), τ(θc) is always larger than τ(θt) . In some cases,
180
+ the θc-flow synchronizes the cell fully whereas the θt-flow fails completely (e.g., at ν = −2
181
+ Hz). Together, these results imply that a flow of given amplitude synchronizes flagellar beating
182
+ more effectively if it selectively loads the cis.
183
+ We repeat the experiments with cells from multiple cultures, captured on different pipettes,
184
+ and with different eyespot orientations (∼50% heading rightward in the imaging plane) to rule
185
+ out possible influence from the setup. τ(ν; θ) of N=11 wt cells tested in the TRIS-minimal
186
+ medium (pH=7.0) are displayed in Fig. 3B (labeled as ”TRIS”). On average, ε(θc) = 2.9 Hz
187
+ and is 70% larger than ε(θt) = 1.7 Hz. It bears emphasis that ε(θc) > ε(θt) holds true for every
188
+ single cell tested (11/11). In Fig. 3C, we show this by representing each cell as a point in the
189
+ ε(θc) - ε(θt) plane. A point being below the first bisector line ( ε(θc) = ε(θt) ) indicates that
190
+ ε(θc) > ε(θt) for this cell. All cells cluster clearly below the line. This asymmetry manifest
191
+ equivalently through τ. In Fig. 3D, each point represents the time fractions of the same cell
192
+ synchronized by the θc-flow and the θt-flow at the same frequency. Most points (>90%) are
193
+ below the first bisector line, meaning that τ(θc) > τ(θt) . Altogether, all results show that
194
+ selectively loading the cis flagellum establishes synchronization with the flow more effectively,
195
+ pointing to cis and trans playing differential roles in the coordinated beating.
196
+ We next study whether this newly observed cis-trans asymmetry is affected by calcium
197
+ depletion. Calcium is a critical second messenger for modulating flagellates motility and is
198
+ deeply involved in phototaxis (45). The depletion of the free environmental calcium is known
199
+ to degrade flagellar synchronization and exacerbate flagellar asymmetry (22). Here we focus
200
+ on whether calcium depletion affects the asymmetry ε(θc) > ε(θt) . We deplete environmental
201
+ calcium by EGTA-chelation, following the protocol in Ref. (46). Similar to previous reports (22,
202
+ 47), the number of freely swimming cells drops significantly in EGTA-containing medium.
203
+ However, the remaining cells beat synchronously for hours after capture. For these beating cells,
204
+ calcium depletion is first confirmed by characterizing their deflagellation behavior. Indeed,
205
+ calcium depletion is reported to inhibit deflagellation (28, 48). In experiments with standard
206
+ calcium concentration, all cells deflagellated under pipette suction (20/20). For experiments
207
+ 7
208
+
209
+ conducted in calcium depleting EGTA-containing medium, we observe deflagellation to occur
210
+ in none but one cell (1/19).
211
+ After confirming the calcium depletion, we perform the same sets of flow synchronization
212
+ experiments. The dashed lines in Fig. 3B show the median synchronization profiles τ(ν; θ)
213
+ (N=6 cells, labeled as ”EGTA”). The flagellar asymmetry is unaffected, see also Fig. 3E. Note
214
+ that ε(θc) > ε(θt) again applies for every single cell tested. The mean values of ε drop slightly.
215
+ However, the different effectiveness between θc-flows and θt-flows, ε(θc) − ε(θt) , is not af-
216
+ fected, see Fig. 3E inset.
217
+ Finally, we determine how the forcing strength of the flow depends on the hydrodynamic
218
+ forces exerted by the flow on the flagella. We compute the hydrodynamic beat-averaged loads,
219
+ F Flow =
220
+ � 2π
221
+ 0
222
+ FFlowdϕ/2π, P Flow =
223
+ � 2π
224
+ 0
225
+ PFlowdϕ/2π, induced by the flow on the trans and on
226
+ the cis flagella, see the horizontal lines in Fig. 2. These loads are computed for the θc-flow,
227
+ θt-flow, θa-flow and we also include experiments and computations performed with flows along
228
+ θ = 90◦ (circles), see SM. Sec.S2. Fig. 3F and G represent ε as a function of the loads on the
229
+ cis and trans flagellum respectively, with each symbol representing one of the four different
230
+ flow directions, see the drawings. We find that the effective forcing strength scales with the
231
+ time-averaged drag on the cis, ε ∼ F
232
+ c
233
+ Flow, while we find no such correlation between ε and
234
+ F
235
+ t
236
+ Flow. The linear relation between ε and F
237
+ c
238
+ Flow has an intercept near zero (ε|F c
239
+ Flow=0 ≈ 0).
240
+ Given the total forces on both flagella (F
241
+ c
242
+ Flow + F
243
+ t
244
+ Flow) for these flows remains almost constant
245
+ (0.74-0.79F0), the zero-intercept implies that for a hypothetical flow that exerts no load on the
246
+ cis but solely forces the trans, it will not be able to synchronize the cell at all. This suggests a
247
+ negligible contribution of the forcing on the trans in establishing synchronization with flows.
248
+ The asymmetry is lost in ptx1 mutants
249
+ Furthermore, we examine the flagellar dominance mutant ptx1. In this mutant, both flagella re-
250
+ spond similarly to changes of calcium concentrations (38) and have similar beating frequencies
251
+ when demembranated and reactivated (23).
252
+ Ptx1 mutants have two modes of coordinated beating, namely, the in-phase (IP) synchro-
253
+ nization and the anti-phase (AP) synchronization (29, 49). First, we apply θa-flow in the same
254
+ frequency and amplitude ranges as for wt cells. We find that the IP mode around f0 ≈ 50 Hz
255
+ is the only mode that can be synchronized by external flows. We focus on this mode and report
256
+ τ as τ = tsync/tIP for this mutant, where tIP is the total time of IP-beating under the applied
257
+ 8
258
+
259
+ flows, see Fig. 4A. Synchronization profiles τ(ν; θ) of ptx1 are shown in Fig. 4B. The median
260
+ profiles are of similar width and height, indistinguishable from each other, and hence indicate
261
+ a loss of asymmetric susceptibility to flow synchronization. The loss is further confirmed by
262
+ the extracted ε(θ) (31) and τ(θ) (Fig. 4C-D). Cells and synchronization attempts are distributed
263
+ evenly across the first bisector lines (7/14 cells are below ε(θc) = ε(θt) in Fig. 4C, and ∼50%
264
+ points are below τ(θc) = τ(θt) in Fig. 4D). Altogether, all results show consistently that the
265
+ asymmetry is lost in ptx1.
266
+ Modeling
267
+ Framework
268
+ To investigate the implications of our experimental results on the coupling between flagella
269
+ and their dynamics, we develop a model for the system (SM. Sec.S3), representing flagella and
270
+ external flows as oscillators with directional couplings:
271
+
272
+
273
+
274
+
275
+
276
+ ˙ϕf = 2πff
277
+ ˙ϕc = 2π[fc − λt sin(ϕc-ϕt) − εc sin(ϕc-ϕf)] + ζc(t)
278
+ ˙ϕt = 2π[ft − λc sin(ϕt-ϕc) − εt sin(ϕt-ϕf)] + ζt(t).
279
+ (2)
280
+ ϕf,c,t(t) respectively represent the phase of the flow, the cis, and the trans flagellum. ff,c,t
281
+ represents the inherent frequency of the forcing (flow), the cis, and the trans respectively. The
282
+ phase dynamics of each flagellum is modulated by its interactions with the other flagellum as
283
+ well as the background flow. Take the cis ( ˙ϕc) for example, the effect of the trans and the forcing
284
+ on the cis are respectively accounted for by the λt-term and the εc-term, see Eq. (2). In other
285
+ words, λt and εc measure the sensitivity of the actual cis-frequency to the phase differences
286
+ between oscillators (ϕc − ϕt,f), see the arrows in Fig. 5A. Lastly, ζc,t represent the white noise
287
+ of the cis and trans flagellum respectively. In the following parts, without loss of generality, the
288
+ noise are assumed equally strong and uncorrelated (⟨ζ2
289
+ c ⟩ = ⟨ζ2
290
+ t ⟩, or T c
291
+ eff = T t
292
+ eff). Nuanced phase
293
+ dynamics under differential noise levels can be found in SM. Sec.S4.
294
+ Eq. (2) can be readily reduced to Eq. (1), which allows us to write the experimentally mea-
295
+ sured values (f0, ε(θ), Teff) analytically with εc,t, λc,t, and ζc,t. The asymptotic behavior of the
296
+ 9
297
+
298
+ model under the condition ϕc ≈ ϕt ≈ ϕf are (SM. Sec.S3):
299
+
300
+
301
+
302
+
303
+
304
+ f0
305
+ = αfc + (1 − α)ft,
306
+ Teff
307
+ = α2T c
308
+ eff + (1 − α)2T t
309
+ eff,
310
+ ε(θ)
311
+ = αεc(θ) + (1 − α)εt(θ),
312
+ (3)
313
+ with α = λc/(λc + λt) representing the dominance of cis. It is then clear that when α ≈ 1, the
314
+ coordinated beating will display dynamic properties of the cis flagellum.
315
+ Fig. 5A illustrates an exemplary modeling scheme describing flagellar beating subjected
316
+ to θc-flows. The direction and thickness of arrows represent coupling direction and strength
317
+ respectively. The selective loading on the cis is represented by εc > εt; while λc > λt reflects
318
+ that the cis has a more dominant role in the coordinated beating. We run Monte-Carlo simulation
319
+ with Eq. (2) using customized MATLAB scripts.
320
+ Coordinated beating under symmetric forcing
321
+ We first model the flow synchronization induced by θa-flow (symmetric flagellar loads). In this
322
+ case, ε(θ) = αεc(θ) + (1 − α)εt(θ) (Eq. (3)) reduces to ε = εc,t and is independent of α. We
323
+ set εc,t as 2.4 Hz to match the measured ε(θa) =2.4 Hz (Fig. 3B).
324
+ At similar detunings as in the experimental results in Fig. 1F, our Monte-Carlo simulations
325
+ reproduces the phase dynamics with: (i) no flow synchronization, (ii-iii) unstable synchroniza-
326
+ tion, and (iv) stable synchronization (Fig. 5B). Repeating the simulations for varying forcing
327
+ strength ε (= εc,t) and frequency ff yields Arnold tongue diagrams in agreement with those
328
+ reported from our experiments. The Arnold Tongue for wt in Fig. 3A and ptx1 in Fig. 4A are
329
+ reproduced with simulations shown in Fig. 5C and D respectively. The only parameter value
330
+ changed between Fig. 5C and D is the level of noise (T c,t
331
+ eff ), which is increased by an order of
332
+ magnitude. The differences in phase dynamics between wt and ptx1, when subjected to sym-
333
+ metric external loading, are therefore accounted by solely varying the noise.
334
+ Coordinated beating under selective loading
335
+ We next model flow synchronization by the θc-flows and the θt-flows. The selective forcing
336
+ (εc ̸= εt) allows the effect of flagellar dominance (λc ̸= λt) to manifest in the effective forcing
337
+ strength ε(θ) and hence in the synchronization profiles τ(ν; θ), Fig. 5E. Similar to our exper-
338
+ imental observations, θc-flow synchronizes the coordinated beating over the broadest range of
339
+ ν (i.e. largest ε). This is directly attributed to the dominance λc > λt: by setting λc = λt,
340
+ 10
341
+
342
+ the differences between τ(θc) and τ(θt) disappear even under selective loading (Fig. 5E inset).
343
+ Fig. 5F details how the asymmetry of inter-flagellar coupling (λc/λt) affects the asymmetry
344
+ between τ(θc) and τ(θt) . The open symbols represent ε(θ) measured from modeled τ(ν; θ)
345
+ and the lines represent Eq. (3). The difference between ε(θc) and ε(θt) increases with λc/λt,
346
+ and they each saturates to reflect only the forcing on the cis (εc, the grey dashed lines). With
347
+ fc = 45 Hz, ft = 65 Hz (23, 26), and f0 ≈ 50 Hz, we deduce from Eq. (3) that λc = 4λt for
348
+ wt cells. For wt cells under calcium depletion, experimental results are reproduced with a lower
349
+ total forcing strength (Fig. 5G). εc + εt is set to 4.08 Hz (15% lower) to reflect the 7% − 20%
350
+ decrease in ε(θ) induced by calcium depletion.
351
+ The ptx1 results are reproduced with a stronger noise (T c,t
352
+ eff = 9.42 rad2/s) and a symmetric
353
+ inter-flagellar coupling λc/λt = 1, see Fig. 5H and Table. 1. Both changes are necessary for
354
+ reproducing the synchronization profiles of ptx1 in Fig. 5H: while the stronger noise lowers
355
+ the maximal values of τ(θ, ν), setting λc/λt = 4 would still result in τ(θc) > τ(θt) in the
356
+ central range (|ν| ≲ 2.4 Hz). Finally, it is noteworthy that the noise in ptx1 increases not only
357
+ because a higher noise value for individual flagella, but also because the cis-trans coupling has
358
+ become symmetric. As shown by Eq. (3), the unilateral coupling promotes not only the cis-
359
+ frequency in the synchrony but also the cis-noise. Given T c
360
+ eff ≪ T t
361
+ eff and λc = 4λt, we confirm
362
+ with simulations that the cis stabilizes the beating frequency of the trans and decreases its
363
+ beating noise. The simulations are in good agreement with experimental noise measurements,
364
+ see SM. Sec.S4 for details.
365
+ Discussion
366
+ The two flagella of C. reinhardtii have long been known to have inherently different dynamic
367
+ properties such as frequency, waveform, level of active noise, and responses to second messen-
368
+ gers (23,25,26,29,30). Intriguingly, when connected by basal fibers and beating synchronously,
369
+ they both adopt the kinematics of the cis-(eyespot) flagellum, which led to the assumption that
370
+ the flagella may have differential roles in coordination. In this work, we test this hypothesis by
371
+ employing oscillatory flows applied from an angle with respect to the cells’ symmetry axis and
372
+ thus exert biased loads on one flagellum.
373
+ Without an exception, in wt cells, θc-flows, the ones that selectively load the cis flagellum,
374
+ are always more effective in synchronizing the flagellar beating than the θt-flows. This is shown
375
+ by the larger effective forcing strengths ( ε(θc) > ε(θt) , Fig. 3B-C) and larger synchronized time
376
+ 11
377
+
378
+ fractions ( τ(θc) > τ(θt) , Fig. 3D). Mapping the measured forcing strength ε(θ) as a function
379
+ of the loads, we find empirically that ε ∝ F
380
+ c
381
+ Flow (Fig. 3F) and that trans-loads appear to mat-
382
+ ter negligibly. These observations all indicate that the cis-loads determine whether an external
383
+ forcing can synchronize the cell. Moreover, this point is further highlighted by an unexpected
384
+ finding: when θt-flows are applied, the trans flagellum always beats against the external flow
385
+ (P t
386
+ Flow < 0) and the only stabilizing factor for flow synchronization is the cis flagellum working
387
+ along with the flow during the recovery stroke (Fig. 2C lower panel). These observations defini-
388
+ tively prove that the two flagella have differential roles in the coordination and interestingly
389
+ imply that flagella are coupled to external flow only through the cis.
390
+ To have a mechanistic understanding of this finding, we model the system with Eq. (2). In
391
+ the model, selective hydrodynamic loading and flagellar dominance in the coordinated beating
392
+ are respectively represented by εc ̸= εt and λc ̸= λt. Setting out from the model, we obtain
393
+ closed-form expressions for observables such as f0 and ε (Eq. (3)), which illustrate how flag-
394
+ ellar dominance and selective loading affect the coordinated flagellar beating. Moreover, with
395
+ Monte-Carlo simulation, we clarified the interplay between flows and flagella (SM. Sec.S3),
396
+ and reproduces all experimental observations.
397
+ With the model, we show that a ”dominance” of the cis (λc > λt) is sufficient to explain
398
+ why the coordinated flagellar beating bears the frequency and the noise level of the cis flag-
399
+ ellum. In the model, such dominance means that the cis-phase is much less sensitive to the
400
+ trans-phase than the other way around. We then reproduce the phase dynamics of flow synchro-
401
+ nization at varying detunings (Fig. 5B), amplitudes (Fig. 5C), and noise (Fig. 5D). Exploiting
402
+ the observation that the coordination between flagella cannot be broken by external flows up
403
+ to the strongest ones tested (εmax ∼ 10 Hz, Fig. 3A), we quantify the lower limit of the total
404
+ basal coupling, λc + λt, to be approximately 40 Hz (deduced in SM. Sec.S3), which is an order
405
+ magnitude larger than the hydrodynamic inter-flagellar coupling (31,50–52).
406
+ The modulation of flagellar dominance mediates tactic behaviors (22, 23, 38, 47). Calcium
407
+ is hypothesized to be underlying the modulation of dominance, as it causes the connecting
408
+ fiber between flagella to contract (53), modulates the cis- and trans activity (e.g. beating am-
409
+ plitude) differentially (22), and calcium influx comprises the initial step of CR’s photo- (54)
410
+ and mechanoresponses (45). We therefore investigate flagellar coupling in the context of tactic
411
+ steering by depleting the environmental free calcium and hence inhibiting signals of calcium
412
+ influxes. Cells are first acclimated to calcium depletion, and then tested with the directional
413
+ flows. Our results show that the cis dominance does not require the involvement of free envi-
414
+ 12
415
+
416
+ ronmental calcium. Calcium depletion merely induces an overall drop in the forcing strength
417
+ perceived by the cell ε(θ) (7% − 20%), which is captured by reducing εc + εt for 15% (mean
418
+ drop) in the model (Fig. 5G). Together, our results indicate that the leading role of cis, is an
419
+ inherent property, that does not require active influx of external calcium, and possibly reflects
420
+ an intrinsic mechanical asymmetry of the cellular mesh that anchors the two flagella into the
421
+ cell body.
422
+ In ptx1 cells, a lack of flagellar dominance (λc = λt) and a stronger noise level help repro-
423
+ duce our experimental observations. Previous studies suggested that both flagella of ptx1 are
424
+ similar to the wildtype trans (23), and that the noise levels of this mutant’s synchronous beating
425
+ are much greater than those of wt (29) (see also SM. Sec.S4). If both flagella and their anchoring
426
+ roots indeed have the composition of the wildtype trans, such symmetry would predict λc = λt.
427
+ This symmetric coupling renders the noise of ptx1 Teff = T t
428
+ eff (Eq. (3)), which is about an order
429
+ of magnitude larger than the noise of wt Teff ≈ T c
430
+ eff.
431
+ The comparison between ptx1 and wt highlights an intriguing advantage of the observed
432
+ unilateral coupling (λc ≫ λt); that is, it strongly suppresses the high noise of the trans. Consid-
433
+ ering that the trans is richer in CAH6 protein and this protein’s possible role in inorganic carbon
434
+ sensing (14,20), the potential sensing role of the trans is worth noticing. Assuming the strong
435
+ noise present in the trans originates from the biochemical processes related to sensing, then
436
+ the unilateral coupling effectively prevents such noise from perturbing the cell’s synchronous
437
+ beating and effective swimming. In this way, the asymmetric coupling may combine the benefit
438
+ of having a stable cis as the driver while equipping a noisy trans as a sensor.
439
+ Material and methods
440
+ Cell culture
441
+ CR wildtype (wt) strain cc125 (mt+) and flagellar dominance mutant ptx1 cc2894 (mt+) are
442
+ cultured in TRIS-minimal medium (pH=7.0) with sterile air bubbling, in a 14h/10h day-night
443
+ cycle. Experiments are performed on the 4th day after inoculating the liquid culture, when the
444
+ culture is still in the exponential growth phase and has a concentration of ∼ 2 × 105 cells/ml.
445
+ Before experiments, cells are collected and resuspended in fresh TRIS-minimal (pH=7.0).
446
+ 13
447
+
448
+ Calcium depletion
449
+ In calcium depletion assays, cells are cultured in the same fashion as mentioned above but
450
+ washed and resuspended in fresh TRIS-minimal medium + 0.5 mM EGTA (pH=7.0). Free
451
+ calcium concentration is estimated to drop from 0.33 mM in the TRIS-minimal medium, to
452
+ 0.01 µM in the altered medium (46). Experiments start at least one hour after the resuspension
453
+ in order to acclimate the cells.
454
+ Experimental setup
455
+ Single cells of CR are studied following a protocol similar to the one described in (31). Cell
456
+ suspensions are filled into a customized flow chamber with an opening on one side. The air-
457
+ water interface on that side is pinned on all edges and is sealed with silicone oil. A micropipette
458
+ held by micromanipulator (SYS-HS6, WPI) enters the chamber and captures single cells by as-
459
+ piration. The manipulator and the captured cell remain stationary in the lab frame of reference,
460
+ while the flow chamber and the fluid therein are oscillated by a piezoelectric stage (Nano-Drive,
461
+ Mad City Labs), such that external flows are applied to the cell. Frequencies and amplitudes of
462
+ the oscillations are individually calibrated by tracking micro-beads in the chamber. Bright field
463
+ microscopy is performed on an inverted microscope (Nikon Eclipse Ti-U, 60× water immersion
464
+ objective). Videos are recorded with a sCMOS camera (LaVision PCO.edge) at 600-1000 Hz.
465
+ Measurement scheme
466
+ The flagellar beating of each tested cell is recorded before, during, and after the application
467
+ of the flows. We measure the cell’s average beating frequency f0 over 2 s (∼100 beats). For
468
+ ptx1 cells, f0 is reported for the in-phase (IP) synchronous beating. Unless otherwise stated,
469
+ directional flows (θ = 0, ±45◦) are of the same amplitude (780±50 µm/s, mean±std), similar
470
+ to those used in Ref. (31). Flow frequencies ff are scanned over [f0 − 7, f0 + 7] Hz for each
471
+ group of directional flows.
472
+ Computation of the flagellar loads
473
+ To quantify the hydrodynamic forces on the flagella, we first track realistic flagellar deforma-
474
+ tion from videos wherein background flows are applied. Then we employ a hybrid method
475
+ combining boundary element method (BEM) and slender-body theory (40, 55) to compute the
476
+ 14
477
+
478
+ drag forces exerted on each flagellum and the forces’ rates of work. In this approach, each flag-
479
+ ellum is represented as a slender-body (55) with 26 discrete points along its centerline and the
480
+ time-dependent velocity of each of the 26 points is calculated by its displacement across frames.
481
+ The cell body and the pipette used to capture the cell are represented as one entity with a com-
482
+ pleted double layer boundary integral equation (56). Stresslet are distributed on cell-pipette’s
483
+ surface; while stokeslet and rotlet of the completion flow are distributed along cell-pipette’s
484
+ centerline (57). The no-slip boundary condition on the cell-pipette surface is satisfied at col-
485
+ location points. Lastly, stokeslets are distributed along the centerlines of the flagella, so that
486
+ no-slip boundary conditions are met on their surfaces. Integrating the distribution of stokeslets
487
+ f(s) over a flagellar shape, one obtains the total drag force F =
488
+
489
+ f(s)ds is obtained. Similarly,
490
+ the force’s rate of work is computed as P =
491
+
492
+ f(s) · U(s)ds, where U(s) is the velocity of the
493
+ flagellum at the position s along the centerline.
494
+ The computations shown in this study are based on videos of a representative cell which
495
+ originally beats at ∼50 Hz. The cell is fully synchronized by flows along different directions
496
+ (θ = 0◦, ±45◦ and 90◦) at 49.2 Hz. In the computations, the applied flows are set to have an
497
+ amplitude of 780 µm/s to reflect the experiments. Computations begin with the onset of the
498
+ background flows (notified experimentally by a flashlight event), and last for ∼30 beats (500
499
+ frames sampled at 801 fps). Additionally, we confirm the results of θt-flow-synchronization,
500
+ that both flagella spend large fractions of time beating against the flows, with other cells and
501
+ with θt-flows at other frequencies.
502
+ Isolate loads of external flows
503
+ The total loads (F and P) computed consist of two parts, one from the flow created by the two
504
+ flagella themselves and the other from the applied flow. In the low Reynolds number regime,
505
+ the loads of the two parts add up directly (linearity): F = FSelf + FFlow, and P = PSelf + PFlow.
506
+ To isolate FFlow and PFlow, we compute F′ = FSelf and P ′ = PSelf by running the computation
507
+ again but without the external flows, and obtain FFlow = F − F′ and PFlow = P − P ′.
508
+ Modeling parameters
509
+ We assume the flagellar intrinsic frequencies fc and ft to be 45 Hz and 65 Hz respectively
510
+ (23, 26, 28). On this basis, λc : λt is assumed to be 4:1 to account for the observed f0 (∼ 50
511
+ Hz). εc : εt is set as 2:1, 1:1, and 1:2 for the θc-flows, the θa-flows, and the θt-flows respectively,
512
+ 15
513
+
514
+ see Fig. 2A-C. Additionally, εc + εt is assumed to be constant to reflect the fact that F
515
+ c
516
+ Flow +
517
+ F
518
+ t
519
+ Flow approximately does not vary with flow directions. We take a typical value of T c,t
520
+ eff =
521
+ 1.57 rad2/s (31). The sum of inter-flagellar coupling λtot = λc + λt is set to be large enough,
522
+ i.e., λtot = 3νct with νct = |ft − fc|, to account for the fact that: 1) the coordinated beating
523
+ is approximated in-phase, and 2) up until the strongest flow applied, the coordinated beating
524
+ cannot be broken (quantitative evaluation is detailed in SM. Sec.S3). To model wt cells under
525
+ calcium depletion, we decrease εc + εt by 15% - which is the mean decrease in the observed
526
+ ε(θc) , ε(θa) , and ε(θt) (Fig. 3E). For ptx1 cells, we assume a symmetric inter-flagellar coupling
527
+ (λc = λt) and a stronger noise level (SM. Sec.S4). The parameters are summarized in Table. 1.
528
+ Table 1: Modeling parameters
529
+ variable
530
+ symbol (unit)
531
+ TRIS
532
+ EGTA
533
+ ptx1
534
+ Intrinsic freq. (23,26)
535
+ fc, ft (Hz)
536
+ 45,65
537
+ 45,65
538
+ 45,65
539
+ Basal coupling∗
540
+ λc + λt (Hz)
541
+ 60
542
+ 60
543
+ 60
544
+ cis dominance (23,38)
545
+ λc : λt (-)
546
+ 4:1
547
+ 4:1
548
+ 1:1
549
+ Flow detuning
550
+ ν (Hz)
551
+ [-10,10]
552
+ [-10,10]
553
+ [-10,10]
554
+ Total forcing (51)
555
+ εc + εt (Hz)
556
+ 4.8
557
+ 4.08
558
+ 4.8
559
+ Noise∗ (31)
560
+ T c,t
561
+ eff (rad2/s)
562
+ 1.57
563
+ 1.57
564
+ 9.42
565
+ ∗ detailed in SM. Sec.S3
566
+ 16
567
+
568
+ References
569
+ 1. Berg, H. C. & Brown, D. A. Chemotaxis in escherichia coli analysed by three-dimensional
570
+ tracking. Nature 239, 500–504 (1972).
571
+ 2. Smriga, S., Fernandez, V. I., Mitchell, J. G. & Stocker, R. Chemotaxis toward phytoplank-
572
+ ton drives organic matter partitioning among marine bacteria. Proceedings of the National
573
+ Academy of Sciences 113, 1576–1581 (2016).
574
+ 3. Hegemann, P. & Berthold, P. Chapter 13 - sensory photoreceptors and light control of
575
+ flagellar activity. In Harris, E. H., Stern, D. B. & Witman, G. B. (eds.) The Chlamydomonas
576
+ Sourcebook (Second Edition), vol. 3, 395–429 (Academic Press, London, 2009).
577
+ 4. Ueki, N., Matsunaga, S., Inouye, I. & Hallmann, A. How 5000 independent rowers coor-
578
+ dinate their strokes in order to row into the sunlight: Phototaxis in the multicellular green
579
+ alga volvox. BMC Biology 8, 103 (2010).
580
+ 5. Stehnach, M. R., Waisbord, N., Walkama, D. M. & Guasto, J. S. Viscophobic turning
581
+ dictates microalgae transport in viscosity gradients. Nature Physics 17, 926–930 (2021).
582
+ 6. Brokaw, C., Josslin, R. & Bobrow, L. Calcium ion regulation of flagellar beat symmetry in
583
+ reactivated sea urchin spermatozoa. Biochemical and Biophysical Research Communica-
584
+ tions 58, 795–800 (1974).
585
+ 7. Gong, A. et al. The steering gaits of sperm. Philosophical Transactions of the Royal Society
586
+ B: Biological Sciences 375, 20190149 (2020).
587
+ 8. Gadˆelha, H., Hern´andez-Herrera, P., Montoya, F., Darszon, A. & Corkidi, G. Human sperm
588
+ uses asymmetric and anisotropic flagellar controls to regulate swimming symmetry and cell
589
+ steering. Science Advances 6, eaba5168 (2020).
590
+ 9. Bennett, R. R. & Golestanian, R. A steering mechanism for phototaxis in Chlamydomonas.
591
+ Journal of The Royal Society Interface 12, 20141164 (2015).
592
+ 10. R¨uffer, U. & Nultsch, W. Flagellar photoresponses of Chlamydomonas cells held on mi-
593
+ cropipettes: II. Change in Flagellar Beat Pattern. Cell Motility and the Cytoskeleton 18,
594
+ 269–278 (1991).
595
+ 17
596
+
597
+ 11. Ueki, N. & Wakabayashi, K.-i. Dynein-mediated photobehavioral responses in Chlamy-
598
+ domonas. In Dyneins, 368–385 (Elsevier Ltd., 2017), second edition edn.
599
+ 12. Naitoh, Y. & Kaneko, H. Reactivated triton-extracted models of paramecium: Modification
600
+ of ciliary movement by calcium ions. Science 176, 523–524 (1972).
601
+ 13. Merchant, S. S. et al. The Chlamydomonas genome reveals the evolution of key animal and
602
+ plant functions. Science 318, 245–250 (2007).
603
+ 14. Choi, H. I., Kim, J. Y. H., Kwak, H. S., Sung, Y. J. & Sim, S. J. Quantitative analysis of the
604
+ chemotaxis of a green alga, Chlamydomonas reinhardtii, to bicarbonate using diffusion-
605
+ based microfluidic device. Biomicrofluidics 10, 014121–014121 (2016).
606
+ 15. R¨uffer, U. & Nultsch, W. Flagellar photoresponses of Chlamydomonas cells held on mi-
607
+ cropipettes: I. change in flagellar beat frequency. Cell Motility and the Cytoskeleton 15,
608
+ 162–167 (1990).
609
+ 16. Holmes, J. & Dutcher, S. Cellular asymmetry in C. reinhardtii. Journal of Cell Science 94,
610
+ 273–285 (1989).
611
+ 17. Dutcher, S. K. & O’Toole, E. T. The basal bodies of C. reinhardtii. Cilia 5, 1–7 (2016).
612
+ 18. Mittelmeier, T. M., Boyd, J. S., Lamb, M. R. & Dieckmann, C. L. Asymmetric properties of
613
+ the Chlamydomonas reinhardtii cytoskeleton direct rhodopsin photoreceptor localization.
614
+ Journal of Cell Biology 193, 741–753 (2011).
615
+ 19. Sakakibara, H., Mitchell, D. R. & Kamiya, R. A Chlamydomonas outer arm dynein mutant
616
+ missing the alpha heavy chain. The Journal of cell biology 113, 615–622 (1991).
617
+ 20. Mackinder, L. C. et al. A spatial interactome reveals the protein organization of the algal
618
+ co2-concentrating mechanism. Cell 171, 133–147.e14 (2017).
619
+ 21. Yu, K., Liu, P., Venkatachalam, D., Hopkinson, B. M. & Lechtreck, K. F. The bbsome
620
+ restricts entry of tagged carbonic anhydrase 6 into the cis-flagellum of chlamydomonas
621
+ reinhardtii. PLOS ONE 15, 1–21 (2020).
622
+ 22. Kamiya, R. & Witman, G. B. Submicromolar levels of calcium control the balance of
623
+ beating between the two flagella in demembranated models of Chlamydomonas. Journal
624
+ of Cell Biology 98, 97–107 (1984).
625
+ 18
626
+
627
+ 23. Okita, N., Isogai, N., Hirono, M., Kamiya, R. & Yoshimura, K. Phototactic activity in
628
+ Chlamydomonas ’non-phototactic’ mutants deficient in Ca2+-dependent control of flagellar
629
+ dominance or in inner-arm dynein. Journal of Cell Science 118, 529–537 (2005).
630
+ 24. Takada, S. & Kamiya, R. Beat frequency difference between the two flagella of Chlamy-
631
+ domonas depends on the attachment site of outer dynein arms on the outer-doublet micro-
632
+ tubules. Cell Motility and the Cytoskeleton 36, 68–75 (1997).
633
+ 25. Saegusa, Y. & Yoshimura, K. cAMP controls the balance of the propulsive forces generated
634
+ by the two flagella of Chlamydomonas. Cytoskeleton 72, 412–421 (2015).
635
+ 26. Kamiya, R. & Hasegawa, E. Intrinsic difference in beat frequency between the two flagella
636
+ of Chlamydomonas reinhardtii. Experimental Cell Research 173, 299–304 (1987).
637
+ 27. R¨uffer, U. & Nultsch, W. Comparison of the beating of cis- and trans-flagella of Chlamy-
638
+ domonas cells held on micropipettes. Cell Motility and the Cytoskeleton 7, 87–93 (1987).
639
+ 28. Wan, K. Y., Leptos, K. C. & Goldstein, R. E. Lag, lock, sync, slip: the many phases of
640
+ coupled flagella. Journal of The Royal Society Interface 11, 20131160 (2014).
641
+ 29. Leptos, K. C. et al. Antiphase synchronization in a flagellar-dominance mutant of Chlamy-
642
+ domonas. Physical Review Letters 111, 1–5 (2013).
643
+ 30. Wan, K. Y. Coordination of eukaryotic cilia and flagella. Essays In Biochemistry 62, 829–
644
+ 838 (2018).
645
+ 31. Quaranta, G., Aubin-Tam, M.-E. & Tam, D. Hydrodynamics versus intracellular coupling
646
+ in the synchronization of eukaryotic flagella. Physical Review Letters 115, 238101 (2015).
647
+ 32. Wan, K. Y. & Goldstein, R. E. Coordinated beating of algal flagella is mediated by basal
648
+ coupling. Proceedings of the National Academy of Sciences 113, E2784–E2793 (2016).
649
+ 33. R¨uffer, U. & Nultsch, W.
650
+ High-speed cinematographic analysis of the movement of
651
+ Chlamydomonas. Cell Motility 5, 251–263 (1985).
652
+ 34. Klindt, G. S., Ruloff, C., Wagner, C. & Friedrich, B. M. In-phase and anti-phase flagellar
653
+ synchronization by waveform compliance and basal coupling. New Journal of Physics 19,
654
+ 113052 (2017).
655
+ 19
656
+
657
+ 35. Liu, Y. et al. Transitions in synchronization states of model cilia through basal connection
658
+ coupling. Journal of The Royal Society Interface 15 (2018).
659
+ 36. Guo, H., Man, Y., Wan, K. Y. & Kanso, E. Intracellular coupling modulates biflagellar
660
+ synchrony. Journal of The Royal Society Interface 18, 20200660 (2021).
661
+ 37. Wan, K. Y. & Goldstein, R. E. Rhythmicity, recurrence, and recovery of flagellar beating.
662
+ Phys. Rev. Lett. 113, 238103 (2014).
663
+ 38. Horst, J. & Witman, G. B. Ptx1, a nonphototactic mutant of Chlamydomonas, lacks control
664
+ of flagellar dominance. The Journal of Cell Biology 120, 733–741 (1993).
665
+ 39. Wei, D., Dehnavi, P. G., Aubin-Tam, M.-E. & Tam, D. Is the zero reynolds number approx-
666
+ imation valid for ciliary flows? Physical Review Letters 122, 124502 (2019).
667
+ 40. Wei, D., Dehnavi, P. G., Aubin-Tam, M.-E. & Tam, D. Measurements of the unsteady flow
668
+ field around beating cilia. Journal of Fluid Mechanics 915, A70 (2021).
669
+ 41. Kralemann, B., Cimponeriu, L., Rosenblum, M., Pikovsky, A. & Mrowka, R. Phase dy-
670
+ namics of coupled oscillators reconstructed from data. Phys. Rev. E 77, 066205 (2008).
671
+ 42. Pikovsky, A., Rosenblum, M. & Kurths, J. Synchronization: A Universal Concept in Non-
672
+ linear Sciences. Cambridge Nonlinear Science Series (Cambridge University Press, 2001).
673
+ 43. Polin, M., Tuval, I., Drescher, K., Gollub, J. P. & Goldstein, R. E. Chlamydomonas swims
674
+ with two “gears” in a eukaryotic version of run-and-tumble locomotion. Science 325, 487–
675
+ 490 (2009).
676
+ 44. Friedrich, B. Hydrodynamic synchronization of flagellar oscillators. The European Physi-
677
+ cal Journal Special Topics 225, 2353–2368 (2016).
678
+ 45. Yoshimura, K. Stimulus Perception and Membrane Excitation in Unicellular Alga Chlamy-
679
+ domonas, 79–91 (Springer Berlin Heidelberg, Berlin, Heidelberg, 2011).
680
+ 46. Wakabayashi, K.-i., Ide, T. & Kamiya, R. Calcium-dependent flagellar motility activation
681
+ in C. reinhardtii in response to mechanical agitation. Cell Motility and the Cytoskeleton
682
+ 66, 736–742 (2009).
683
+ 20
684
+
685
+ 47. Pazour, G. J., Agrin, N., Leszyk, J. & Witman, G. B. Proteomic analysis of a eukaryotic
686
+ cilium. The Journal of Cell Biology 170, 103–113 (2005).
687
+ 48. Quarmby, L. & Hartzell, H. Two distinct, calcium-mediated, signal transduction pathways
688
+ can trigger deflagellation in C. reinhardtii.
689
+ The Journal of Cell Biology 124, 807–815
690
+ (1994).
691
+ 49. R¨uffer, U. & Nultsch, W.
692
+ Flagellar coordination in Chlamydomonas cells held on mi-
693
+ cropipettes. Cell Motility and the Cytoskeleton 41, 297–307 (1998).
694
+ 50. Brumley, D. R., Wan, K. Y., Polin, M. & Goldstein, R. E. Flagellar synchronization through
695
+ direct hydrodynamic interactions. eLife 3, e02750 (2014).
696
+ 51. Klindt, G. S., Ruloff, C., Wagner, C. & Friedrich, B. M. Load response of the flagellar beat.
697
+ Physical Review Letters 258101, 1–5 (2016).
698
+ 52. Pellicciotta, N. et al. Entrainment of mammalian motile cilia in the brain with hydrody-
699
+ namic forces. Proceedings of the National Academy of Sciences 117, 8315–8325 (2020).
700
+ 53. Hayashi, M., Yagi, T., Yoshimura, K. & Kamiya, R. Real-time observation of Ca2+-induced
701
+ basal body reorientation in Chlamydomonas. Cell Motility and the Cytoskeleton 41, 49–56
702
+ (1998).
703
+ 54. Harz, H. & Hegemann, P. Rhodopsin-regulated calcium currents in Chlamydomonas. Na-
704
+ ture 351, 489–491 (1991).
705
+ 55. Keller, J. B. & Rubinow, S. I. Slender-body theory for slow viscous flow. Journal of Fluid
706
+ Mechanics 75, 705–714 (1976).
707
+ 56. Power, H. & Miranda, G. Second kind integral equation formulation of stokes’ flows past
708
+ a particle of arbitrary shape. SIAM Journal on Applied Mathematics 47, 689–698 (1987).
709
+ 57. Keaveny, E. E. & Shelley, M. J. Applying a second-kind boundary integral equation for
710
+ surface tractions in stokes flow. Journal of Computational Physics 230, 2141 – 2159 (2011).
711
+ 58. Kamiya, R. Analysis of cell vibration for assessing axonemal motility in Chlamydomonas.
712
+ Methods 22, 383–387 (2000).
713
+ 21
714
+
715
+ Acknowledgments
716
+ The authors thank Roland Kieffer for technical support. D.W. thanks Ritsu Kamiya for helpful
717
+ discussions. The authors acknowledge support by the European Research Council (ERC starting
718
+ grants no. 716712 and no. 101042612).
719
+ Author Contributions
720
+ D.W. performed experiments, computations, designed the model, and drafted the manuscript.
721
+ G.Q. performed early experiments and obtained preliminary results. M.A. and D.T. conceived
722
+ the study, supervised the project and critically revised the manuscript.
723
+ Competing interests
724
+ Authors declare that they have no competing interests.
725
+ Supplementary materials
726
+ Supplementary Text
727
+ Figs. S1 to S5
728
+ References (23,26,28,29,31,38,43,58)
729
+ 22
730
+
731
+ Figure 1: Experimental workflow. (A) Captured CR cells are subjected to sinusoidal flows of
732
+ frequency ff along given angles (θ) in the xy-plane. Flows along θ = −45◦, 0◦, 45◦ of same
733
+ amplitude (780±50 µm/s, mean±std.) are used and termed as shown. (B-E) Extracting flagellar
734
+ phase ϕc and ϕt by image processing. Raw images (B) are thresholded and contrast-adjusted to
735
+ highlight the flagella (C). Mean pixel values within the user-defined interrogation windows (red
736
+ and blue circles) capture the raw phases of beating (D), which are then converted to observable-
737
+ independent phases (E). Inset: phase difference ϕc − ϕt. (F) Flagella-flow phase dynamics at
738
+ decreasing detuning ν = ff − f0 with f0 the cell’s beating frequency without external flow.
739
+ Traces i to iv are taken at detunings marked in the inset. Plateaus marked black represent
740
+ flow synchronization, whose time fractions τ = tsync/ttot are noted. ttot is the total time of
741
+ recording. Inset: the flow synchronization profile, τ(ν), reports the effective forcing strength
742
+ 2ε by its width.
743
+ 23
744
+
745
+ A
746
+ Oa- flow,0
747
+ B
748
+ eyespot
749
+ Ot- flow, 45
750
+ c -flow,-45°
751
+ A
752
+ (s/ur)
753
+ -780μm/s
754
+ D
755
+ D
756
+ t ()
757
+ E
758
+ c
759
+ pt
760
+ (2元)
761
+ 0.5
762
+ 0.5S
763
+ 0
764
+ 4444442
765
+ 0.5
766
+ 0
767
+ 4
768
+ 8
769
+ 12
770
+ 16
771
+ t (beat)
772
+ F 12
773
+ IV
774
+ ii
775
+ () - )
776
+ i, T=0
777
+ T
778
+ 2
779
+ 8
780
+ ii
781
+ ii, T = 0.18
782
+ 1
783
+ 0
784
+ -6
785
+ 0
786
+ ii, T = 0.80
787
+ v (Hz)
788
+ 4
789
+ iv, T = 1.00
790
+ 0
791
+ 0
792
+ 2
793
+ 4
794
+ 6
795
+ 8
796
+ 10
797
+ t (s)Figure 2: External flagellar loads when beating is synchronized. Force magnitude (upper pan-
798
+ els) and power (lower panels) exerted by external flows of θ = 0◦ (A, θa-flow), −45◦ (B,
799
+ θc-flow), and +45◦ (C, θt-flow). The medians (solid lines) and interquartile ranges (shadings)
800
+ are computed over ∼20 synchronized beats. Dashed horizontal lines: loads averaged over a
801
+ synchronized beat. Force magnitudes and powers are scaled by F0=9.9 pN and P0=1.1 fW
802
+ respectively. Flagellar phase corresponds to the displayed shapes in the middle x-axis.
803
+ 24
804
+
805
+ I cis loads
806
+ trans loads
807
+ = Median
808
+ Interquartile
809
+ :.: Beat-averaged
810
+ A
811
+ B
812
+ FFlow/Fo
813
+ 0.5
814
+ 0
815
+ H
816
+ /Po
817
+ 10
818
+ 0
819
+ -10
820
+ 0元/2元3元/22元
821
+ 0元/2元3元/2 2元
822
+ 0元/2元3元/2 2元
823
+ Flagellar phase (rad)
824
+ Flagellar phase (rad)
825
+ Flagellar phase (rad)Figure 3: Flow synchronization of wt cells. (A) Arnold tongue of a representative cell tested
826
+ with θa-flow. The contour is interpolated from N=132 measurements (6 equidistant amplitudes
827
+ × 22 equidistant frequencies), and color-coded by the entrained time fraction τ. (B) The syn-
828
+ chronization profiles τ(ν; θ) of a representative wt cell (inset), the median profile of the TRIS
829
+ group wt cells (N=11, solid lines) and the EGTA group (N=6, dashed lines), with either θc-
830
+ flows (red), θa-flows (yellow) or θt-flows (blue). Shaded areas are the interquartile ranges for
831
+ the TRIS group. (C) Tested wt cells represented on the ε(θc) − ε(θt) plane (TRIS group). Solid
832
+ line: the first bisector line (y = x). (D) Comparing τ(ν; θc) and τ(ν; θt) for each cell at each ap-
833
+ plied frequency. N=132 pairs of experiments are represented on the τ(θc) − τ(θt) plane. More
834
+ than 90% of them are below the first bisector line. (E) The coupling strengths ε(θ) of the TRIS
835
+ group (black) and the EGTA group (gray). Bars and error bars: mean and 1 std., respectively.
836
+ Inset: δε = ε(θc) − ε(θt) . NS: not significant, p>0.05, Kruskal-Wallis test, One-Way ANOVA.
837
+ Relations between the forcing strength ε and the loads on the cis (F) and the trans flagellum
838
+ (G). Markers represent different flow angles, see the drawings.
839
+ 25
840
+
841
+ 20
842
+ C
843
+ 4
844
+ 0
845
+ (zH)
846
+ (10)3
847
+ 2
848
+ 8
849
+ 0
850
+ 0
851
+ -10-5
852
+ 0
853
+ 510
854
+ 15
855
+ 2025
856
+ 0
857
+ 2
858
+ 4
859
+ 6
860
+ v (Hz)
861
+ ε(0c) (Hz)
862
+ B
863
+ D
864
+ A single celi
865
+ 4
866
+ 0
867
+ 4
868
+ Median over population
869
+ T=1
870
+ TRIS
871
+ 0
872
+ EGTA :
873
+ 2 = 6.0 Hz
874
+ T
875
+ 0
876
+ t(Gc) (-)
877
+ 1
878
+ E
879
+ 6
880
+ (zH)
881
+ 3NS
882
+ 5.3 Hz
883
+ E
884
+ 3
885
+ 3.8 Hz
886
+ m
887
+ 0
888
+ -8
889
+ -4
890
+ 0
891
+ 4
892
+ 8
893
+ v (Hz)
894
+ TRIS EGTA
895
+ cis loads
896
+ trans loads
897
+ F
898
+ 4
899
+ G 4
900
+ 2
901
+ 2
902
+
903
+ m
904
+ 3
905
+ 0
906
+ 0
907
+ 0
908
+ 0.5
909
+ 0
910
+ 5
911
+ 0
912
+ 0.5
913
+ 0
914
+ 5
915
+ FFlow/Fo
916
+ Plow/Po
917
+ FFlow/Fo
918
+ Pflow/PoFigure 4: The asymmetric susceptibility to flow synchronization is lost in the flagellar domi-
919
+ nance mutant ptx1. (A) Arnold tongue of a representative ptx1 cell tested with θa-flow. The
920
+ contour is interpolated from N=132 measurements (6 equidistant amplitudes × 22 equidistant
921
+ frequencies). Color bar: the entrained time fraction τ = tsync/tIP. (B) Flow synchronization
922
+ profiles τ(ν; θ) of N=14 ptx1 cells, tested with θc-flows (red), θa-flows (yellow) and θt-flows
923
+ (blue). (C) ε(θc) and ε(θt) of the tested cells. The first bisector line (solid): y = x. (D)
924
+ τ(ν; θc,t) for each cell at each applied frequency. N=154 points are present.
925
+ 26
926
+
927
+ A
928
+ 20
929
+ 6
930
+ () n/n
931
+
932
+ 10
933
+ (zH)
934
+ 4
935
+
936
+
937
+
938
+ 8
939
+
940
+ @2
941
+
942
+
943
+ -10
944
+ -5
945
+ 0
946
+ 5
947
+ 10
948
+ 15
949
+ 20 25
950
+
951
+ 3
952
+ v (Hz)
953
+ 0
954
+ B
955
+ Median
956
+ Interquartile
957
+ 0
958
+ 2
959
+ 4
960
+ 6
961
+ (0c) (Hz)
962
+ 0
963
+ (-) (0)1
964
+ T
965
+ 0
966
+ 1
967
+ 0
968
+ 0
969
+ -8
970
+ -4
971
+ 0
972
+ 4
973
+ 8
974
+ 0
975
+ 1
976
+ v (Hz)
977
+ T(c) (-)Figure 5: Modeling the asymmetric flow synchronization. (A) Modeling scheme describing a
978
+ cell beating under directional flow (θc-flow as an example). Arrows represent the directional
979
+ coupling coefficients with line thickness representing the relative strength. For example, λc
980
+ points from cis to trans, representing how the latter (ϕc) is sensitive to the former (ϕt); mean-
981
+ while, the arrow of λc being thicker than λt means that ϕt is much more sensitive to ϕc than
982
+ the other way around. (B) Modeled phase dynamics of flow synchronization under θa-flows,
983
+ analogous to Fig. 1F. Reproducing the Arnold tongue diagrams at the noise level of wt (C)
984
+ and ptx1 (D), analogous to Fig. 3A and Fig. 4A respectively. (E) Flow synchronization profiles
985
+ τ(ν; θ) obtained experimentally (upper panel) and by modeling (lower panel). Inset: the mod-
986
+ eling results with symmetric inter-flagellar coupling. (F) Effective forcing strength ε(θ) as a
987
+ function of the inter-flagellar coupling asymmetry λc/λt. Points: measured from simulation;
988
+ lines: analytical approximation (Eq. (3)); dashed lines: εc respectively for the θc-flow, θa-flow,
989
+ and θt-flow (from top to bottom). (G) Reproducing the flow synchronization of wt cells under
990
+ calcium depletion (H) Reproducing results of ptx1. See Table. 1 for the modeling parameters.
991
+ 27
992
+
993
+ A
994
+ B
995
+ trans
996
+ cis
997
+ (fe, pc)
998
+ 12
999
+ (ft, Pt)
1000
+ 111
1001
+ 1
1002
+ ii
1003
+ 1
1004
+ ()
1005
+ fi, -0
1006
+ 6
1007
+ 0
1008
+ 6
1009
+ Idh.
1010
+ v (Hz)
1011
+ Λt
1012
+ ji, {-0.27
1013
+ i, t=0.85
1014
+ Et
1015
+ Ec
1016
+ iv, T=1.00
1017
+ (fr, Pf)
1018
+ 0
1019
+ 0
1020
+ 2
1021
+ 4
1022
+ 6
1023
+ 8
1024
+ Induced flow (0=-45°)
1025
+ t (s)
1026
+ C
1027
+ D
1028
+ 10
1029
+ 10
1030
+ Ec,t (Hz)
1031
+ (zH)
1032
+ 5
1033
+ Low noise level
1034
+ Ec,t
1035
+ High noise level
1036
+ 0
1037
+ 0
1038
+ -10-5
1039
+ ¥05101520
1040
+ 25
1041
+ -10-5
1042
+ 051015 20 25
1043
+ v (Hz)
1044
+ v (Hz)
1045
+ E
1046
+ F
1047
+ Exp, wt
1048
+ 3
1049
+ 0
1050
+ (zH)
1051
+ 2
1052
+ 0
1053
+ 0
1054
+ Model,wt
1055
+ m
1056
+ Analytical
1057
+ 2e/t=4
1058
+ 1
1059
+ 6
1060
+ Monte-Carlo
1061
+ Ac=入t
1062
+ 000
1063
+ Ec
1064
+ 0
1065
+ 0
1066
+ -6
1067
+ 6
1068
+ 0
1069
+ 0
1070
+ 5
1071
+ 10
1072
+ 15
1073
+ 20
1074
+ 8
1075
+ 入/M
1076
+ v (Hz)
1077
+ G
1078
+ H
1079
+ Exp,wt
1080
+ Exp, ptx1
1081
+ EGTA
1082
+ 0
1083
+ 0
1084
+ T
1085
+ 1
1086
+ L
1087
+ [Model,wt
1088
+ Model, ptx1
1089
+ EGTA
1090
+ 0
1091
+ 0
1092
+ 6
1093
+ 0
1094
+ 6
1095
+ -6
1096
+ 0
1097
+ 6
1098
+ v (Hz)
1099
+ v (Hz)Supplementary materials for
1100
+ The younger flagellum coordinates the beating in
1101
+ C. reinhardtii
1102
+ Da Wei1,3,Greta Quaranta2, Marie-Eve Aubin-Tam1†, Daniel S.W. Tam2∗
1103
+ 1Department of Bionanoscience, Delft University of Technology,
1104
+ 2628CJ Delft, Netherlands.
1105
+ 2Laboratory for Aero and Hydrodynamics, Delft University of Technology,
1106
+ 2628CD Delft, Netherlands.
1107
+ 3Beijing National Laboratory for Condensed Matter Physics, Institute of Physics,
1108
+ Chinese Academy of Sciences; Beijing 100190, China.
1109
+ †Corresponding author. Email: [email protected];
1110
+ ∗Corresponding author. Email: [email protected].
1111
+ 1
1112
+ arXiv:2301.13278v1 [physics.bio-ph] 30 Jan 2023
1113
+
1114
+ S1
1115
+ Extracting coupling strength by fitting phase dynamics
1116
+ In the work described in the manuscript, the flagellum-flow coupling strength ε in wt cells is
1117
+ mainly extracted by the synchronization profile τ(ν) ≥50%. Meanwhile, in previous works [1,
1118
+ 2], fitting the distribution of phase dynamics is employed to extract ε. In the latter approach, the
1119
+ idea is that the phase locking during synchronization leads to a peaked probability distribution
1120
+ of ∆ϕ, whose width is affected by the effective noise Teff. The distribution, P(∆ϕ), can be
1121
+ derived from the Adler equation Eq. (1) as:
1122
+ P(∆ϕ) =
1123
+ � ∆ϕ+2π
1124
+ δct
1125
+ exp(V (∆ϕ′) − V (∆ϕ)
1126
+ Teff
1127
+ )d∆ϕ′.
1128
+ (S1)
1129
+ Here V (∆ϕ) = ν∆ϕ + ε cos(∆ϕ) is a wash-board potential, Teff is the noise, and ∆ϕ is the
1130
+ difference between the flagellar phase and the flow’s phase.
1131
+ Here, we demonstrate that these two approaches are equivalent in extracting ε. For all
1132
+ wt cells tested in the TRIS-minimal medium (N=11), their ε(θ) measured by the τ(ν) width
1133
+ and extracted from fitting are plotted against each other, Fig. S1. All points center around the
1134
+ identity line, showing the equivalence in obtaining ε by the two methods. For the ptx1 dataset,
1135
+ ε are extracted from fitting the phase dynamics.
1136
+ 2
1137
+
1138
+ ��������������
1139
+ �����������������
1140
+ ����������
1141
+ ���������
1142
+ �������������
1143
+ Figure S1: Equivalence of extracting coupling strength ε by different methods. Each point
1144
+ represents one cell under either the θa-flow (green square), the θc-flow (red circle), or the θt-
1145
+ flow (blue triangle). The x coordinate is the coupling strength ε measured by the half width of
1146
+ synchronization profile τ(ν) ≥ 50%; and the y coordinate is obtained by fitting the flagellar
1147
+ phase dynamics.
1148
+ 3
1149
+
1150
+ S2
1151
+ Hydrodynamic computation for flow along 90 degree
1152
+ Similar to Fig. 2 in the main text, we present the computed drag force and power for the
1153
+ flow along 90◦. The solid lines and the shadings represent the median and the interquartile
1154
+ range of FFlow and PFlow over the flow-synchronized beats, respectively. Force magnitudes
1155
+ are scaled by F0 = 6πµRU0 = 9.9 pN, which is the Stokes drag on a typical free-swimming
1156
+ cell (radius R = 5 µm, swim velocity U0 = 110 µm/s); while the viscous powers are scaled by
1157
+ P0 = F0U0 = 6πµRU 2
1158
+ 0 = 1.1 fW. Here µ = 0.95 mPa·s is the dynamic viscosity of water at
1159
+ 22 oC. Quantitatively, the mean force is 0.37F0 and 0.34F0 (Fig. S2 top panel) while the mean
1160
+ power is -0.2P0 and -0.4P0 (Fig. S2 bottom panel), for the cis and the trans respectively.
1161
+ Figure S2: Computed hydrodynamic loads on the flagella. Computation results of the drag
1162
+ force (upper panel) and the force’s rate of work (lower panel) on the cis (red) and the trans
1163
+ (blue) flagellum during synchronized cycles, when the cell is subjected to the flow with θ = 90◦.
1164
+ Scaling factors F0=9.9 pN and P0=1.1 fW.
1165
+ 4
1166
+
1167
+ FFlow/Fo
1168
+ 1
1169
+ 0.5
1170
+ 0
1171
+ cis loads
1172
+ 10
1173
+ PFlow/Po
1174
+ - trans loads
1175
+ 0
1176
+ -10
1177
+ 0
1178
+ 元/2
1179
+ 2元3元/22元
1180
+ Flagellar phase (rad)S3
1181
+ The model
1182
+ The external flow and the two flagella are described by three coupled ordinary differential equa-
1183
+ tions (ODEs). Phase dynamics of these equations are examined by Monte-Carlo simulation.
1184
+ The temporal resolution of simulation (dt) is 1 ms, which corresponds to the experimental
1185
+ frame rates (801 Hz).
1186
+
1187
+
1188
+
1189
+
1190
+
1191
+
1192
+
1193
+
1194
+
1195
+
1196
+
1197
+
1198
+
1199
+ dϕf
1200
+ dt = 2πff
1201
+ (S2a)
1202
+ dϕc
1203
+ dt = 2πfc − 2πλt sin(��c − ϕt) − 2πεc sin(ϕc − ϕf) + ζc(t)
1204
+ (S2b)
1205
+ dϕt
1206
+ dt = 2πft − 2πλc sin(ϕt − ϕc) − 2πεt sin(ϕt − ϕf) + ζt(t).
1207
+ (S2c)
1208
+ The cis, the trans, and the external flow are described as oscillators, whose intrinsic fre-
1209
+ quencies are fc,t,f and phases ϕc,t,f, respectively. The flow is assumed to be noise free and the
1210
+ two flagella are assumed to have the same level of noise (ζc = ζt). The noises are assumed to
1211
+ be Gaussian, ⟨ζc,t(τ + t)ζc,t(τ)⟩ = 2T c,t
1212
+ eff δ(t).
1213
+ S3.1
1214
+ Flagellar synchronization
1215
+ Setting εc and εt to 0, the interaction between the two flagella in the absence of the flow is
1216
+ modeled by:
1217
+
1218
+
1219
+
1220
+
1221
+
1222
+ dϕt
1223
+ dt = 2πfc − 2πλt sin(ϕc − ϕt) + ζc(t)
1224
+ (S3a)
1225
+ dϕc
1226
+ dt = 2πft − 2πλc sin(ϕt − ϕc) + ζt(t).
1227
+ (S3b)
1228
+ When the two flagella are able to beat synchronously, dϕc
1229
+ dt = dϕt
1230
+ dt = f0, we can obtain the
1231
+ analytical expression of f0 by adding up λc×Eq. (S3a) and λt×Eq. (S3b):
1232
+ f0 = λtft + λcfc
1233
+ λc + λt
1234
+ .
1235
+ (S4)
1236
+ 5
1237
+
1238
+ Meanwhile, the steady-state phase difference δct = ϕc −ϕt is obtained by subtracting Eq. (S3a)
1239
+ from Eq. (S3b):
1240
+ sin(δct) = fc − ft
1241
+ λc + λt
1242
+ = νct
1243
+ λtot
1244
+ .
1245
+ (S5)
1246
+ It is therefore obvious that the two flagella can only beat at the same frequency (dϕc/dt =
1247
+ dϕt/dt = f0) if |νct/λtot| ≤ 1.
1248
+ S3.2
1249
+ Interaction between three oscillators
1250
+ Now we put the flow back into the picture. According to experimental observations, the two
1251
+ flagella mostly beat synchronously, we therefore focus on this case and first simplify the equa-
1252
+ tions. By adding up λc×Eq. (S2b) and λt×Eq. (S2c), and substituting ϕc,t as ϕ0 = ϕc −δct/2 =
1253
+ ϕt + δct/2, we obtain:
1254
+ dϕ0
1255
+ dt = 2πf0−2π λcεc
1256
+ λc + λt
1257
+ sin
1258
+
1259
+ ϕ0 − ϕf − δct
1260
+ 2
1261
+
1262
+ −2π
1263
+ λtεt
1264
+ λc + λt
1265
+ sin
1266
+
1267
+ ϕ0 − ϕf + δct
1268
+ 2
1269
+
1270
+ +λtζt + λcζc
1271
+ λc + λt
1272
+ .
1273
+ (S6)
1274
+ Given different choices of coupling constants (λc,t, εc,t), this equation would generate com-
1275
+ plex phase dynamics - as we shall see in the following sections. We first limit the discussion to
1276
+ small δct - as it is observed in our experiment as well as in [3]. The model’s asymptotic behavior
1277
+ at δct ≈ 0 is:
1278
+ dϕ0
1279
+ dt = 2πf0 − 2πε sin(ϕ0 − ϕf) + ζ0(t),
1280
+ (S7)
1281
+ where
1282
+ f0 = λtft + λcfc
1283
+ εtc + λt
1284
+ , ε = λtεt + λcεc
1285
+ λc + λt
1286
+ , ζ0 = λtζt + λcζc
1287
+ λc + λt
1288
+ .
1289
+ (S8)
1290
+ In this strong-coupling limit (δct ≈ 0, or equivalently, λtot ≫ νct), the coupled flagella
1291
+ behaves as a single oscillator whose beating frequency f0 will not be interfered by the external
1292
+ flow. The analytical form well captures the system’s behavior, as shown by Fig. 5F. Next we
1293
+ explore the model’s behaviors when λtot − νct is comparable with ε.
1294
+ 6
1295
+
1296
+ ���
1297
+ ���
1298
+ ���
1299
+ ���
1300
+ ���
1301
+ ���
1302
+ ������������
1303
+ f�
1304
+ f�
1305
+ ��������cis���������
1306
+ ��������trans���������
1307
+ ��������trans ����cis
1308
+ ���
1309
+ ���
1310
+ ���
1311
+ ������
1312
+ ������
1313
+ ������
1314
+ Figure S3: Determine the lower limit of λtot. The time fractions of the cis (a) and the trans
1315
+ flagellum (b) synchronized by the flow. (c) The time fraction of where cis and trans are syn-
1316
+ chronized. Arrows points towards increasing (λtot − ν)/ε.
1317
+ S3.3
1318
+ Lower limit of inter-flagellar coupling
1319
+ The value (λtot − νct)/ε determines if the flow can disrupt the synchronization between cis
1320
+ and trans. We assume νct = 20 Hz[4, 5, 6, 3] and focus on synchronization of the θa-flow.
1321
+ We plot the synchronization time fractions with increasing λtot in Fig. S3. When it satisfies
1322
+ (λtot − νct)/ε ≥ 2, external flows cease to affect the flagellar synchronization observably. As
1323
+ the strongest flow (21U0) applied experimentally corresponds to ε ≈ 10 Hz, altogether, we
1324
+ conclude that λtot ≳ νct + 2εmax = 40 Hz. In the main text, we set λtot = 60 = 3νct Hz, which
1325
+ satisfies this relation and matches the observation that the phase lag between the flagella (δct) is
1326
+ small.
1327
+ 7
1328
+
1329
+ S4
1330
+ Flagellar noise of the ptx1 mutant
1331
+ Here we show an as-yet uncharacterized strong noise present in the synchronous beating of the
1332
+ mutant ptx1. The in-phase (IP) mode of ptx1 cells and the breaststroke beating of the wt cells
1333
+ are similar in waveform and frequency [7, 8]. However, the former has a much stronger noise.
1334
+ Figure S4: Stronger frequency fluctuation of the IP mode of ptx1 cells. (a-e) Representative
1335
+ probability distributions of the beating frequency of a wt (a) and four ptx1 cells (b-e) over 30
1336
+ seconds. Probability distributions of the IP (purple) and AP mode (yellow) are respectively nor-
1337
+ malized for better visualization. The time fractions of the AP mode are noted in each panel. (f)
1338
+ The wt and ptx1 cells represented by its mean beating frequency ⟨f0⟩ and the standard deviation
1339
+ of the beating frequencies over time σ(f0).
1340
+ The strong noises show obviously in fluctuations of IP beating frequencies [8].
1341
+ In Fig. S4, we display the distribution of beating frequency of a representative wt cell (panel
1342
+ a) and four representative ptx1 cells (panels b-e). The broad peaks of the IP (purple) and AP
1343
+ (yellow) beating of ptx1 sharply contrast the narrow peak of wt. We quantify the frequency
1344
+ fluctuations of all the cells in the main text (N=11 for wt and N=14 for ptx1), Fig. S4f. The
1345
+ cells are represented by its mean beating frequency over time ⟨f0⟩ and the frequency’s standard
1346
+ deviation σ(f0). Clearly, the breaststroke beating of wt, the IP, and the AP mode of ptx1 each
1347
+ forms a cluster. The wt cluster is at (⟨f0⟩, σ(f0)) = (50.5 ± 2.6, 0.8 ± 0.3) Hz (mean± 1 std.
1348
+ the over cell population); and it is evidently less dispersed than both the IP and the AP mode
1349
+ 8
1350
+
1351
+ 0.1
1352
+ (a)
1353
+ (f)
1354
+ wt
1355
+ wildtype
1356
+ ptxl, IP mode
1357
+ 0
1358
+ 40
1359
+ 50
1360
+ 60
1361
+ 70
1362
+ 80
1363
+ ptxl, AP mode
1364
+ 0.06
1365
+ 4
1366
+ AP: 6.9%
1367
+ (b)
1368
+ ptx1
1369
+ 0
1370
+ (zH) (°J)o
1371
+ 0.04
1372
+ AP: 10.2%.
1373
+ (c)
1374
+ PDF
1375
+ -
1376
+ 0
1377
+ 2
1378
+ 0.06
1379
+ AP: 39.7%
1380
+ (d)
1381
+
1382
+ 0
1383
+
1384
+ 0.04
1385
+ AP: 20.4%
1386
+
1387
+ (e)
1388
+ 0
1389
+ 0
1390
+ 40
1391
+ 50
1392
+ 60
1393
+ 70
1394
+ 80
1395
+ 40
1396
+ 50
1397
+ 60
1398
+ 70
1399
+ 80
1400
+ Frequency (Hz)
1401
+ (fo) (Hz)of ptx1, which are at (47.4 ± 3.1, 3.4 ± 0.9) Hz and (67.6 ± 2.1, 1.9 ± 0.7) Hz, respectively.
1402
+ Under the assumption of a white (Gaussian) noise, σ(f0) is proportional to the noise level ζ,
1403
+ and thus scales with √Teff. Consider that σ(f0) for ptx1 is 3-5 folds larger than that of wt,
1404
+ we therefore conclude that the noise level in ptx1 is an order of magnitude larger than wt,
1405
+ T ptx1
1406
+ eff
1407
+ /T wt
1408
+ eff ∼ O(10).
1409
+ Figure S5: Effect of a low-noise cis in stabilizing the beating of the trans (a) Fluctuations in
1410
+ beating frequency (σ(f0)) under different coupling schemes and flagellar noises. Other model
1411
+ parameters are the same as used in the main text. The red and blue shaded area represent the
1412
+ experimentally observed range for ptx1 and wt cells, respectively, with short bars marking the
1413
+ mean values. (b) the rate of slip under the conditions. Error bars correspond to 1 std. over N=9
1414
+ repetitions.
1415
+ The stronger noise in ptx1 can be attributed to two sources, namely, the loss of a stable
1416
+ cis and the loss of the unilateral coupling, Fig. S5. We perform Monte-Carlo simulations of
1417
+ the coupled beating of cis and trans under three conditions: (1) a stable cis (T c
1418
+ eff = T 0
1419
+ eff =
1420
+ 1.57 rad/s2) coupled with the trans unilaterally (λc = 4λt), (2) a stable cis coupled with the
1421
+ trans bilaterally (λc = λt), and (3) an equally noisy cis (T c
1422
+ eff = T t
1423
+ eff) bilaterally coupled with
1424
+ trans, see the blue, yellow, and red data in Fig. S5 respectively. It is obvious that, when the trans
1425
+ is coupled to a stable cis, varying its noise over an order of magnitude only leads to a ∼ 20%
1426
+ stronger frequency fluctuation (the blue line in Fig. S5(a)). On the contrary, lacking either
1427
+ the unilateral coupling or the low-noised cis would increase the fluctuation for 200% (yellow
1428
+ 9
1429
+
1430
+ = =
1431
+ =
1432
+ =
1433
+ 0.1
1434
+ (a)
1435
+ (b)
1436
+ 3
1437
+ ptx1
1438
+ 0.08
1439
+ Slip rate (Hz)
1440
+ o(fo) (Hz)
1441
+ 0.06
1442
+ 2
1443
+ 0.04
1444
+ 1
1445
+ 0.02
1446
+ im
1447
+ 0
1448
+ 0
1449
+ 100
1450
+ 101
1451
+ 10°
1452
+ 101
1453
+ Teff / Teff
1454
+ Teff / Teffline) or 300% (red line). Qualitatively, simulation results are in agreement with experimental
1455
+ measurements assuming that T t
1456
+ eff/T c
1457
+ eff ∼ O(10), see the red and blue shaded areas in Fig. S5(a).
1458
+ Moreover, a low-noise cis is already sufficient to prevent slips from interrupting the synchrony
1459
+ between cis and trans, even for bilateral coupling. In Fig. S5(b), as long as the cis-noise remains
1460
+ low, slips will be sparse (< 0.01 Hz). Together, these simulation results highlight the stabilizing
1461
+ effect of a low-noise cis flagellum, and illustrates the contribution of unilateral coupling in
1462
+ further enhancing the stabilization.
1463
+ References
1464
+ [1] Polin, M., Tuval, I., Drescher, K., Gollub, J. P. & Goldstein, R. E. Chlamydomonas swims
1465
+ with two “gears” in a eukaryotic version of run-and-tumble locomotion. Science 325, 487–
1466
+ 490 (2009).
1467
+ [2] Quaranta, G., Aubin-Tam, M.-E. & Tam, D. Hydrodynamics versus intracellular coupling
1468
+ in the synchronization of eukaryotic flagella. Physical Review Letters 115, 238101 (2015).
1469
+ [3] Wan, K. Y., Leptos, K. C. & Goldstein, R. E. Lag, lock, sync, slip: the many phases of
1470
+ coupled flagella. Journal of The Royal Society Interface 11, 20131160 (2014).
1471
+ [4] Kamiya, R. & Hasegawa, E. Intrinsic difference in beat frequency between the two flagella
1472
+ of Chlamydomonas reinhardtii. Experimental Cell Research 173, 299–304 (1987).
1473
+ [5] Kamiya, R. Analysis of cell vibration for assessing axonemal motility in Chlamydomonas.
1474
+ Methods 22, 383–387 (2000).
1475
+ [6] Okita, N., Isogai, N., Hirono, M., Kamiya, R. & Yoshimura, K. Phototactic activity in
1476
+ Chlamydomonas ’non-phototactic’ mutants deficient in Ca2+-dependent control of flagellar
1477
+ dominance or in inner-arm dynein. Journal of Cell Science 118, 529–537 (2005).
1478
+ 10
1479
+
1480
+ [7] Horst, J. & Witman, G. B. Ptx1, a nonphototactic mutant of Chlamydomonas, lacks control
1481
+ of flagellar dominance. The Journal of Cell Biology 120, 733–741 (1993).
1482
+ [8] Leptos, K. C. et al. Antiphase synchronization in a flagellar-dominance mutant of Chlamy-
1483
+ domonas. Physical Review Letters 111, 1–5 (2013).
1484
+ 11
1485
+
DNFQT4oBgHgl3EQfPTY7/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
EdE0T4oBgHgl3EQfywKq/content/tmp_files/2301.02664v1.pdf.txt ADDED
@@ -0,0 +1,636 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Lindbladian-Induced Alignment in Quantum
2
+ Measurements
3
+ R. Englman and A. Yahalom
4
+ Ariel University, Ariel 40700,Israel
5
+ January 10, 2023
6
+ Keywords: Quantum measurement theory, Density matrix evolution, Quan-
7
+ tum state resolution, Lindblad operators, Quantum speed limit.
8
+ Abstract
9
+ An expression of the Lindbladian form is proposed that ensures an un-
10
+ ambiguous time-continuous reduction of the initial system-pointer wave-
11
+ packet to one in which the readings and the observable’s values are aligned,
12
+ formalized as the transition from an outer product to an inner product of
13
+ the system’s and apparatus’ density matrices. The jump operators are in
14
+ the basis of the observables, with uniquely determined parameters derived
15
+ from the measurement set-up (thereby differing from S. Weinberg’s Lind-
16
+ bladian resolution of wave-packet formalism) and conforming to Born’s
17
+ probability rules. The novelty lies in formalising the adaptability of the
18
+ surroundings (including the measuring device) to the mode of observa-
19
+ tion. Accordingly, the transition is of finite duration (in contrast to its
20
+ instantaneousness in the von Neumann’s formulation). This duration is
21
+ estimated for a simple half-spin-like model.
22
+ 1
23
+ Introduction
24
+ In the century-run of quantum physics (plus 4 years, if one marks its beginning
25
+ with the award of a Nobel Prize in 1918 to Max Planck for ”his discovery of
26
+ quanta”) a single shadow of non-sequitur has darkened its glorious achievements,
27
+ one that goes variously under the names of wave-function collapse, reduction of
28
+ the wave-packet, quantum measurement, einselection, etc. Aspects of the prob-
29
+ lem (or its articulations) were manifold, such as the breakdown of the predicted
30
+ time-development in accordance with the Schr¨odinger equation, the abruptness
31
+ of change in a measurement (”natura non facit saltum”, where art thou?), the
32
+ apparent non-applicability of quantum rules to macroscopic systems, imputed
33
+ arbitrariness of Born’s probability rules, the requirement of ”infinite regress”
34
+ 1
35
+ arXiv:2301.02664v1 [quant-ph] 6 Jan 2023
36
+
37
+ for the measuring apparatus and others. Numerous papers enlarged on these is-
38
+ sues [1, 2] and various proposals for resolution of the problem were put forward.
39
+ These include the observer’s cognition [3], stochastic effects [4], in particular
40
+ spontaneous localization [2, 5, 6], a many world scenario [7], non-linearity addi-
41
+ tion to the Schr¨odinger equation [8], Poincar´e recurrent state [9], gravitationally
42
+ induced collapse [10, 11, 12], etc.
43
+ Common to these works, and with the specific purpose of providing a
44
+ blue-print for measurements compatible with the Copenhagen formulation of
45
+ quantum theory, was the need to give expression to the coupling of the micro-
46
+ scopic system with its macroscopic environment. Standing apart from these and
47
+ belonging to the field of non-equilibrium thermodynamics and to the establish-
48
+ ment of equilibrium, a general form for this interaction was given by Lindblad
49
+ [13] and by Gorin , Kossakowski and Sudarshan [14], satisfying some necessary
50
+ conditions. Constructing a merger between the two separately oriented fields,
51
+ S. Weinberg recently proposed a Lindblad-operator mechanism for the collapse
52
+ of the density matrix (DM) in the course of a complete measurement [15]. No-
53
+ tably, the mechanism was linear in the state’s DM. The collapsed state (Eq. (1)
54
+ in [15]) comprises the set of projection operators of the measurable item; the
55
+ system’s Hamiltonian is described by a spectral decomposition onto the same
56
+ operators (Eq.
57
+ (16) in [15]) (although in the verbal discussion a more gen-
58
+ eral situation is considered): collapse is achieved ”independent[ly] of the details
59
+ of these [Lindblad] operators”. Decay between energy eigenstates had earlier
60
+ been treated by the Lindblad formalism (for a pedagogical presentation the
61
+ volume [16], Chapter 8 may be consulted) employing the interaction represen-
62
+ tation. However, this is not convenient for treating measurements of observables
63
+ that do not commute with the Hamiltonian. Detailed theories relate to the out-
64
+ come (”mapping”) of quantum operations, including measurements; the present
65
+ work describes the process of these happening.(For a pedagogical introduction
66
+ to stochasticity-induced wave-packet- reduction, obviating pointer reading, one
67
+ may refer to [17].)
68
+ 2
69
+ Overview of the Method and Terms
70
+ 2.1
71
+ The leading idea, also in review
72
+ While the concept of unity of observer and observation had already featured
73
+ in Bohr’s view: ”The answer that we get is built up from the combined interac-
74
+ tion of [the observer’s] state and the object of interrogation.” [18], this was not
75
+ given a formal expression in the Copenhagen interpretation. It was more em-
76
+ phatically asserted both by J. Bell: ”I meant that the ’apparatus’ should not be
77
+ severed from the rest of the world in boxes ...[19]” and A. Peres: ” A measure-
78
+ ment both creates and records a property of the system [20]”. This change in
79
+ 2
80
+
81
+ the course of a measurement a���ects also the environment outside the observed
82
+ system ; in the words of A. Leggett ”...under these conditions the macroscopic
83
+ apparatus, and more generally any part of the macro-world which has suffered
84
+ changes in the course of the measurement process, does not end up in a state
85
+ with definite macroscopic properties at all,... [1]”.
86
+ The same line of thought appears to motivate S. Weinberg, who wrote in his
87
+ preamble to a 2016 Lindbladian formulation of the masurement process[15], that
88
+ ”We will instead [of the original formulation of the Copenhagen interpretation,
89
+ (which we will not dwell on here)] adopt the popular modern
90
+ view that the
91
+ Copenhagen interpretation refers to open systems in which the transition is
92
+ driven by the ineraction of the microscopic system under study (which may
93
+ include an observer) chosen to bring the transition about.” (Our italics.)
94
+ These developments indicate the justification for a formulation in which the
95
+ effect of the apparatus is incorporated in the equation defining the evolution of
96
+ the system, rather than one in which the two entities are separate, barring an
97
+ interaction between them.
98
+ 2.1.1
99
+ ”Alignment”
100
+ The process whereby the pointer readings become in correspondence with the
101
+ possible values of the observable. Formally, for I possible values, the combined
102
+ density matrix reduces from comprising I2 terms to one having I terms. (E.g.,
103
+ equation (2.5) in [1].)
104
+ 2.1.2
105
+ ”Dissipator”
106
+ Added term (in the form of sums of appropriately weighted jump-operators)
107
+ to the standard time dependent Schr¨odinger equation, inducing non-unitary
108
+ evolution in the system, accompanied by changes of its information entropy.
109
+ 2.2
110
+ Motivation for the choice of formalism
111
+ Thermalization of open systems can be described by a Lindbladian formal-
112
+ ism in which Gibbsian probabilities are so inserted as parameters, that the
113
+ ”Dissipator” vanishes at these values of the density matrix. Replacement of
114
+ the Gibbsian probabilities by Born probabilities achieves alignment in a state
115
+ reduction and does so continuously.
116
+ Limitations: Born’s probability rules are assumed, not derived; the interac-
117
+ tion term is not traced to a microscopic mechanism.
118
+ The source of this interaction term, shown in Eqn. 6 below, incorporating
119
+ the coupling between the observed system and its surroundings (including the
120
+ 3
121
+
122
+ measuring device) is an open question (also raised by a referee). In its appli-
123
+ cation to a thermalization process, the Lindbladian jump operators have been
124
+ derived, though with the aids of several approximations (e.g.,[21]), as well as,
125
+ more recently, for the dissipation in a Dicke system with a bosonic background
126
+ [22]. We are not in the position to provide such first principle derivation for
127
+ the Lindbladian jump-operators bringing about a transition and incorporat-
128
+ ing the Born rules. It seems to be specific to the type of measurement under
129
+ consideration and it is clear that just any jump operator, as in Weinberg’s
130
+ Lindbladian formulation will not do the job . Likely, one would need to in-
131
+ clude non-Markovian dynamics, so that the coupling to the device and eventual
132
+ pointer reading are two separate consecutive events. Inclusion of such dynamics
133
+ is outside the scope of the present work.
134
+ 3
135
+ Assumptions
136
+ We explore the time (t)-development of the combined density matrix ρ(t) of
137
+ the measured system (S) and of the reading (pointer, dial, etc.) on the mea-
138
+ suring apparatus (A) for a complete and discrete measurement , expressing the
139
+ underlying assumptions by three propositions.
140
+ Proposition 1. In accord with the long-time historical approach, the mea-
141
+ sured object S and the pointer of the measuring set-up A are treated on equal
142
+ footings as subject to microscopic quantum laws, and formally describable by
143
+ their respective Hamiltonians. Aware of the difficulties connected with an ”in-
144
+ finite regress”, the effects of the rest of the Universe on S+A are not included
145
+ in the formalism; instead, for a phenomenological, approximative description, a
146
+ Lindbladian term appears in the master equation.
147
+ Proposition 2. Prior to the measurement with A and S decoupled, and being
148
+ free of external influence for a long time, both are in energy quantum states,
149
+ pure or mixed. After the measurement, the state is not an energy eigenstate
150
+ and subsequently it will spread over to a superposition of energy eigenstates.
151
+ The fast decoherence case treated below in section 5 is akin to the Zeno effect
152
+ [23].
153
+ Proposition 3. Only those states of the reading apparatus (e.g., the right
154
+ or left positions of a pointer) that may be in direct correspondence with the
155
+ measured states of the system (e.g., spin up or down) are given expression in
156
+ the formalism. (At a beginning, the case treated is one in which there is a one-
157
+ to-one correspondence between the states of the system and the readings of the
158
+ apparatus; a generalization is given subsequently.) A discussion in section
159
+ 8
160
+ touches on the epistemological status of the Lindbladian terms in a measurement
161
+ process.
162
+ 4
163
+
164
+ 4
165
+ Analysis
166
+ Considering (for simplicity) a pure state for the system, its initial state-vector
167
+ written in the basis of the observed property |S, i > takes the form
168
+ ψS(t = 0) =
169
+
170
+ i=1,..,I
171
+ cS
172
+ i |S, i >
173
+ (1)
174
+ Born’s rule for the probability of observing the i-component is |cS
175
+ i ]2 ≡ pi, sum-
176
+ ming to unity. Likewise, for the apparatus readings j, numbering J, one has
177
+ the superposition with (complex and normalized) coefficients cA
178
+ j
179
+ ψA(t = 0) =
180
+
181
+ j=1,..,J
182
+ cA
183
+ j |A, j >
184
+ (2)
185
+ We start with the one-to-one correspondence situation, for which I = J, and the
186
+ reading j on A establishes uniquely the value i = j for the system’s measured
187
+ property.
188
+ For the combined state-vector the density operator has the outer-product
189
+ form (where the stars denote complex conjugates):
190
+
191
+ i,j,i′,j′
192
+ |S, i > |A, j > cS∗
193
+ i cA∗
194
+ j cA
195
+ j′cS
196
+ i′ < A, j′| < S, i′| ≡
197
+
198
+ i,j,i′,j′
199
+ |i, j > Ciji′j′ < i′, j′|
200
+ (3)
201
+ the right hand side written in an obvious shortened notation, in which Ciji′j′ =
202
+ cS∗
203
+ i cA∗
204
+ j cA
205
+ j′cS
206
+ i′. After collapse, the density operator takes the aligned, single-sum
207
+ form
208
+
209
+ i
210
+ |S, i > |A, i > |cS
211
+ i |2 < S, i| < A, i|
212
+ (4)
213
+ It will be now shown that this is the time-asymptotic solution of the Lind-
214
+ bladian master equation properly parametrized.
215
+ We recall Lindblad’s equation for the time varying density of states operator
216
+ ρ ≡ ρ(t), as being of the following general form:
217
+ ∂ρ
218
+ ∂t = − i
219
+ ¯h[H, ρ] +
220
+
221
+ n
222
+ γn⟨LnρL†
223
+ n − 1
224
+ 2(L†
225
+ nLnρ + ρL†
226
+ nLn)⟩
227
+ (5)
228
+ The second term, here named the ”Lindblad term” [13, 14] though in different
229
+ contexts also referred to as the Dissipator [24], contains Ln’s that are Lindblad
230
+ jump-operators. We shall consistently work in the observable + pointer’s basis
231
+ (i.e., not in an energy basis). In this basis, neither the density operator ρ = ρ(t),
232
+ nor the A+S Hamiltonian H is diagonal at the beginning or in the course of
233
+ the development. But, as will be demonstrated, the Lindbladian formalism, by
234
+ a proper choice of its form, drives A+S to the desired diagonal form for the
235
+ combined observable +pointer basis. We postulate just one single term in the
236
+ previous n-sum, as well as off-diagonal forms, namely |i, j >< i′, j′|, (i, j ̸=
237
+ 5
238
+
239
+ i′, j′), for the jump-operators in the observable basis, leading to the following
240
+ parametrized form of the Lindblad term
241
+
242
+
243
+ ΓΩ
244
+
245
+ i′,j′̸=i,j
246
+ r(i, j)
247
+ r(i′, j′)⟨|i, j >< i′, j′|ρ|i′, j′ >< i, j|
248
+
249
+ 1
250
+ 2(|i′, j′ >< i, j|i, j >< i′, j′|ρ + ρ|i′, j′ >< i, j|i, j >< i′, j′|⟩
251
+ (6)
252
+ Here a circular frequency Ω is inserted, so as to make Γ , that quantifies the
253
+ strength of the system-environment coupling, dimensionless. One notes that in
254
+ the pre-factor appear the parameters r(i, j), r(i′, j′)(i, j, i′, j′ = 1, ..., I) whose
255
+ significance will be clear by deriving the matrix elements of the above operator.
256
+ These are
257
+ Lρi,j,i′,j′
258
+ =
259
+ δi,i′δj,j′r(i, j)
260
+
261
+ k,l
262
+ r−1(k, l)ρk,l,k,l
263
+
264
+ 1
265
+ 2[r−1(i, j) + r−1(i′, j′)]ρi,j,i′,j′
266
+
267
+ k,l
268
+ r(k, l)
269
+ (7)
270
+ It can be seen that the trace of the above vanishes and that each matrix element
271
+ vanishes upon the substitution
272
+ ρi,j,i′,j′ = δi,i′δj,j′r2(i, j)
273
+ (8)
274
+ While these properties hold for any arbitrary r(ij), the observable-pointer align-
275
+ ment is achieved by identifying the r parameters with the system’s superposition
276
+ coefficient: r(ij) = |cS
277
+ i |δi,j, or
278
+ r(i, j)2 = |cS
279
+ i |2 ≡ piδi,j
280
+ (9)
281
+ the last being the Born probabilities appearing in the collapsed state. As already
282
+ noted, this identification of probabilities relates to the well known procedure for
283
+ the Lindblad-induced thermalization of open systems, for which detailed balance
284
+ imposes the relation between the pre-factors γ(δE)/γ(−δE) = e−βδE/Z, the
285
+ latter being the canonical probabilities (with β = 1/kBT, kB the Boltzmann
286
+ constant, T the ambient temperature and Z the partition function [24, 25, 21])
287
+ .
288
+ [It also seems fair to point out that also in the standard (Copenhagen, or von
289
+ Neumannian) description of the alignment stage, as appears in e.g. Eq.(2.5) of
290
+ [1], this development is summarily stated, without specification of the underlying
291
+ mechanism.]
292
+ 5
293
+ Fast Decoherence Limit
294
+ We now consider the case that the time development in the state is predom-
295
+ inantly due to the coupling to the environment, rather than to the unitary
296
+ 6
297
+
298
+ change induced by the Hamiltonian, meaning that the second term on the right
299
+ hand side in Eq. 5 dominates the first. Quantitatively: Γ >> ||H||/¯hΩ. Ne-
300
+ glecting the commutator we now form matrix elements of the Lindblad term in
301
+ Eq. 5 in the observable+pointer basis. Because of the approximation made, the
302
+ off-diagonal matrix elements are decoupled from the diagonal ones. The master
303
+ equation of the off-diagonal terms reads (with a notation simplified by writing for
304
+ the index pairs i, j → r, i′, j′ → s and consequently for ρi,j,i′,j′ → ρrs ≡ ρrs(t)
305
+ dρrs
306
+ dt
307
+ = −ΓΩ
308
+ ��√pr + √ps
309
+ 2
310
+ ρrs
311
+
312
+ m
313
+ √pm
314
+
315
+ , r ̸= s
316
+ (10)
317
+ This shows that off-diagonal matrix elements decay exponentially in time (de-
318
+ cohere), maintaining their real character that they had initially. Had we kept
319
+ the (imaginary) commutator term, we would have found that the decay is mod-
320
+ ulated by the eigen-energies of the Hamiltonian.
321
+ For the diagonal matrix elements we find,
322
+ dρrr
323
+ dt
324
+ = ΓΩ
325
+
326
+ √pr
327
+
328
+ m
329
+ ρmm
330
+ √pm
331
+ − ρrr
332
+ √pr
333
+
334
+ m
335
+ √pm
336
+
337
+ (11)
338
+ Again, it can be seen that the trace of the last expression vanishes, and so
339
+ does the right-hand side under the substitution ρrr → pr. With these taking the
340
+ values as in Eq. 9, one arrives at the aligned form (written out in the original,
341
+ system-pointer indexes)
342
+ ρ(t → ∞) =
343
+
344
+ i
345
+ |ψA
346
+ i > |ψS
347
+ i > |cS
348
+ i |2 < ψS
349
+ i | < ψA
350
+ i |
351
+ (12)
352
+ 5.1
353
+ Illustrative example for a two-way experiment
354
+ Exemplifying the foregoing for a two-valued system (such as a 1
355
+ 2-spin electron),
356
+ prepared as an eigenstate of a Zeeman-field with the magnetic field inclined at an
357
+ angle 2αS to the vertical, in conjunction with an apparatus pointer, represented
358
+ as being likewise in an eigenstate of a quasi-Zeeman field inclined at an angle 2αA
359
+ to the vertical. The eigenstates are linear superpositions of their z- spins; these
360
+ are the observables that are to be determined by the measurement. Initially, the
361
+ system and the pointer are in the superposition states as shown above in Eqs.
362
+ 1 and 2 and whose superposition coefficients cS
363
+ i and cA
364
+ j now have the values,
365
+ sin / cos(αS) and sin / cos(αA), respectively. The DM in the observable basis
366
+ is now a 4x4 matrix, in which appear all the combinations of the products of
367
+ the above circular functions. As the outcome of the application of the Lindblad
368
+ operator in the rate equation, at long times the matrix becomes reduced to the
369
+ diagonal form discussed earlier. In these, cos2(αS) = p1 and sin2(αS) = p4
370
+ belonging to the aligned observable lie on the diagonal and are non-zero; the
371
+ other two diagonal entries for the anti-aligned situations are zero.
372
+ Plotted in Figure 1 are computed DM eigenvalues as functions of time
373
+ (in red and blue), normalized to their respective Born probabilities, showing
374
+ 7
375
+
376
+ -4
377
+ -3
378
+ -2
379
+ -1
380
+ 0
381
+ 1
382
+ 2
383
+ Log10time[invfrequn]
384
+ 0.2
385
+ 0.4
386
+ 0.6
387
+ 0.8
388
+ 1.0
389
+ 1.2
390
+ 1.4
391
+ P1,P2,Decoh
392
+ Figure 1: Density matrix eigenvalues normalized to their asymptotic (pointer-
393
+ aligned) values for the two aligned terms in the illustrative example (in red and
394
+ blue), plotted against time in inverse circular frequency unit. In green is shown
395
+ a decohering off-diagonal matrix element. Lindblad coupling strength Γ = 5, α
396
+ angles .37 π and .65 π.
397
+ their asymptotic convergence.
398
+ In green, the typical decohering tendency of
399
+ an off-diagonal element is demonstrated. Figure 2 depicts the entropy S(t) =
400
+ − �
401
+ r Pr(t) log Pr(t) of the system and apparatus-pointer, (in which Pr(t) are
402
+ computed eigenvalues of the DM.) The non-monotonic behavior is characteristic
403
+ of of the Lindblad formalism, in which the environment’s entropy change is not
404
+ taken into account.
405
+ [In numerical work, based on forward integration, putting zeros for some
406
+ of the pi’s introduces singularities, eventually algebraically cancelling out, but
407
+ preventing flow of computation. Therefore, instead, one puts arbitrarily small
408
+ values for these and obtains for the aligned DM one that is arbitrarily close to,
409
+ but not exactly equal to the true one.]
410
+ 6
411
+ Eigenvalue analysis
412
+ An alternative to the numerical solution of the differential rate equation is eigen-
413
+ value analysis, already treated in [15], based on the Landbladian term being a
414
+ linear function of the diagonals in the density matrix. Thereby, the resulting
415
+ rate equations have solution of the form
416
+ ρnn(t) =
417
+
418
+ k
419
+ vn,keλkt
420
+ (13)
421
+ in which λk and vn,k are the diagonalized eigenvalues and eigenvectors of the
422
+ Lindbladian matrix diagonals in Eq. 11. Calculation shows that for the 4 x
423
+ 8
424
+
425
+ -4
426
+ -3
427
+ -2
428
+ -1
429
+ 0
430
+ Log10time[invfrequn]
431
+ 0.05
432
+ 0.10
433
+ 0.15
434
+ 0.20
435
+ S
436
+ Figure 2: Entropy of the combined system plus apparatus. Noteworthy is the
437
+ initial peak common to the Lindblad formalism .
438
+ 4 matrix considered above there are three negative eigenvalues and one zero
439
+ eigenvalue, which alone is of interest at the long term behavior. Belonging to
440
+ this eigenvalue, the (transposed) eigenvector is found to be {p1, p2, p3, p4} ≈
441
+ {cS
442
+ 1 , 0, 0, cS
443
+ 4 }, as required for the alignment between the quantum states and the
444
+ reading in the measuring apparatus.
445
+ 6.1
446
+ Measurement speed
447
+ Figure 1 shows that alignment is achieved for the model with the chosen strength
448
+ parameter (Γ = 5) by a time of cca. 0.1/Ω. By varying the strength in the
449
+ computed model, we find a shortening of this time that is inversely proportional
450
+ to the strength. This is expected from the quantum speed limit (QSL) results
451
+ that border quantum transition times τ from below.
452
+ Essentially, QSL is the ratio of two norms [26, 27], that of the ”quantum
453
+ distance” [28] and of the speed of the state evolution. Formally
454
+ τ > ||ρ(t → ∞) − ρ(t = 0)||
455
+ || dρ(t)
456
+ dt ||
457
+ = ||ρ(t → ∞) − ρ(t = 0)||
458
+ ||[Lρ(t)]||
459
+ (14)
460
+ Ways of calculating the norms vary, e.g., [29, 30]. Recently, for a system de-
461
+ veloping due to a Lindbalian operator, three contributions to the speed were
462
+ discerned [24]. To estimate ||ρ(t → ∞) − ρ(t = 0)|, we have used the ”Trace
463
+ Distance ”defined as
464
+ T(ρ, σ) = 1
465
+ 2Tr[
466
+
467
+ (ρ − σ)] = 1
468
+ 2
469
+
470
+ i
471
+ |µi|
472
+ (15)
473
+ [31], where µi are the eigenvalues of the matrix differences. The DM velocity,
474
+ as defined above , changes (decreases) with time, ultimately vanishing at the
475
+ 9
476
+
477
+ fulfilment of alignment; we have taken the root-mean-square sum of the rate of
478
+ the diagonal matrix elements at initial times. These yield a very low limit of
479
+ τ > 1
480
+ 2
481
+ 1.41
482
+ 5 ∗ 69.03 = .0045/Ω
483
+ (16)
484
+ to be compared with the actually computed value, about 20 times longer. Better
485
+ (higher) limits of transition times may be generated by different ways of forming
486
+ the norm for the DM velocity (e.g. not at the beginning).
487
+ 6.2
488
+ Multiple Reading-System correspondence
489
+ A simple generalization of the foregoing applies when each (eigen-)value of the
490
+ observable is in correspondence with not just one reading of the pointer, but
491
+ with several (say, R) readings, all of the same significance for the outcome. Then
492
+ one simply inserts pi/R into the corresponding Lindblad term, in place of just
493
+ pi. In the more complex case, that not all readings have the same likelihood,
494
+ pi would have to be weighted by a probability factot, rather than by a constant
495
+ denominator.
496
+ 7
497
+ The Lindbladian, ”Who ordered this?”
498
+ Historically, Lindblad terms were introduced as the most general forms that
499
+ maintain complete positivity of the DM’s and preserve their trace [13, 14]. The
500
+ various derivations that have been presented (and among these a recent one by
501
+ [32]), involve several approximations for the coupling between the system and its
502
+ environment. Insomuch that the derivation involves also tracing over the degrees
503
+ of freedom of the environment, much detail of the latter is lost and of course
504
+ it is impossible to work backwards from the Lindbladian to the environment.
505
+ What is remarkable is that for special purposes the appropriate Lindbladian
506
+ operators take a very special, practically unique form. Such is the case for the
507
+ accepted description of thermalization [21, 24] by a Lindblad formalism. The
508
+ parametrization of the Lindblad term employed in the present work, though it
509
+ may appear arbitrary and particular for each case, is in fact identical to the
510
+ one used for thermalization subject to the relabelling of the Gibbsian thermal
511
+ distribution function as (the Born) probabilities (p1, p2, ...),with the proviso of
512
+ working in the observable, rather than in the energy basis.
513
+ (This contrasts
514
+ with the different approach in [15], which claims attainment of collapse for any
515
+ Lindbladian operator.) At the same time, it needs to be noted that the analog
516
+ of detailed balance is missing in wave-function collapse.
517
+ How come to have
518
+ such a specific Lindbladian, whose source may be any measurement device and
519
+ procedure? One is left to wonder about the possibility of a special meta-physical
520
+ status of the Lindblad terms, or query with Wheeler ”Who ordered this?”
521
+ 10
522
+
523
+ 8
524
+ Conclusion
525
+ The well-known Lindbladian extension to the quantum theory of motion to
526
+ environmental effects is here adapted to establish the resolution of a wave-packet
527
+ in a measurement as a smooth process. This is enabled by an unambiguous
528
+ parametrization of the jump-operators describing the interaction of the broad
529
+ environment with the observed system, both regarded as quantal entities.
530
+ Above, in section 2.1, a brief historically oriented preview has been provided
531
+ for the distinct approach in this work, namely, one based on the wholeness of
532
+ the entities (observed system and observing device), through the (Lindbladian)
533
+ equation yielding the evolution of the system.
534
+ A main result emerging from the formalism, and capable of experimental
535
+ verification, is the finitely temporal variation of the system, and this in a de-
536
+ terministic way rather than just statistically, on the average, contrasting also
537
+ with the instantaneous collapse description by (e.g.) von Neumann. Such tem-
538
+ poral variation in continuous-thermalization processes has been proposed quite
539
+ recently [33, 34], also by employment of a Lindbladian formalism and within a
540
+ Markovian framework.
541
+ Experimentally, verification of the time dependence of the transition in any
542
+ particular measurement, implicit in our formulae, could be observed by re-
543
+ peated observation performed on the system subject to non-demolition tran-
544
+ sitions. These observations would be akin to the Zeno-effect measurement, such
545
+ as has been achieved in the form of quasi-periodic oscillation of the result for
546
+ a Superconducting flux cubit[35]. Further work is needed for quantifying the
547
+ information-entropy change in the environment [31, 36].
548
+ Acknowledgement
549
+ The authors thank the referee for meticulous reading and insightful ques-
550
+ tioning of a previous version of this paper.
551
+ References
552
+ [1] A.J. Leggett, Macroscopic quantum systems and the quantum theory of
553
+ measurement. Suppl. Progress Theor. Phys. 69, 80 (1980)
554
+ [2] G.C. Ghirardi, P. Pearle and A. Rimini, Markov processes in Hilbert space
555
+ and continuous spontaneous localization of systems of identical particles.
556
+ Phys. Rev. A 42, 78 (1990)
557
+ [3] E.P. Wigner, Remarks on the mind-body question. In I. J. Good (ed.) ”The
558
+ Scientist Speculates”(Heinemann, London,1961)
559
+ 11
560
+
561
+ [4] W.H. Louisell, Quantum Statistical Properties of Radiation. (Wiley, New
562
+ York, 1973)
563
+ [5] G.C. Ghirardi, A. Rimini and T.Weber, Unified dynamics for microscopic
564
+ and macroscopic systems, Phys. Rev. D 34, 470 (1986)
565
+ [6] P. Pearle, Combining stochastic, dynamical state reduction with sponta-
566
+ neous localization, Phys. Rev. A 39 2277-89 (1989)
567
+ [7] H. Everett, Relative state formulation of quantum mechanics, Rev. Mod.
568
+ Phys. 29 454-462 (1957)
569
+ [8] I. Byalinicki-Birula and J. Mycielski, Nonlinear wave mechanics, Annals
570
+ Phys. (NY) 108 62-83 (1976).
571
+ [9] G. Sewell, On the mathematical structure of quantum measurement theory,
572
+ Repts. Math. Phys. 56 271-90 (2005)
573
+ [10] L.Di´osi, Models for universal reduction of macroscopic quantum fuctua-
574
+ tions. Phys. Rev. A 40 165 (1989)
575
+ [11] R. Penrose, On gravity’s role in quantum state reduction. Gen. Relativ.
576
+ Gravit. 28, 581 (1996).
577
+ [12] S. Donadi, K. Piscicchia, C. Curceanu, L. Di´osi, M. Laubenstein and A.
578
+ Bassi, Underground test of gravity-related wave function collapse. Nat. Phys.
579
+ 17 74 (2021)
580
+ [13] G. Lindblad, On the generators of quantum dynamical semigroups, Com-
581
+ mun. Math. Phys., 48 112-30 (1976).
582
+ [14] V. Gorini, A. Kossakowski and E.C.G. Sudarshan, Completely positive
583
+ dynamical semigroups of N-level systems, J. Math. Phys. 17 821-5 (1976).
584
+ [15] S. Weinberg, What happens in a measurement?, Phys. Rev. A 93, 032124
585
+ 1-6 (2016)
586
+ [16] M.A. Nielsen and I.L. Chuang, Quantun Computation and Quantum In-
587
+ formation (Cambridge University Press, Cambridge, U.K., 2010)
588
+ [17] K. Jacobs and D.A. Steck, A straightforward introduction to continu-
589
+ ous quantum measurement, Contemporary Physics, 47 (5) 279-303 (2006).
590
+ Arxiv.org/abs/quant-ph/0611067.
591
+ [18] N. Bohr, Atomic Physics and Human Knowledge, (Wiley, New York, 1958)
592
+ [19] J. Bell, Physics World, 3 33-4 (1990).
593
+ [20] A. Peres, Quantum Theory: Concepts and Methods, (Kluwer Academic,
594
+ Norwell MA, 1998)
595
+ 12
596
+
597
+ [21] H.-P. Breuer and F. F. Petruccione, The Theory of Open Quantum Systems
598
+ (Oxford University Press, New York, 2002)
599
+ [22] S. B. J¨ager , T. Schmit, G. Morigi , M. J. Holland and R. Betzholz, Lindblad
600
+ Master Equations for Quantum Systems Coupled to Dissipative Bosonic
601
+ Modes, Phys. Rev. Lett. 129 063601 1-7 (2022)
602
+ [23] D. Leibfried, R. Blatt, C. Monroe, and D. Wineland, (2003). Quantum
603
+ dynamics of single trapped ions. Rev. Mod. Phys. 75 (1), 281–324 (2003).
604
+ Doi:10.1103/RevModPhys.75.281.
605
+ [24] K. Funo, N. Shiraishi and K. Saito, Speed limit for open quantum systems,
606
+ New J. Phys. 21 0300611 1-9 (2019).
607
+ [25] J.M. Horowitz and J.M.R. Parrondo, Entropy production along non-
608
+ equilibrium jump trajectories, New J. Phys. 15 220601 1-23 (2013). )
609
+ [26] L. I. Mandelshtam and I. E. Tamm, The uncertainty relation between en-
610
+ ergy and time in nonrelativistic quantum mechanics. J. Phys. (USSR) 9, 249
611
+ (1945).
612
+ [27] N. Margulus and L. B. Levitin, The maximum speed of dynamical evolu-
613
+ tion. Physica D 120, 188 (1998).
614
+ [28] S. L. Braunstein and C. M. Caves, Statistical distance and the geometry
615
+ of quantum states. Phys. Rev. Lett. 72 3439-44 (1994)
616
+ [29] S. Deffner and S. Campbell, Quantum speed limits: from Heisenberg’s
617
+ uncertainty principle to optimal quantum control. J. Phys. A: Math. Theor.
618
+ 50 453001 (2017)
619
+ [30] J. Anandan and Y. Aharonov, Geometry of quantum evolution. Phys. Rev.
620
+ Lett. 65 1697-701 (1990)
621
+ [31] S.M. Barnett, Quantum Information,(University Press, Oxford, 2009),
622
+ Chapter 4.
623
+ [32] D. Manzano, A short introduction to the Lindblad master equation. AIP
624
+ Adv. 10 (2) 1063 (2020); doi.org/10.1063/1.5115323
625
+ [33] K. Korzekwa and M.Lostaglio, Optimizing thermalization, Phys. Rev.Lett.
626
+ 129 040602 1-7 (2022)
627
+ [34] M.Lostaglio and K. Korzekwa, Continuous thermomajorizaton and a com-
628
+ plete set of Markovian thermal processes. Phys. Rev, A 106 012426 1-18
629
+ (2022)
630
+ [35] K. Kakuyanagi, T. Baba, Y. Matsuzaki, H. Nakano, S. Saito and K. Semba,
631
+ Observation of quantum Zeno effect in a superconducting flux qubit, New
632
+ J. Phys. 17 063035 1-10 (2015)
633
+ [36] M. Keyl, Fundamentals of quantum information theory, Physics Reports
634
+ 369 (5) 431-58 (2002)
635
+ 13
636
+
EdE0T4oBgHgl3EQfywKq/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,471 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf,len=470
2
+ page_content='Lindbladian-Induced Alignment in Quantum Measurements R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
3
+ page_content=' Englman and A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
4
+ page_content=' Yahalom Ariel University, Ariel 40700,Israel January 10, 2023 Keywords: Quantum measurement theory, Density matrix evolution, Quan- tum state resolution, Lindblad operators, Quantum speed limit.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
5
+ page_content=' Abstract An expression of the Lindbladian form is proposed that ensures an un- ambiguous time-continuous reduction of the initial system-pointer wave- packet to one in which the readings and the observable’s values are aligned, formalized as the transition from an outer product to an inner product of the system’s and apparatus’ density matrices.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
6
+ page_content=' The jump operators are in the basis of the observables, with uniquely determined parameters derived from the measurement set-up (thereby differing from S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
7
+ page_content=' Weinberg’s Lind- bladian resolution of wave-packet formalism) and conforming to Born’s probability rules.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
8
+ page_content=' The novelty lies in formalising the adaptability of the surroundings (including the measuring device) to the mode of observa- tion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
9
+ page_content=' Accordingly, the transition is of finite duration (in contrast to its instantaneousness in the von Neumann’s formulation).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
10
+ page_content=' This duration is estimated for a simple half-spin-like model.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
11
+ page_content=' 1 Introduction In the century-run of quantum physics (plus 4 years, if one marks its beginning with the award of a Nobel Prize in 1918 to Max Planck for ”his discovery of quanta”) a single shadow of non-sequitur has darkened its glorious achievements, one that goes variously under the names of wave-function collapse, reduction of the wave-packet, quantum measurement, einselection, etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
12
+ page_content=' Aspects of the prob- lem (or its articulations) were manifold, such as the breakdown of the predicted time-development in accordance with the Schr¨odinger equation, the abruptness of change in a measurement (”natura non facit saltum”, where art thou?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
13
+ page_content=' ), the apparent non-applicability of quantum rules to macroscopic systems, imputed arbitrariness of Born’s probability rules, the requirement of ”infinite regress” 1 arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
14
+ page_content='02664v1 [quant-ph] 6 Jan 2023 for the measuring apparatus and others.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
15
+ page_content=' Numerous papers enlarged on these is- sues [1, 2] and various proposals for resolution of the problem were put forward.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
16
+ page_content=' These include the observer’s cognition [3], stochastic effects [4], in particular spontaneous localization [2, 5, 6], a many world scenario [7], non-linearity addi- tion to the Schr¨odinger equation [8], Poincar´e recurrent state [9], gravitationally induced collapse [10, 11, 12], etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
17
+ page_content=' Common to these works, and with the specific purpose of providing a blue-print for measurements compatible with the Copenhagen formulation of quantum theory, was the need to give expression to the coupling of the micro- scopic system with its macroscopic environment.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
18
+ page_content=' Standing apart from these and belonging to the field of non-equilibrium thermodynamics and to the establish- ment of equilibrium, a general form for this interaction was given by Lindblad [13] and by Gorin , Kossakowski and Sudarshan [14], satisfying some necessary conditions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
19
+ page_content=' Constructing a merger between the two separately oriented fields, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
20
+ page_content=' Weinberg recently proposed a Lindblad-operator mechanism for the collapse of the density matrix (DM) in the course of a complete measurement [15].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
21
+ page_content=' No- tably, the mechanism was linear in the state’s DM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
22
+ page_content=' The collapsed state (Eq.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
23
+ page_content=' (1) in [15]) comprises the set of projection operators of the measurable item;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
24
+ page_content=' the system’s Hamiltonian is described by a spectral decomposition onto the same operators (Eq.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
25
+ page_content=' (16) in [15]) (although in the verbal discussion a more gen- eral situation is considered): collapse is achieved ”independent[ly] of the details of these [Lindblad] operators”.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
26
+ page_content=' Decay between energy eigenstates had earlier been treated by the Lindblad formalism (for a pedagogical presentation the volume [16], Chapter 8 may be consulted) employing the interaction represen- tation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
27
+ page_content=' However, this is not convenient for treating measurements of observables that do not commute with the Hamiltonian.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
28
+ page_content=' Detailed theories relate to the out- come (”mapping”) of quantum operations, including measurements;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
29
+ page_content=' the present work describes the process of these happening.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
30
+ page_content=' (For a pedagogical introduction to stochasticity-induced wave-packet- reduction, obviating pointer reading, one may refer to [17].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
31
+ page_content=') 2 Overview of the Method and Terms 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
32
+ page_content='1 The leading idea, also in review While the concept of unity of observer and observation had already featured in Bohr’s view: ”The answer that we get is built up from the combined interac- tion of [the observer’s] state and the object of interrogation.” [18], this was not given a formal expression in the Copenhagen interpretation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
33
+ page_content=' It was more em- phatically asserted both by J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
34
+ page_content=' Bell: ”I meant that the ’apparatus’ should not be severed from the rest of the world in boxes .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
35
+ page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
36
+ page_content='[19]” and A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
37
+ page_content=' Peres: ” A measure- ment both creates and records a property of the system [20]”.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
38
+ page_content=' This change in 2 the course of a measurement affects also the environment outside the observed system ;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
39
+ page_content=' in the words of A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
40
+ page_content=' Leggett ”.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
41
+ page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
42
+ page_content='under these conditions the macroscopic apparatus, and more generally any part of the macro-world which has suffered changes in the course of the measurement process, does not end up in a state with definite macroscopic properties at all,.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
43
+ page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
44
+ page_content=' [1]”.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
45
+ page_content=' The same line of thought appears to motivate S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
46
+ page_content=' Weinberg, who wrote in his preamble to a 2016 Lindbladian formulation of the masurement process[15], that ”We will instead [of the original formulation of the Copenhagen interpretation, (which we will not dwell on here)] adopt the popular modern view that the Copenhagen interpretation refers to open systems in which the transition is driven by the ineraction of the microscopic system under study (which may include an observer) chosen to bring the transition about.” (Our italics.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
47
+ page_content=') These developments indicate the justification for a formulation in which the effect of the apparatus is incorporated in the equation defining the evolution of the system, rather than one in which the two entities are separate, barring an interaction between them.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
48
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
49
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
50
+ page_content='1 ”Alignment” The process whereby the pointer readings become in correspondence with the possible values of the observable.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
51
+ page_content=' Formally, for I possible values, the combined density matrix reduces from comprising I2 terms to one having I terms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
52
+ page_content=' (E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
53
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
54
+ page_content=', equation (2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
55
+ page_content='5) in [1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
56
+ page_content=') 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
57
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
58
+ page_content='2 ”Dissipator” Added term (in the form of sums of appropriately weighted jump-operators) to the standard time dependent Schr¨odinger equation, inducing non-unitary evolution in the system, accompanied by changes of its information entropy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
59
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
60
+ page_content='2 Motivation for the choice of formalism Thermalization of open systems can be described by a Lindbladian formal- ism in which Gibbsian probabilities are so inserted as parameters, that the ”Dissipator” vanishes at these values of the density matrix.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
61
+ page_content=' Replacement of the Gibbsian probabilities by Born probabilities achieves alignment in a state reduction and does so continuously.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
62
+ page_content=' Limitations: Born’s probability rules are assumed, not derived;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
63
+ page_content=' the interac- tion term is not traced to a microscopic mechanism.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
64
+ page_content=' The source of this interaction term, shown in Eqn.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
65
+ page_content=' 6 below, incorporating the coupling between the observed system and its surroundings (including the 3 measuring device) is an open question (also raised by a referee).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
66
+ page_content=' In its appli- cation to a thermalization process, the Lindbladian jump operators have been derived, though with the aids of several approximations (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
67
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
68
+ page_content=',[21]), as well as, more recently, for the dissipation in a Dicke system with a bosonic background [22].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
69
+ page_content=' We are not in the position to provide such first principle derivation for the Lindbladian jump-operators bringing about a transition and incorporat- ing the Born rules.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
70
+ page_content=' It seems to be specific to the type of measurement under consideration and it is clear that just any jump operator, as in Weinberg’s Lindbladian formulation will not do the job .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
71
+ page_content=' Likely, one would need to in- clude non-Markovian dynamics, so that the coupling to the device and eventual pointer reading are two separate consecutive events.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
72
+ page_content=' Inclusion of such dynamics is outside the scope of the present work.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
73
+ page_content=' 3 Assumptions We explore the time (t)-development of the combined density matrix ρ(t) of the measured system (S) and of the reading (pointer, dial, etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
74
+ page_content=') on the mea- suring apparatus (A) for a complete and discrete measurement , expressing the underlying assumptions by three propositions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
75
+ page_content=' Proposition 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
76
+ page_content=' In accord with the long-time historical approach, the mea- sured object S and the pointer of the measuring set-up A are treated on equal footings as subject to microscopic quantum laws, and formally describable by their respective Hamiltonians.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
77
+ page_content=' Aware of the difficulties connected with an ”in- finite regress”, the effects of the rest of the Universe on S+A are not included in the formalism;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
78
+ page_content=' instead, for a phenomenological, approximative description, a Lindbladian term appears in the master equation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
79
+ page_content=' Proposition 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
80
+ page_content=' Prior to the measurement with A and S decoupled, and being free of external influence for a long time, both are in energy quantum states, pure or mixed.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
81
+ page_content=' After the measurement, the state is not an energy eigenstate and subsequently it will spread over to a superposition of energy eigenstates.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
82
+ page_content=' The fast decoherence case treated below in section 5 is akin to the Zeno effect [23].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
83
+ page_content=' Proposition 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
84
+ page_content=' Only those states of the reading apparatus (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
85
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
86
+ page_content=', the right or left positions of a pointer) that may be in direct correspondence with the measured states of the system (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
87
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
88
+ page_content=', spin up or down) are given expression in the formalism.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
89
+ page_content=' (At a beginning, the case treated is one in which there is a one- to-one correspondence between the states of the system and the readings of the apparatus;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
90
+ page_content=' a generalization is given subsequently.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
91
+ page_content=') A discussion in section 8 touches on the epistemological status of the Lindbladian terms in a measurement process.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
92
+ page_content=' 4 4 Analysis Considering (for simplicity) a pure state for the system, its initial state-vector written in the basis of the observed property |S, i > takes the form ψS(t = 0) = � i=1,.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
93
+ page_content='.,I cS i |S, i > (1) Born’s rule for the probability of observing the i-component is |cS i ]2 ≡ pi, sum- ming to unity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
94
+ page_content=' Likewise, for the apparatus readings j, numbering J, one has the superposition with (complex and normalized) coefficients cA j ψA(t = 0) = � j=1,.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
95
+ page_content='.,J cA j |A, j > (2) We start with the one-to-one correspondence situation, for which I = J, and the reading j on A establishes uniquely the value i = j for the system’s measured property.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
96
+ page_content=' For the combined state-vector the density operator has the outer-product form (where the stars denote complex conjugates): � i,j,i′,j′ |S, i > |A, j > cS∗ i cA∗ j cA j′cS i′ < A, j′| < S, i′| ≡ � i,j,i′,j′ |i, j > Ciji′j′ < i′, j′| (3) the right hand side written in an obvious shortened notation, in which Ciji′j′ = cS∗ i cA∗ j cA j′cS i′.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
97
+ page_content=' After collapse, the density operator takes the aligned, single-sum form � i |S, i > |A, i > |cS i |2 < S, i| < A, i| (4) It will be now shown that this is the time-asymptotic solution of the Lind- bladian master equation properly parametrized.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
98
+ page_content=' We recall Lindblad’s equation for the time varying density of states operator ρ ≡ ρ(t), as being of the following general form: ∂ρ ∂t = − i ¯h[H, ρ] + � n γn⟨LnρL† n − 1 2(L† nLnρ + ρL† nLn)⟩ (5) The second term, here named the ”Lindblad term” [13, 14] though in different contexts also referred to as the Dissipator [24], contains Ln’s that are Lindblad jump-operators.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
99
+ page_content=' We shall consistently work in the observable + pointer’s basis (i.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
100
+ page_content='e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
101
+ page_content=', not in an energy basis).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
102
+ page_content=' In this basis, neither the density operator ρ = ρ(t), nor the A+S Hamiltonian H is diagonal at the beginning or in the course of the development.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
103
+ page_content=' But, as will be demonstrated, the Lindbladian formalism, by a proper choice of its form, drives A+S to the desired diagonal form for the combined observable +pointer basis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
104
+ page_content=' We postulate just one single term in the previous n-sum,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
105
+ page_content=' as well as off-diagonal forms,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
106
+ page_content=' namely |i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
107
+ page_content=' j >< i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
108
+ page_content=' j′|,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
109
+ page_content=' (i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
110
+ page_content=' j ̸= 5 i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
111
+ page_content=' j′),' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
112
+ page_content=' for the jump-operators in the observable basis,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
113
+ page_content=' leading to the following parametrized form of the Lindblad term Lρ ≡ ΓΩ � i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
114
+ page_content='j′̸=i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
115
+ page_content='j r(i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
116
+ page_content=' j) r(i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
117
+ page_content=' j′)⟨|i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
118
+ page_content=' j >< i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
119
+ page_content=' j′|ρ|i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
120
+ page_content=' j′ >< i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
121
+ page_content=' j| − 1 2(|i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
122
+ page_content=' j′ >< i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
123
+ page_content=' j|i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
124
+ page_content=' j >< i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
125
+ page_content=' j′|ρ + ρ|i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
126
+ page_content=' j′ >< i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
127
+ page_content=' j|i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
128
+ page_content=' j >< i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
129
+ page_content=' j′|⟩ (6) Here a circular frequency Ω is inserted,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
130
+ page_content=' so as to make Γ ,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
131
+ page_content=' that quantifies the strength of the system-environment coupling,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
132
+ page_content=' dimensionless.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
133
+ page_content=' One notes that in the pre-factor appear the parameters r(i, j), r(i′, j′)(i, j, i′, j′ = 1, .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
134
+ page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
135
+ page_content=', I) whose significance will be clear by deriving the matrix elements of the above operator.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
136
+ page_content=' These are Lρi,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
137
+ page_content='j,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
138
+ page_content='i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
139
+ page_content='j′ = δi,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
140
+ page_content='i′δj,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
141
+ page_content='j′r(i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
142
+ page_content=' j) � k,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
143
+ page_content='l r−1(k,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
144
+ page_content=' l)ρk,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
145
+ page_content='l,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
146
+ page_content='k,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
147
+ page_content='l − 1 2[r−1(i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
148
+ page_content=' j) + r−1(i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
149
+ page_content=' j′)]ρi,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
150
+ page_content='j,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
151
+ page_content='i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
152
+ page_content='j′ � k,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
153
+ page_content='l r(k,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
154
+ page_content=' l) (7) It can be seen that the trace of the above vanishes and that each matrix element vanishes upon the substitution ρi,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
155
+ page_content='j,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
156
+ page_content='i′,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
157
+ page_content='j′ = δi,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
158
+ page_content='i′δj,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
159
+ page_content='j′r2(i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
160
+ page_content=' j) (8) While these properties hold for any arbitrary r(ij),' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
161
+ page_content=' the observable-pointer align- ment is achieved by identifying the r parameters with the system’s superposition coefficient: r(ij) = |cS i |δi,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
162
+ page_content='j,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
163
+ page_content=' or r(i,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
164
+ page_content=' j)2 = |cS i |2 ≡ piδi,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
165
+ page_content='j (9) the last being the Born probabilities appearing in the collapsed state.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
166
+ page_content=' As already noted, this identification of probabilities relates to the well known procedure for the Lindblad-induced thermalization of open systems, for which detailed balance imposes the relation between the pre-factors γ(δE)/γ(−δE) = e−βδE/Z, the latter being the canonical probabilities (with β = 1/kBT, kB the Boltzmann constant, T the ambient temperature and Z the partition function [24, 25, 21]) .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
167
+ page_content=' [It also seems fair to point out that also in the standard (Copenhagen, or von Neumannian) description of the alignment stage, as appears in e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
168
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
169
+ page_content=' Eq.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
170
+ page_content=' (2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
171
+ page_content='5) of [1], this development is summarily stated, without specification of the underlying mechanism.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
172
+ page_content='] 5 Fast Decoherence Limit We now consider the case that the time development in the state is predom- inantly due to the coupling to the environment, rather than to the unitary 6 change induced by the Hamiltonian, meaning that the second term on the right hand side in Eq.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
173
+ page_content=' 5 dominates the first.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
174
+ page_content=' Quantitatively: Γ >> ||H||/¯hΩ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
175
+ page_content=' Ne- glecting the commutator we now form matrix elements of the Lindblad term in Eq.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
176
+ page_content=' 5 in the observable+pointer basis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
177
+ page_content=' Because of the approximation made, the off-diagonal matrix elements are decoupled from the diagonal ones.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
178
+ page_content=' The master equation of the off-diagonal terms reads (with a notation simplified by writing for the index pairs i, j → r, i′, j′ → s and consequently for ρi,j,i′,j′ → ρrs ≡ ρrs(t) dρrs dt = −ΓΩ �√pr + √ps 2 ρrs � m √pm � , r ̸= s (10) This shows that off-diagonal matrix elements decay exponentially in time (de- cohere), maintaining their real character that they had initially.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
179
+ page_content=' Had we kept the (imaginary) commutator term, we would have found that the decay is mod- ulated by the eigen-energies of the Hamiltonian.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
180
+ page_content=' For the diagonal matrix elements we find, dρrr dt = ΓΩ � √pr � m ρmm √pm − ρrr √pr � m √pm � (11) Again, it can be seen that the trace of the last expression vanishes, and so does the right-hand side under the substitution ρrr → pr.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
181
+ page_content=' With these taking the values as in Eq.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
182
+ page_content=' 9, one arrives at the aligned form (written out in the original, system-pointer indexes) ρ(t → ∞) = � i |ψA i > |ψS i > |cS i |2 < ψS i | < ψA i | (12) 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
183
+ page_content='1 Illustrative example for a two-way experiment Exemplifying the foregoing for a two-valued system (such as a 1 2-spin electron), prepared as an eigenstate of a Zeeman-field with the magnetic field inclined at an angle 2αS to the vertical, in conjunction with an apparatus pointer, represented as being likewise in an eigenstate of a quasi-Zeeman field inclined at an angle 2αA to the vertical.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
184
+ page_content=' The eigenstates are linear superpositions of their z- spins;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
185
+ page_content=' these are the observables that are to be determined by the measurement.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
186
+ page_content=' Initially, the system and the pointer are in the superposition states as shown above in Eqs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
187
+ page_content=' 1 and 2 and whose superposition coefficients cS i and cA j now have the values, sin / cos(αS) and sin / cos(αA), respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
188
+ page_content=' The DM in the observable basis is now a 4x4 matrix, in which appear all the combinations of the products of the above circular functions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
189
+ page_content=' As the outcome of the application of the Lindblad operator in the rate equation, at long times the matrix becomes reduced to the diagonal form discussed earlier.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
190
+ page_content=' In these, cos2(αS) = p1 and sin2(αS) = p4 belonging to the aligned observable lie on the diagonal and are non-zero;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
191
+ page_content=' the other two diagonal entries for the anti-aligned situations are zero.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
192
+ page_content=' Plotted in Figure 1 are computed DM eigenvalues as functions of time (in red and blue), normalized to their respective Born probabilities, showing 7 4 3 2 1 0 1 2 Log10time[invfrequn] 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
193
+ page_content='2 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
194
+ page_content='4 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
195
+ page_content='6 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
196
+ page_content='8 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
197
+ page_content='0 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
198
+ page_content='2 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
199
+ page_content='4 P1,P2,Decoh Figure 1: Density matrix eigenvalues normalized to their asymptotic (pointer- aligned) values for the two aligned terms in the illustrative example (in red and blue), plotted against time in inverse circular frequency unit.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
200
+ page_content=' In green is shown a decohering off-diagonal matrix element.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
201
+ page_content=' Lindblad coupling strength Γ = 5, α angles .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
202
+ page_content='37 π and .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
203
+ page_content='65 π.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
204
+ page_content=' their asymptotic convergence.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
205
+ page_content=' In green, the typical decohering tendency of an off-diagonal element is demonstrated.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
206
+ page_content=' Figure 2 depicts the entropy S(t) = − � r Pr(t) log Pr(t) of the system and apparatus-pointer, (in which Pr(t) are computed eigenvalues of the DM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
207
+ page_content=') The non-monotonic behavior is characteristic of of the Lindblad formalism, in which the environment’s entropy change is not taken into account.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
208
+ page_content=' [In numerical work, based on forward integration, putting zeros for some of the pi’s introduces singularities, eventually algebraically cancelling out, but preventing flow of computation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
209
+ page_content=' Therefore, instead, one puts arbitrarily small values for these and obtains for the aligned DM one that is arbitrarily close to, but not exactly equal to the true one.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
210
+ page_content='] 6 Eigenvalue analysis An alternative to the numerical solution of the differential rate equation is eigen- value analysis, already treated in [15], based on the Landbladian term being a linear function of the diagonals in the density matrix.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
211
+ page_content=' Thereby, the resulting rate equations have solution of the form ρnn(t) = � k vn,keλkt (13) in which λk and vn,k are the diagonalized eigenvalues and eigenvectors of the Lindbladian matrix diagonals in Eq.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
212
+ page_content=' 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
213
+ page_content=' Calculation shows that for the 4 x 8 4 3 2 1 0 Log10time[invfrequn] 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
214
+ page_content='05 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
215
+ page_content='10 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
216
+ page_content='15 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
217
+ page_content='20 S Figure 2: Entropy of the combined system plus apparatus.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
218
+ page_content=' Noteworthy is the initial peak common to the Lindblad formalism .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
219
+ page_content=' 4 matrix considered above there are three negative eigenvalues and one zero eigenvalue, which alone is of interest at the long term behavior.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
220
+ page_content=' Belonging to this eigenvalue, the (transposed) eigenvector is found to be {p1, p2, p3, p4} ≈ {cS 1 , 0, 0, cS 4 }, as required for the alignment between the quantum states and the reading in the measuring apparatus.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
221
+ page_content=' 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
222
+ page_content='1 Measurement speed Figure 1 shows that alignment is achieved for the model with the chosen strength parameter (Γ = 5) by a time of cca.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
223
+ page_content=' 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
224
+ page_content='1/Ω.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
225
+ page_content=' By varying the strength in the computed model, we find a shortening of this time that is inversely proportional to the strength.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
226
+ page_content=' This is expected from the quantum speed limit (QSL) results that border quantum transition times τ from below.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
227
+ page_content=' Essentially, QSL is the ratio of two norms [26, 27], that of the ”quantum distance” [28] and of the speed of the state evolution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
228
+ page_content=' Formally τ > ||ρ(t → ∞) − ρ(t = 0)|| || dρ(t) dt || = ||ρ(t → ∞) − ρ(t = 0)|| ||[Lρ(t)]|| (14) Ways of calculating the norms vary, e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
229
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
230
+ page_content=', [29, 30].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
231
+ page_content=' Recently, for a system de- veloping due to a Lindbalian operator, three contributions to the speed were discerned [24].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
232
+ page_content=' To estimate ||ρ(t → ∞) − ρ(t = 0)|, we have used the ”Trace Distance ”defined as T(ρ, σ) = 1 2Tr[ � (ρ − σ)] = 1 2 � i |µi| (15) [31], where µi are the eigenvalues of the matrix differences.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
233
+ page_content=' The DM velocity, as defined above , changes (decreases) with time, ultimately vanishing at the 9 fulfilment of alignment;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
234
+ page_content=' we have taken the root-mean-square sum of the rate of the diagonal matrix elements at initial times.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
235
+ page_content=' These yield a very low limit of τ > 1 2 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
236
+ page_content='41 5 ∗ 69.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
237
+ page_content='03 = .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
238
+ page_content='0045/Ω (16) to be compared with the actually computed value, about 20 times longer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
239
+ page_content=' Better (higher) limits of transition times may be generated by different ways of forming the norm for the DM velocity (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
240
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
241
+ page_content=' not at the beginning).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
242
+ page_content=' 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
243
+ page_content='2 Multiple Reading-System correspondence A simple generalization of the foregoing applies when each (eigen-)value of the observable is in correspondence with not just one reading of the pointer, but with several (say, R) readings, all of the same significance for the outcome.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
244
+ page_content=' Then one simply inserts pi/R into the corresponding Lindblad term, in place of just pi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
245
+ page_content=' In the more complex case, that not all readings have the same likelihood, pi would have to be weighted by a probability factot, rather than by a constant denominator.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
246
+ page_content=' 7 The Lindbladian, ”Who ordered this?”' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
247
+ page_content=' Historically, Lindblad terms were introduced as the most general forms that maintain complete positivity of the DM’s and preserve their trace [13, 14].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
248
+ page_content=' The various derivations that have been presented (and among these a recent one by [32]), involve several approximations for the coupling between the system and its environment.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
249
+ page_content=' Insomuch that the derivation involves also tracing over the degrees of freedom of the environment, much detail of the latter is lost and of course it is impossible to work backwards from the Lindbladian to the environment.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
250
+ page_content=' What is remarkable is that for special purposes the appropriate Lindbladian operators take a very special, practically unique form.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
251
+ page_content=' Such is the case for the accepted description of thermalization [21, 24] by a Lindblad formalism.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
252
+ page_content=' The parametrization of the Lindblad term employed in the present work, though it may appear arbitrary and particular for each case, is in fact identical to the one used for thermalization subject to the relabelling of the Gibbsian thermal distribution function as (the Born) probabilities (p1, p2, .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
253
+ page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
254
+ page_content='),with the proviso of working in the observable, rather than in the energy basis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
255
+ page_content=' (This contrasts with the different approach in [15], which claims attainment of collapse for any Lindbladian operator.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
256
+ page_content=') At the same time, it needs to be noted that the analog of detailed balance is missing in wave-function collapse.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
257
+ page_content=' How come to have such a specific Lindbladian, whose source may be any measurement device and procedure?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
258
+ page_content=' One is left to wonder about the possibility of a special meta-physical status of the Lindblad terms, or query with Wheeler ”Who ordered this?”' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
259
+ page_content=' 10 8 Conclusion The well-known Lindbladian extension to the quantum theory of motion to environmental effects is here adapted to establish the resolution of a wave-packet in a measurement as a smooth process.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
260
+ page_content=' This is enabled by an unambiguous parametrization of the jump-operators describing the interaction of the broad environment with the observed system, both regarded as quantal entities.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
261
+ page_content=' Above, in section 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
262
+ page_content='1, a brief historically oriented preview has been provided for the distinct approach in this work, namely, one based on the wholeness of the entities (observed system and observing device), through the (Lindbladian) equation yielding the evolution of the system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
263
+ page_content=' A main result emerging from the formalism, and capable of experimental verification, is the finitely temporal variation of the system, and this in a de- terministic way rather than just statistically, on the average, contrasting also with the instantaneous collapse description by (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
264
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
265
+ page_content=') von Neumann.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
266
+ page_content=' Such tem- poral variation in continuous-thermalization processes has been proposed quite recently [33, 34], also by employment of a Lindbladian formalism and within a Markovian framework.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
267
+ page_content=' Experimentally, verification of the time dependence of the transition in any particular measurement, implicit in our formulae, could be observed by re- peated observation performed on the system subject to non-demolition tran- sitions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
268
+ page_content=' These observations would be akin to the Zeno-effect measurement, such as has been achieved in the form of quasi-periodic oscillation of the result for a Superconducting flux cubit[35].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
269
+ page_content=' Further work is needed for quantifying the information-entropy change in the environment [31, 36].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
270
+ page_content=' Acknowledgement The authors thank the referee for meticulous reading and insightful ques- tioning of a previous version of this paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
271
+ page_content=' References [1] A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
272
+ page_content='J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
273
+ page_content=' Leggett, Macroscopic quantum systems and the quantum theory of measurement.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
274
+ page_content=' Suppl.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
275
+ page_content=' Progress Theor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
276
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
277
+ page_content=' 69, 80 (1980) [2] G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
278
+ page_content='C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
279
+ page_content=' Ghirardi, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
280
+ page_content=' Pearle and A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
281
+ page_content=' Rimini, Markov processes in Hilbert space and continuous spontaneous localization of systems of identical particles.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
282
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
283
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
284
+ page_content=' A 42, 78 (1990) [3] E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
285
+ page_content='P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
286
+ page_content=' Wigner, Remarks on the mind-body question.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
287
+ page_content=' In I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
288
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
289
+ page_content=' Good (ed.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
290
+ page_content=') ”The Scientist Speculates”(Heinemann, London,1961) 11 [4] W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
291
+ page_content='H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
292
+ page_content=' Louisell, Quantum Statistical Properties of Radiation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
293
+ page_content=' (Wiley, New York, 1973) [5] G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
294
+ page_content='C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
295
+ page_content=' Ghirardi, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
296
+ page_content=' Rimini and T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
297
+ page_content='Weber, Unified dynamics for microscopic and macroscopic systems, Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
298
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
299
+ page_content=' D 34, 470 (1986) [6] P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
300
+ page_content=' Pearle, Combining stochastic, dynamical state reduction with sponta- neous localization, Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
301
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
302
+ page_content=' A 39 2277-89 (1989) [7] H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
303
+ page_content=' Everett, Relative state formulation of quantum mechanics, Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
304
+ page_content=' Mod.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
305
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
306
+ page_content=' 29 454-462 (1957) [8] I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
307
+ page_content=' Byalinicki-Birula and J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
308
+ page_content=' Mycielski, Nonlinear wave mechanics, Annals Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
309
+ page_content=' (NY) 108 62-83 (1976).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
310
+ page_content=' [9] G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
311
+ page_content=' Sewell, On the mathematical structure of quantum measurement theory, Repts.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
312
+ page_content=' Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
313
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
314
+ page_content=' 56 271-90 (2005) [10] L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
315
+ page_content='Di´osi, Models for universal reduction of macroscopic quantum fuctua- tions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
316
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
317
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
318
+ page_content=' A 40 165 (1989) [11] R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
319
+ page_content=' Penrose, On gravity’s role in quantum state reduction.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
320
+ page_content=' Gen.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
321
+ page_content=' Relativ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
322
+ page_content=' Gravit.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
323
+ page_content=' 28, 581 (1996).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
324
+ page_content=' [12] S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
325
+ page_content=' Donadi, K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
326
+ page_content=' Piscicchia, C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
327
+ page_content=' Curceanu, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
328
+ page_content=' Di´osi, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
329
+ page_content=' Laubenstein and A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
330
+ page_content=' Bassi, Underground test of gravity-related wave function collapse.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
331
+ page_content=' Nat.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
332
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
333
+ page_content=' 17 74 (2021) [13] G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
334
+ page_content=' Lindblad, On the generators of quantum dynamical semigroups, Com- mun.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
335
+ page_content=' Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
336
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
337
+ page_content=', 48 112-30 (1976).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
338
+ page_content=' [14] V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
339
+ page_content=' Gorini, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
340
+ page_content=' Kossakowski and E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
341
+ page_content='C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
342
+ page_content='G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
343
+ page_content=' Sudarshan, Completely positive dynamical semigroups of N-level systems, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
344
+ page_content=' Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
345
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
346
+ page_content=' 17 821-5 (1976).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
347
+ page_content=' [15] S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
348
+ page_content=' Weinberg, What happens in a measurement?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
349
+ page_content=', Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
350
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
351
+ page_content=' A 93, 032124 1-6 (2016) [16] M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
352
+ page_content='A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
353
+ page_content=' Nielsen and I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
354
+ page_content='L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
355
+ page_content=' Chuang, Quantun Computation and Quantum In- formation (Cambridge University Press, Cambridge, U.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
356
+ page_content='K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
357
+ page_content=', 2010) [17] K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
358
+ page_content=' Jacobs and D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
359
+ page_content='A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
360
+ page_content=' Steck, A straightforward introduction to continu- ous quantum measurement, Contemporary Physics, 47 (5) 279-303 (2006).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
361
+ page_content=' Arxiv.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
362
+ page_content='org/abs/quant-ph/0611067.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
363
+ page_content=' [18] N.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
364
+ page_content=' Bohr, Atomic Physics and Human Knowledge, (Wiley, New York, 1958) [19] J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
365
+ page_content=' Bell, Physics World, 3 33-4 (1990).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
366
+ page_content=' [20] A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
367
+ page_content=' Peres, Quantum Theory: Concepts and Methods, (Kluwer Academic, Norwell MA, 1998) 12 [21] H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
368
+ page_content='-P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
369
+ page_content=' Breuer and F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
370
+ page_content=' F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
371
+ page_content=' Petruccione, The Theory of Open Quantum Systems (Oxford University Press, New York, 2002) [22] S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
372
+ page_content=' B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
373
+ page_content=' J¨ager , T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
374
+ page_content=' Schmit, G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
375
+ page_content=' Morigi , M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
376
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
377
+ page_content=' Holland and R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
378
+ page_content=' Betzholz, Lindblad Master Equations for Quantum Systems Coupled to Dissipative Bosonic Modes, Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
379
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
380
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
381
+ page_content=' 129 063601 1-7 (2022) [23] D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
382
+ page_content=' Leibfried, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
383
+ page_content=' Blatt, C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
384
+ page_content=' Monroe, and D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
385
+ page_content=' Wineland, (2003).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
386
+ page_content=' Quantum dynamics of single trapped ions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
387
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
388
+ page_content=' Mod.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
389
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
390
+ page_content=' 75 (1), 281–324 (2003).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
391
+ page_content=' Doi:10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
392
+ page_content='1103/RevModPhys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
393
+ page_content='75.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
394
+ page_content='281.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
395
+ page_content=' [24] K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
396
+ page_content=' Funo, N.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
397
+ page_content=' Shiraishi and K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
398
+ page_content=' Saito, Speed limit for open quantum systems, New J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
399
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
400
+ page_content=' 21 0300611 1-9 (2019).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
401
+ page_content=' [25] J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
402
+ page_content='M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
403
+ page_content=' Horowitz and J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
404
+ page_content='M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
405
+ page_content='R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
406
+ page_content=' Parrondo, Entropy production along non- equilibrium jump trajectories, New J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
407
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
408
+ page_content=' 15 220601 1-23 (2013).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
409
+ page_content=' ) [26] L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
410
+ page_content=' I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
411
+ page_content=' Mandelshtam and I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
412
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
413
+ page_content=' Tamm, The uncertainty relation between en- ergy and time in nonrelativistic quantum mechanics.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
414
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
415
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
416
+ page_content=' (USSR) 9, 249 (1945).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
417
+ page_content=' [27] N.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
418
+ page_content=' Margulus and L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
419
+ page_content=' B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
420
+ page_content=' Levitin, The maximum speed of dynamical evolu- tion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
421
+ page_content=' Physica D 120, 188 (1998).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
422
+ page_content=' [28] S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
423
+ page_content=' L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
424
+ page_content=' Braunstein and C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
425
+ page_content=' M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
426
+ page_content=' Caves, Statistical distance and the geometry of quantum states.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
427
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
428
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
429
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
430
+ page_content=' 72 3439-44 (1994) [29] S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
431
+ page_content=' Deffner and S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
432
+ page_content=' Campbell, Quantum speed limits: from Heisenberg’s uncertainty principle to optimal quantum control.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
433
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
434
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
435
+ page_content=' A: Math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
436
+ page_content=' Theor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
437
+ page_content=' 50 453001 (2017) [30] J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
438
+ page_content=' Anandan and Y.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
439
+ page_content=' Aharonov, Geometry of quantum evolution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
440
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
441
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
442
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
443
+ page_content=' 65 1697-701 (1990) [31] S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
444
+ page_content='M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
445
+ page_content=' Barnett, Quantum Information,(University Press, Oxford, 2009), Chapter 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
446
+ page_content=' [32] D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
447
+ page_content=' Manzano, A short introduction to the Lindblad master equation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
448
+ page_content=' AIP Adv.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
449
+ page_content=' 10 (2) 1063 (2020);' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
450
+ page_content=' doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
451
+ page_content='org/10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
452
+ page_content='1063/1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
453
+ page_content='5115323 [33] K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
454
+ page_content=' Korzekwa and M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
455
+ page_content='Lostaglio, Optimizing thermalization, Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
456
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
457
+ page_content='Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
458
+ page_content=' 129 040602 1-7 (2022) [34] M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
459
+ page_content='Lostaglio and K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
460
+ page_content=' Korzekwa, Continuous thermomajorizaton and a com- plete set of Markovian thermal processes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
461
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
462
+ page_content=' Rev, A 106 012426 1-18 (2022) [35] K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
463
+ page_content=' Kakuyanagi, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
464
+ page_content=' Baba, Y.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
465
+ page_content=' Matsuzaki, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
466
+ page_content=' Nakano, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
467
+ page_content=' Saito and K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
468
+ page_content=' Semba, Observation of quantum Zeno effect in a superconducting flux qubit, New J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
469
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
470
+ page_content=' 17 063035 1-10 (2015) [36] M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
471
+ page_content=' Keyl, Fundamentals of quantum information theory, Physics Reports 369 (5) 431-58 (2002) 13' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/EdE0T4oBgHgl3EQfywKq/content/2301.02664v1.pdf'}
EdE1T4oBgHgl3EQf-gYV/content/tmp_files/2301.03568v1.pdf.txt ADDED
@@ -0,0 +1,1867 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Quasi-equilibrium configurations of binary systems of dark matter admixed neutron
2
+ stars
3
+ Hannes R. R¨uter
4
+ ,1 Violetta Sagun
5
+ ,1 Wolfgang Tichy
6
+ ,2 and Tim Dietrich
7
+ 3, 4
8
+ 1CFisUC, Department of Physics, University of Coimbra, 3004-516 Coimbra, Portugal
9
+ 2Department of Physics, Florida Atlantic University, Boca Raton, FL 33431, USA
10
+ 3Institut f¨ur Physik und Astronomie, Universit¨at Potsdam, Haus 28, Karl-Liebknecht-Str. 24/25, Potsdam, Germany
11
+ 4Max Planck Institute for Gravitational Physics (Albert Einstein Institute), Am M¨uhlenberg 1, Potsdam 14476, Germany
12
+ (Dated: January 10, 2023)
13
+ Using an adapted version of the SGRID code, we construct for the first time consistent quasi-
14
+ equilibrium configurations for a binary system consisting of two neutron stars in which each is
15
+ admixed with dark matter. The stars are modelled as a system of two non-interacting fluids min-
16
+ imally coupled to gravity. For the fluid representing baryonic matter the SLy equation of state is
17
+ used, whereas the second fluid, which corresponds to dark matter, is described using the equation
18
+ of state of a degenerate Fermi gas. We consider two different scenarios for the distribution of the
19
+ dark matter. In the first scenario the dark matter is confined to the core of the star, whereas in the
20
+ second scenario the dark matter extends beyond the surface of the baryonic matter, forming a halo
21
+ around the baryonic star. The presence of dark matter alters the star’s reaction to the companion’s
22
+ tidal forces, which we investigate in terms of the coordinate deformation and mass shedding pa-
23
+ rameters. The constructed quasi-equilibrium configurations mark the first step towards consistent
24
+ numerical-relativity simulations of dark matter admixed neutron star binaries.
25
+ I.
26
+ INTRODUCTION
27
+ In the present era of gravitational wave (GW) astron-
28
+ omy, the internal properties of compact stars can be
29
+ probed during their mergers. Using numerical-relativity
30
+ (NR) simulations of the last stages of a binary coales-
31
+ cence, it is possible to relate observational GW data to
32
+ properties of the source. While these simulations have
33
+ undergone significant improvements in the past, the im-
34
+ pact of dark matter (DM) on the binary neutron star
35
+ (NS) dynamics has not yet been investigated in detail
36
+ and is not taken into account in standard GW analyses.
37
+ In fact, considering a coalescence of compact objects to
38
+ occur in pure vacuum, could be an oversimplification that
39
+ may lead to incorrect conclusions.
40
+ Due to their high compactness, NSs can trap and ac-
41
+ cumulate DM in their interior throughout the star’s evo-
42
+ lution. DM alters the compact star’s properties, e. g., its
43
+ mass, its radius, its tidal deformability, its energy density
44
+ and speed of sound profiles [1–15]. Its effect depends on
45
+ the relative fraction of DM and on the exact equation of
46
+ state (EoS) for the DM and baryonic matter (BM). For an
47
+ extended discussion of the impact of DM on compact star
48
+ properties and its smoking gun signals, see Refs. [16–18].
49
+ While the effect of DM on isolated NSs can be probed
50
+ through electromagnetic observations, GW observations
51
+ of binary systems of DM admixed compact stars open up
52
+ a new observational window and the possibility to probe
53
+ a density and temperature range larger that of isolated
54
+ stars. To push forward our understanding of the imprint
55
+ of DM, we construct quasi-equilibrium configurations of
56
+ DM admixed NS binary system and study the impact of
57
+ DM focusing on quantities pertaining to binary system,
58
+ such as the orbital binding energy and the tidal deforma-
59
+ tions.
60
+ It is worth noting that not only NSs, but also black
61
+ holes could be embedded into DM. A step towards un-
62
+ derstanding the impact of DM on black hole mergers was
63
+ made in [19], where the behaviour of wave DM around
64
+ equal mass black hole binaries was studied in numerical
65
+ simulations. Furthermore, GW signals from binary coa-
66
+ lescences contain information of the binaries surrounding
67
+ medium [20].
68
+ The effect of DM on the inspiral and post-merger
69
+ phases of DM admixed NSs has been studied by a few
70
+ groups. A first study by Ellis et al. [21] used a simple
71
+ mechanical model, and showed that a DM core can lead
72
+ to the appearance of additional peaks in the post-merger
73
+ GW spectrum. In [22] NR simulations of equal-mass bi-
74
+ naries consisting of BM admixed with a bosonic Klein-
75
+ Gordon field were performed. For a DM mass fraction of
76
+ 10%, a redistribution of fermionic matter by the bosonic
77
+ cores was found, followed by the formation of a one-arm
78
+ spiral instability. Another approach approximating com-
79
+ pact dark component as test particles was studied in [23].
80
+ The simulations show the DM component to remain grav-
81
+ itationally bound after the merger of BM and orbit the
82
+ center of the remnant with an orbital separation of a few
83
+ km. The DM core and a host star are likely to spin at
84
+ different rotational frequencies just after the merger due
85
+ to the absence of non-gravitational interaction. Further
86
+ on, they may synchronise via the gravitational angular
87
+ momentum transfer, including tidal effects [24].
88
+ Up to our knowledge, the first two-fluid NR simulations
89
+ describing binaries of DM admixed NSs were performed
90
+ by Emma et al. [25] for a mixture of BM and mirror DM
91
+ only interacting via the gravitational field. The results
92
+ demonstrate that these systems tend to have a longer in-
93
+ spiral phase with increasing amount of DM, which could
94
+ be associated to the lower deformability of DM admixed
95
+ NSs. These simulations however, did not start from ini-
96
+ tial data satisfying the Hamiltonian and momentum con-
97
+ arXiv:2301.03568v1 [gr-qc] 9 Jan 2023
98
+
99
+ 2
100
+ straints [26–28] and the fluids did not start in an equilib-
101
+ rium configuration. Instead the initial data was approx-
102
+ imated by superimposing TOV-like solutions of isolated
103
+ DM admixed NSs. In this work we construct consistent,
104
+ constraint-solved, quasi-equilibrium conditions for a two-
105
+ fluid system of BM and DM.
106
+ One possible scenario for the formation of DM admixed
107
+ NSs is the capture of DM particles during the lifetime of
108
+ the star, from a progenitor to the equilibrated NS stages.
109
+ The core of a NS is very dense and hence the chance of
110
+ a DM particle experiencing scattering is relatively high.
111
+ In this scattering process the particle transfers its kinetic
112
+ energy to the star, becoming gravitationally bound [29–
113
+ 31].
114
+ This process is more efficient towards the Galac-
115
+ tic center, where the density of DM is many orders of
116
+ magnitude greater than in the galaxy’s arms [32, 33]. A
117
+ conservative estimate of DM capture in the most cen-
118
+ tral part of the Galaxy shows that stars accumulate up
119
+ to 0.01% of DM during the main sequence and equili-
120
+ brated NS stages combined [8]. However, also higher DM
121
+ factions inside compact stars can be achieved through
122
+ other scenarios, e.g., DM production during a supernova
123
+ explosion, accretion of DM clumps formed at the early
124
+ stage of the Universe, or initial star formation on a pre-
125
+ existing DM seed or local DM rich environments [34, 35].
126
+ If DM is symmetric, it cannot reach a high fraction due
127
+ to self-annihilation, producing an electromagnetic or neu-
128
+ trino signal [36]. The latter scenario could lead to addi-
129
+ tional heating of isolated NSs as well as post-merger rem-
130
+ nants [37, 38], modification of kinematic properties [39].
131
+ Moreover, production of light DM particles, e.g., axions,
132
+ in nucleon bremsstrahlung or in Cooper pair breaking
133
+ and formation processes in the NS interior [40–43], could
134
+ speed up the thermal evolution of a star by contributing
135
+ an additional cooling channel.
136
+ We consider DM to be either concentrated in a core or
137
+ extending beyond the surface of BM, forming a DM halo
138
+ around it. As a first step, we consider non-interacting,
139
+ fermonic DM with spin 1
140
+ 2. The single star properties of
141
+ this DM candidate have been discussed in Ref. [8]. The
142
+ baryonic component is modelled through a piecewiese-
143
+ polytropic fit [44] of the SLy EoS [45] that reproduces
144
+ nuclear matter ground state properties, fulfils heaviest
145
+ pulsars measurements [46, 47], X-ray observations by
146
+ NICER [48–52], and tidal deformability constraints from
147
+ GW170817 [53] and GW190425 [54] binary NS mergers.
148
+ The two components interact only through gravity, and
149
+ therefore do not repel each other, but overlap due to the
150
+ absence of non-gravitational interaction. This assump-
151
+ tion is in very good agreement with the observations of
152
+ the Bullet Cluster [55, 56] and direct DM searches [57],
153
+ which show that the DM-BM cross section to be many
154
+ orders of magnitude lower than the typical nuclear one,
155
+ σDM−BM ≈ 10−45 cm2 ≪ σBM ∼ 10−24 cm2.
156
+ By varying the particle mass and relative fraction of
157
+ DM, we obtain either a core configuration with a ra-
158
+ dius of the DM component less or equal to the baryonic
159
+ one, RD ≤ RB, or a halo with RD > RB [58].
160
+ For
161
+ both scenarios, we construct initial configurations em-
162
+ ploying SGRID [59, 60]. Many other codes exist for the
163
+ construction of quasi-equilibrium NS binary systems, no-
164
+ tably the spectral codes LORENE [61, 62], Spells [63],
165
+ FUKA [64, 65], Elliptica [66], and the finite difference
166
+ based code COCAL [67, 68]. Up to our knowledge, these
167
+ codes are only able to solve systems consisting of a sin-
168
+ gle fluid.
169
+ Here we construct for the first time quasi-
170
+ equilibrium binary configurations with two fluids.
171
+ The formalism and results are presented in geometric
172
+ units in which the gravitational constant G = 1 and the
173
+ speed of light c = 1. In these units, lengths are given
174
+ as multiples of the solar mass, M⊙. For the conversion
175
+ to SI units a spatial length must be multiplied by L0 =
176
+ 1476.6250 m/M⊙ and a time by T0 = 4.9254909 × 10−6
177
+ s/M⊙. Where appropriate we also use MeV to specify en-
178
+ ergy and mass of particles, as well as SI units. Through-
179
+ out the paper, Greek letter indices denote four dimen-
180
+ sional, spacetime indices, whereas Latin indices denote
181
+ three-dimensional, spatial indices.
182
+ The paper is organized as follows.
183
+ In Section II we
184
+ summarize the two-fluid formalism and DM distribu-
185
+ tion regimes. Its implementation to the SGRID code is
186
+ described in Section III. In Section IV we analyse the
187
+ convergence properties of the constructed configurations,
188
+ quantify the difference in the velocities of the two flu-
189
+ ids and investigate some physical properties of the quasi-
190
+ equilibrium configuration over a sequence of separations.
191
+ Section V summarizes the results and discusses future
192
+ perspectives.
193
+ II.
194
+ FORMALISM
195
+ We describe the matter as a system of two non-
196
+ interacting perfect fluids only indirectly coupled through
197
+ the gravitational field. This model is well justified, be-
198
+ cause the interaction between standard model BM and
199
+ DM is weak. Utilisation of the perfect fluid model for DM
200
+ is also justified, as the mean free path and the scattering
201
+ time scale of DM particles can be small compared to the
202
+ characteristic time scales of the binary. In the following,
203
+ we estimate the mean free path and scattering time in
204
+ a semi-classical approach for a degenerate Fermi gas of
205
+ particles with the mass of 170 MeV (≈ 3 × 10−28 kg).
206
+ The Fermi gas consists of non-interacting fermions, for
207
+ which a self-scattering cross section σDM formally does
208
+ not exist. Instead, we use the value of the upper limit
209
+ obtained from observations of merging galaxies, which
210
+ yield σDM/m(DM)
211
+ p
212
+ < 1.25 cm2/g, with m(DM)
213
+ p
214
+ the mass
215
+ of the DM particles [56, 69]. In this work we construct
216
+ configurations with a particle density n(DM) of 0.7 fm−3
217
+ in the center of the star. Together with the upper limit
218
+ for σDM this yields a mean free path λ = 1/(n(DM)σDM)
219
+ of 3.7 × 10−17 m, much smaller than the typical length
220
+ scale of a NS, which is on the order of 104 m. The scatter-
221
+ ing time scale can be estimated using the Fermi velocity,
222
+ which reaches values up to 0.8 c in the centre of the star.
223
+
224
+ 3
225
+ Finally, using the value of the mean free path, this yields
226
+ a scattering time of tc = λ/vDM = 1.5 × 10−25 s, much
227
+ smaller than for example the orbital period of the binary,
228
+ which in our configurations is a small as 3 × 10−4 s. At
229
+ the surface of the stars DM reaches the free streaming
230
+ limit and the perfect fluid limit breaks down, but there
231
+ the density is so small, that the impact on the gravita-
232
+ tional field is low and hence the matter in this region can
233
+ be neglected.
234
+ For non-interacting fluids, the energy-momentum ten-
235
+ sor can be split into the two individual fluid components
236
+ given by:
237
+ T (s)
238
+ µν = (e(s) + p(s))u(s)
239
+ µ u(s)
240
+ ν
241
+ + p(s)gµν ,
242
+ (1)
243
+ where e is the proper energy density, p is the pressure, uµ
244
+ is the four velocity of the fluid and the label (s) denotes
245
+ the particles species, which is either BM or DM. The
246
+ Einstein field equations are then given by
247
+ Rµν + 1
248
+ 2gµνR = 8π(T (BM)
249
+ µν
250
+ + T (DM)
251
+ µν
252
+ )
253
+ (2)
254
+ and, because the two particle species do not interact,
255
+ each fluid satisfies the equations of motion of a single
256
+ fluid. Consequently, each fluid satisfies energy momen-
257
+ tum conservation separately: ∇µT (s)
258
+ µν = 0.
259
+ For each fluid, we also define the rest mass density ρ(s)
260
+ 0 ,
261
+ which is computed from the number density n(s) by
262
+ ρ(s)
263
+ 0
264
+ = m(s)
265
+ p n(s) ,
266
+ (3)
267
+ with m(s)
268
+ p
269
+ being the mass of the particles. Furthermore,
270
+ the specific enthalpy is then given by
271
+ h(s) = e(s) + p(s)
272
+ ρ(s)
273
+ 0
274
+ .
275
+ (4)
276
+ To make the equations tractable, the spacetime metric
277
+ gµν is decomposed into a temporal and a spatial part by
278
+ introducing the spatial metric γij, the lapse α, and the
279
+ shift βi [27, 70, 71]. The line element in this 3+1 split
280
+ reads
281
+ ds2 = −α dt2 + γij (βidt + dxi)(βjdt + dxj) .
282
+ (5)
283
+ The extrinsic curvature Kij is related to the time deriva-
284
+ tive of γij, by the formula
285
+ Kij = − 1
286
+ 2α(∂tγij − Diβj − Djβi) ,
287
+ (6)
288
+ where Di denotes the covariant derivative compatible
289
+ with the spatial metric γij.
290
+ We construct the partial differential equations govern-
291
+ ing quasi-equilibrium by following the derivation in [72],
292
+ which is trivially applied to a system of non-interacting
293
+ fluids. To generate quasi-equilibrium configurations, we
294
+ solve equations for velocity potentials φ(s), which are de-
295
+ fined through the following split of the four-velocity
296
+ γi
297
+ µu(s)µ =
298
+ 1
299
+ h(s) (Diφ(s) + w(s)i) ,
300
+ (7)
301
+ where w(s)i is a divergence free vector, i.e., Diw(s)i = 0,
302
+ describing the rotational part of the fluid. Following the
303
+ derivation of [72], we fix the time derivatives of the fields
304
+ by assuming the existence of an approximate Killing vec-
305
+ tor ξ and a set of quasi-equilibrium conditions for the two
306
+ fluids
307
+ Lξe(s) ≈ 0 ,
308
+ (8)
309
+ Lξp(s) ≈ 0 ,
310
+ (9)
311
+ γi
312
+ µLξ(∇µφ(s)) ≈ 0 ,
313
+ (10)
314
+ γi
315
+ µL
316
+ ∇φ(s)
317
+ h(s)u(s)0 w(s)
318
+ µ
319
+ ≈ 0 .
320
+ (11)
321
+ We omit further details of the derivation, since for non-
322
+ interacting fluids everything can be directly carried over
323
+ to the individual fluid components, and we state only
324
+ the resulting partial differential equation for the velocity
325
+ potentials φ(s):
326
+ Di
327
+
328
+ ρ(s)
329
+ 0 α
330
+ h(s) (Diφ(s) + w(s)i) − ρ(s)
331
+ 0 αu(s)0(βi + ξi)
332
+
333
+ = 0 ,
334
+ (12)
335
+ where the boost factor u(s)0 is given by
336
+ u(s)0 =
337
+
338
+ h(s)2 + (Diφ(s) + w(s)
339
+ i )(Diφ(s) + w(s)i)
340
+ αh(s)
341
+ ,
342
+ (13)
343
+ and the specific enthalpy is given by the expression
344
+ h(s) =
345
+
346
+ L(s)2 − (Diφ(s) + w(s)
347
+ i )(Diφ(s) + w(s)i) , (14)
348
+ with
349
+ L(s)2 =
350
+ b(s) +
351
+
352
+ b(s)2 − 4α4((Diφ(s) + w(s)
353
+ i )w(s)i)2
354
+ 2α2
355
+ (15)
356
+ and
357
+ b(s) = ((ξi+βi)Diφ(s)−C(s))2+2α2(Diφ(s)+w(s)
358
+ i )w(s)i .
359
+ (16)
360
+ The variable C(s) is a constant, which can vary for each
361
+ star and controls the mass of the fluid component.
362
+ For the approximate Killing vector ξi we make the fol-
363
+ lowing ansatz:
364
+ ξi = Ω(−y, x − xCM, 0) + vr
365
+ D (ri − ri
366
+ CM) ,
367
+ (17)
368
+ where Ω is the instantaneous orbital frequency, D is the
369
+ separation between the star centres, vr is the radial ve-
370
+ locity, and xCM is the x-coordinate of the centre of mass.
371
+
372
+ 4
373
+ At apsis the orbital frequency together with the sepa-
374
+ ration of the stars control the orbital parameters like ec-
375
+ centricity and length of the semi-major axis. Away from
376
+ apsis there is a non-vanishing radial component of the
377
+ velocity to be taken into account. In cases like the “circu-
378
+ lar” inspiral there is no apsis, but there is an always non-
379
+ vanishing radially inward directed velocity component.
380
+ The configurations presented in this work are constructed
381
+ within the quasi-circular approximation for which the ra-
382
+ dial component is neglected, vr = 0.
383
+ We set the value of Ω to its value at second
384
+ Post-Newtonian order in Arnowitt-Deser-Misner (ADM)
385
+ gauge [73–75] using the sum of the rest masses of the
386
+ two fluids as the mass estimates of the stars, which are
387
+ computed by
388
+ m(s)
389
+ 0i =
390
+
391
+ Vi
392
+ ρ(s)
393
+ i u(s)0α
394
+
395
+ det(γjk)d3x ,
396
+ (18)
397
+ where Vi is the spatial volume over which the i-th star
398
+ extends. The value of xCM is then given by
399
+ xCM = (m(BM)
400
+ 01
401
+ + m(DM)
402
+ 01
403
+ )xc1 + (m(BM)
404
+ 02
405
+ + m(DM)
406
+ 02
407
+ )xc2
408
+ m(BM)
409
+ 01
410
+ + m(DM)
411
+ 01
412
+ + m(BM)
413
+ 02
414
+ + m(DM)
415
+ 02
416
+ ,
417
+ (19)
418
+ where xc1/2 are the x-coordinates of the centres of the
419
+ stars.
420
+ In this work, we present results for equal-mass
421
+ configurations only, i. e., xCM = 0.
422
+ Besides the continuity equation (Eq. (12)) governing
423
+ the fluid velocity potentials φ(s), the metric must be fixed
424
+ in a way satisfying the ADM constraints. To this end we
425
+ choose a conformally flat ansatz for the spatial metric,
426
+ i. e., γij = ψ4¯γij, with γij = δij and ∂tγij = 0, and con-
427
+ struct the data on maximally sliced hypersurfaces, i. e.,
428
+ the trace of the extrinsic curvature vanishes: K = 0 and
429
+ ∂tK = 0.
430
+ The free metric components are the lapse,
431
+ shift, and conformal factor ψ and their governing equa-
432
+ tions are formulated in terms of the extended conformal
433
+ thin sandwich equations (XCTS) [27, 28]. Together with
434
+ Eq. (12), the data is constrained by a set of seven coupled
435
+ partial differential equations, which are solved iteratively
436
+ one-by-one in a self-consistent manner.
437
+ III.
438
+ SGRID
439
+ We have adapted the pseudo-spectral SGRID code [59,
440
+ 60] to generate quasi-equilibrium configurations for two
441
+ fluid systems.
442
+ We use the same iteration scheme that
443
+ is used in [60] for single-fluid NSs. We sketch the iter-
444
+ ation scheme in the following with an emphasis on the
445
+ adaptions and changes made.
446
+ 1. To ensure the convergence of the solver, it is nec-
447
+ essary to provide an initial guess sufficiently close
448
+ to the true solution. This initial guess is chosen as
449
+ a superposition of two boosted TOV-like two fluid
450
+ stars of a given mass. To generate solutions with
451
+ particular rest masses for the fluid components,
452
+ one has to find the central pressures for which the
453
+ masses are realized. Since we are dealing with two
454
+ fluids, this is a two-dimensional root finding prob-
455
+ lem. In our tests, we found that using the Newton-
456
+ Raphson method is not always reliable, because the
457
+ masses are not a monotonous function of the central
458
+ pressures, hence, a Newton-Raphson solver easily
459
+ gets caught in a local extremum of the mass func-
460
+ tion. Instead, we employ a series of bisections on
461
+ the central pressure of one fluid component while
462
+ keeping the central pressure of the other fluid fixed.
463
+ The series of bisections iterates between the two
464
+ fluid components in a self-consistent manner until
465
+ the fluid masses are sufficiently close to the target
466
+ parameters.
467
+ 2. If the residuals of Eq. (12) are larger than 10%
468
+ of the combined residuals of the XCTS equations,
469
+ we solve Eq. (12) and set the new φ(s) to be the
470
+ average of the old solution φ(s)
471
+ old and the just ob-
472
+ tained solution φ(s)
473
+ ell , using the following weights
474
+ φ(s) = 0.8φ(s)
475
+ old + 0.2φ(s)
476
+ ell .
477
+ 3. We proceed by solving the XCTS equations and
478
+ update α, β, and ψ in the same way, averaging the
479
+ old and new solution.
480
+ 4. We do not adjust the values of Ω and xCM as in [60].
481
+ The value of Ω would be fixed within an eccentric-
482
+ ity reduction scheme. xCM is left at its Newtonian
483
+ value, Eq. (19).
484
+ 5. We adjust the constants C(s), such that the rest
485
+ masses of each component and in each star match
486
+ our desired target masses. We then update the val-
487
+ ues of h(s) keeping it fixed until the end of the next
488
+ iteration.
489
+ 6. If the sum of the residuals is below a certain toler-
490
+ ance or a prescribed maximum number of iterations
491
+ is reached, the iteration ends here and is concluded
492
+ with a final solving of the XCTS equations.
493
+ 7. The system of partial differential equations does
494
+ not fix the position of the stars and, hence, they will
495
+ slowly drift if not kept under control. To keep the
496
+ stars in place, the center of the stars are driven back
497
+ to the desired position. For single fluids, the center
498
+ is usually defined in an unambiguous way as the
499
+ point of maximum density. For two fluids the defi-
500
+ nition is ambiguous, because the tidal deformations
501
+ due to the companion star are different for each
502
+ fluid component and, consequently, the maximum
503
+ densities are at different points. In most cases, how-
504
+ ever, the two maximum points will still be close.
505
+ The results shown in this work are obtained by
506
+ choosing the point with the maximum of the to-
507
+ tal proper energy density, e(tot) = e(BM) + e(DM),
508
+ as the center of the stars. We have chosen e(tot),
509
+
510
+ 5
511
+ 1
512
+ 1.05 1.1 1.15 1.2 1.25 1.3
513
+ 1
514
+ 1.05
515
+ 1.1
516
+ 1.15
517
+ 1.2
518
+ -25
519
+ -20
520
+ -15
521
+ -10
522
+ -5
523
+ 0
524
+ 5
525
+ 10
526
+ 15
527
+ 20
528
+ 25
529
+ X
530
+ -10
531
+ -8
532
+ -6
533
+ -4
534
+ -2
535
+ 0
536
+ 2
537
+ 4
538
+ 6
539
+ 8
540
+ 10
541
+ Y
542
+ -25
543
+ -20
544
+ -15
545
+ -10
546
+ -5
547
+ 0
548
+ 5
549
+ 10
550
+ 15
551
+ 20
552
+ 25
553
+ X
554
+ -10
555
+ -8
556
+ -6
557
+ -4
558
+ -2
559
+ 0
560
+ 2
561
+ 4
562
+ 6
563
+ 8
564
+ 10
565
+ Y
566
+ -25
567
+ -20
568
+ -15
569
+ -10
570
+ -5
571
+ 0
572
+ 5
573
+ 10
574
+ 15
575
+ 20
576
+ 25
577
+ X
578
+ -10
579
+ -8
580
+ -6
581
+ -4
582
+ -2
583
+ 0
584
+ 2
585
+ 4
586
+ 6
587
+ 8
588
+ 10
589
+ Y
590
+ -25
591
+ -20
592
+ -15
593
+ -10
594
+ -5
595
+ 0
596
+ 5
597
+ 10
598
+ 15
599
+ 20
600
+ 25
601
+ X
602
+ -10
603
+ -8
604
+ -6
605
+ -4
606
+ -2
607
+ 0
608
+ 2
609
+ 4
610
+ 6
611
+ 8
612
+ 10
613
+ Y
614
+ FIG. 1.
615
+ Specific enthalpy in the z = 0 plane for a config-
616
+ uration with DM halo. In the upper halves only the specific
617
+ enthalpy of DM is shown, whereas in the lower halves the
618
+ BM component lies on top of it. The black lines indicate the
619
+ boundaries of the spectral elements. Each NS is comprised of
620
+ a central cubical element and six cubed sphere elements (of
621
+ which only four intersect the z = 0 plane). The separation
622
+ between the NS centres amounts to 32 M⊙ (47.3 km).
623
+ in particular, because it is a covariant scalar and it
624
+ is the major quantity determining the gravitational
625
+ potential, hence giving an estimate for the center of
626
+ mass of the star. To drive the center of mass back,
627
+ the values of h(s) are transformed by
628
+ h(s),new = h(s) + ∆ri∂ih(s) ,
629
+ (20)
630
+ where ∆ri = ri
631
+ current − ri
632
+ desired.
633
+ 8. Continue with step 2.
634
+ The SGRID code uses surface-fitted coordinates to re-
635
+ duce the Runge phenomenon at the surface of the star.
636
+ Each time we update the specific enthalpy h(s) (step 5
637
+ in the iteration), we adapt the grid such that the bound-
638
+ aries of spectral elements coincide with the new surface
639
+ of the outer fluid. That means we only construct con-
640
+ figurations in which the surfaces of the two fluids do not
641
+ intersect, which would in principle be possible given the
642
+ different deformabilities of the fluids. Furthermore, we do
643
+ not construct domains that are adapted to the surface of
644
+ the inner fluid.
645
+ Therefore, at the surface of the inner
646
+ fluid one can expect to observe the Runge phenomenon
647
+ and a slight degradation of the convergence in the trun-
648
+ cation error. Fig. 1 shows a visualisation of the deformed
649
+ spectral elements inside the NS and the distribution of
650
+ matter in terms of the specific enthalpy.
651
+ To close the system, the EoS is required to relate e(s),
652
+ p(s), ρ(s)
653
+ 0 , and h(s). For the EoS, SGRID reads in either
654
+ parameters of piecewise polytropes or EoS tables. EoS
655
+ tables are interpolated in a thermodynamically consis-
656
+ tent manner [76] using a cubic Hermite interpolation. To
657
+ find the thermodynamic quantities for a given specific
658
+ enthalpy a Newton-Raphson root finder is used. At low
659
+ densities we use a polytrope that is matched at the lowest
660
+ density of the table.
661
+ IV.
662
+ RESULTS
663
+ A.
664
+ Parameters of Constructed Configurations
665
+ We consider two different two-fluid configurations, one
666
+ in which DM is confined to the core of the NS, the dark
667
+ core configuration, and one in which DM extends beyond
668
+ the surface of the BM, so that the NS has a DM halo,
669
+ which we will refer to as the dark halo configuration.
670
+ Furthermore we compare to configurations consisting of
671
+ BM only: the single fluid configuration.
672
+ We describe BM by a piecewise-polytropic fit [44] to
673
+ the SLy EoS [45]. As a model of DM, we investigate the
674
+ degenerate, relativistic Fermi gas of spin- 1
675
+ 2 particles at
676
+ zero temperature, for which the EoS is read in as tabu-
677
+ lated data. EoSs at zero temperature are sufficient for
678
+ our calculations, because the Fermi energy of the sys-
679
+ tem is much higher than its temperature. The typical
680
+ temperature T0 of NS cores is of the order of 106 − 108
681
+ K [77, 78]. We assume that DM has the same tempera-
682
+ ture as the BM, because the captured DM particles keep
683
+ scattering with baryons, rarely but often enough to ther-
684
+ malise with the BM component. A core temperature of
685
+ approximately 108 K is much lower than the Fermi en-
686
+ ergy of BM. This is also true for the Fermi gas EoS we
687
+ consider, e. g. in the dark halo case the Fermi energy of
688
+ DM reaches 403 MeV in the center of the star, an energy
689
+ smaller than that of the BM, but still much larger than
690
+ the temperature of the star, kBT0 �� 0.009 MeV.
691
+ For the dark core configuration the DM particles have a
692
+ mass of 1000 MeV and DM provides 5% of the NSs’ total
693
+ rest mass.
694
+ Fermionic DM particles with mass of 1000
695
+ MeV present an interesting case, because they resemble
696
+ nucleons.
697
+ In the dark halo case we model DM by particles with
698
+ a mass of 170 MeV, for which the fluid is less dense and
699
+ hence easily forms a halo. Furthermore in the dark halo
700
+ configuration DM only contributes 0.5% of the total rest
701
+ mass. The choice of these values for the particle masses
702
+ is motivated by the results of [8], where it was shown that
703
+ for the DM particle masses below 174 MeV DM admixed
704
+ NSs are in agreement with astrophysical observations of
705
+ the heaviest NSs for arbitrary relative fraction of DM.
706
+ Moreover, the chosen mass of 170 MeV and the fraction
707
+ of 0.5% leads to a relatively small halo of approximately
708
+ twice the radius of the BM component, which is easy to
709
+ model. When the size of the halos is big enough so that
710
+ they touch each other, it is no longer possible to fit the
711
+ element surfaces to the outer fluid of a star. Hence, we
712
+ are discarding configurations with separations at which
713
+ the two halos merge. In all configurations the individual
714
+ NSs have the same total rest mass, i. e., the combined
715
+ rest mass of BM and DM is 1.4M⊙. In all setups, the
716
+ NSs have equal masses and are irrotational, i. e., they
717
+ have zero spins.
718
+
719
+ 6
720
+ B.
721
+ Convergence
722
+ To validate the code, we check the convergence of the
723
+ Hamiltonian constraint for a dark halo configuration of
724
+ NSs with a separation of 44 M⊙ (65.0 km) on a quasi-
725
+ circular orbit.
726
+ Fig. 2 shows the magnitude of the Hamiltonian con-
727
+ straint H on the z = 0 plane. The constraint violations
728
+ are largest in the interior of the star, where they reach
729
+ values up to 4×10−5, whereas in the vacuum regions the
730
+ error drops to values below 10−9, but with some spikes
731
+ on the order of 10−7 at the element boundaries. A be-
732
+ haviour commonly seen for spectral codes. The Hamil-
733
+ tonian constraint is largest in the region where the inner
734
+ fluid is non-vanishing. In Fig. 2 one can observe a clear
735
+ transition on the surface of the baryonic fluid to lower
736
+ constraint violations in the DM halo.
737
+ Fig. 3 demonstrates the development of the volume-
738
+ normalised L2-norm of the Hamiltonian constraint for the
739
+ inner cube of one of the stars during the iterative solv-
740
+ ing process. The figure shows the behaviour for different
741
+ number of points n in each dimension, which is the same
742
+ for each spectral element. All curves show a saturation
743
+ in the norm of the Hamiltonian constraint towards the
744
+ end of the iteration process, which for all configurations
745
+ is stopped after 40 iterations. Furthermore, it is visible
746
+ that higher resolution leads to smaller violations of the
747
+ Hamiltonian constraint in the final solution. For compar-
748
+ ison Fig. 3 also shows the sequence for a corresponding
749
+ single fluid configuration with the same mass and sepa-
750
+ ration. After 40 iterations its Hamiltonian constraint is a
751
+ factor 10 smaller than the dark halo configurations and it
752
+ does not show any signs of saturation, i. e. it would prob-
753
+ ably reach even smaller constraint violations if iterated
754
+ further.
755
+ The convergence in the final solution is further inves-
756
+ tigated in Fig. 4, which shows its L2-norm of the Hamil-
757
+ tonian constraint with respect to the number of colloca-
758
+ tion points in the spectral elements. The figure shows
759
+ the constraint violation for the inner cube element and
760
+ for the cubed sphere facing towards the companion star,
761
+ which is also representative for all other cubed sphere el-
762
+ ements inside the NSs. The curves are almost straight
763
+ lines on the log-log-plot of Fig. 4, which is compatible
764
+ with a polynomial convergence of the constraints, i. e.,
765
+ |H|L2 ∼ n−p, with p the order of convergence.
766
+ This
767
+ is the expected convergence behaviour for non-smooth
768
+ data, which we have due to the surface of the inner fluid.
769
+ Using the highest and lowest resolution we can estimate
770
+ the order of convergence in the inner cube element to be
771
+ p ≈ log22/10(|H|L2,n=10/|H|L2,n=22) ≈ 2.7.
772
+ To investigate the convergence of the actual solution
773
+ variables we interpolate the data from different resolu-
774
+ tions on a common set of points and compute norms
775
+ of the estimated errors on these points.
776
+ We interpo-
777
+ late the solution onto a 10 × 10 × 10-grid equidistant
778
+ in each direction, with coordinate components given by
779
+ ri ∈ {20m/9, m ∈ [0..9]}. This grid includes some points
780
+ 1e-13
781
+ 1e-12
782
+ 1e-11
783
+ 1e-10
784
+ 1e-9
785
+ 1e-8
786
+ 1e-7
787
+ 1e-6
788
+ 1e-5
789
+ 1.00
790
+ 1.33
791
+ 1.05 1.1 1.15 1.2 1.25
792
+ 1.00
793
+ 1.22
794
+ 1.05
795
+ 1.1
796
+ 1.15
797
+ FIG. 2.
798
+ Hamiltonian constraint in a dark halo configuration
799
+ in the z = 0 plane. In the lower half the specific enthalpy of
800
+ the two fluids is overlaid.
801
+ with pure vacuum, points with only one fluid present
802
+ and points with both fluids present.
803
+ The error in the
804
+ solution is estimated by taking the difference to the so-
805
+ lution with the highest resolution, i. e., the solution that
806
+ has 22 points in each dimension of the spectral elements.
807
+ In Fig. 5 we show the convergence of the 1-norm and
808
+ the maximum norm over the set of interpolated points
809
+ for the gxx component of the metric and the lapse α.
810
+ Both quantities do not show a monotonic decay of the er-
811
+ ror, but there is an overall trend of decaying error. This
812
+ somewhat broken convergence behaviour can again be
813
+ attributed to the presence of non-smooth fields on the
814
+ surface of the inner fluid. Fig. 6 shows the convergence
815
+ of the error in the specific enthalpy.
816
+ The DM in this
817
+ configuration is fitted to the element boundaries and its
818
+ specific enthalpy displays a relatively clear convergence
819
+ behaviour.
820
+ The BM fluid on the other hand shows a
821
+ very broken convergence and only very little improve-
822
+ ment from the lowest to the highest number of points.
823
+ The maximum norm of the error is actually growing for
824
+ the two largest number of points, whereas the 1-norm
825
+ of the error is also slightly broken, but with an overall
826
+ behaviour similar to that of gxx and α.
827
+ It should be noted, that it is not clear whether the
828
+ formalism of Sec. II actually possesses a unique solution.
829
+ The partial differential equation (12) is not strictly ellip-
830
+ tic on the fluid surface and hence the standard theorems
831
+ for the uniqueness of the solution can not be applied. In-
832
+ stead our algorithm might find a solution of many possi-
833
+ ble, which is another possible explanation for the slightly
834
+ broken convergence behaviour.
835
+ C.
836
+ Difference in the Fluid Velocities
837
+ It is worth pointing out that even if the BM and
838
+ DM fluid components are both irrotational, i. e., non-
839
+ spinning, the exact velocity profiles are not the same.
840
+ The reason for this does not lie in the notion of an irro-
841
+ tational fluid, but is caused by differences in the fluids’
842
+ equations of motion. An irrotational fluid is defined by
843
+
844
+ 7
845
+ 0
846
+ 5
847
+ 10
848
+ 15
849
+ 20
850
+ 25
851
+ 30
852
+ 35
853
+ 40
854
+ iteration
855
+ 10
856
+ -5
857
+ 10
858
+ -4
859
+ 10
860
+ -3
861
+ 10
862
+ -2
863
+ |H|
864
+ L
865
+ 2
866
+ /V
867
+ element
868
+ n
869
+ =
870
+ 22, single fluid
871
+ n
872
+ =
873
+ 12, dark halo
874
+ n
875
+ =
876
+ 14, dark halo
877
+ n
878
+ =
879
+ 18, dark halo
880
+ n
881
+ =
882
+ 22, dark halo
883
+ FIG. 3.
884
+ L2-norm over the inner cube in one of the stars,
885
+ normalised by the volume of the inner cube. The different
886
+ lines show configurations with different number of points n in
887
+ each dimension.
888
+ 10
889
+ 12
890
+ 14
891
+ 16
892
+ 18
893
+ 20
894
+ 22
895
+ numer of points in each dimension n
896
+ 10
897
+ -4
898
+ |H|
899
+ L
900
+ 2
901
+ /V
902
+ element
903
+ inner cube
904
+ left cubed sphere
905
+ FIG. 4.
906
+ Normalised L2-norm of the Hamiltonian constraint
907
+ in a dark halo configuration for a different number of points
908
+ per dimension. The norm is normalised by the volume of the
909
+ spectral element. Note that the x-axis and y-axis are scaled
910
+ logarithmically.
911
+ 10
912
+ 12
913
+ 14
914
+ 16
915
+ 18
916
+ 20
917
+ numer of points in each dimension n
918
+ 10
919
+ -4
920
+ 10
921
+ -3
922
+ 10
923
+ -2
924
+ 10
925
+ -1
926
+ 10
927
+ 0
928
+ 10
929
+ 1
930
+ error norm of variable Y, ||Y
931
+ n
932
+
933
+ Y
934
+ 22
935
+ ||
936
+ g
937
+ xx
938
+ , 1-norm
939
+ g
940
+ xx
941
+ , maximum norm
942
+ α, 1-norm
943
+ α, maximum norm
944
+ FIG. 5.
945
+ Self-convergence of metric variables in dark halo
946
+ configurations. Black: error norm of the gxx component of
947
+ the metric. Blue: error norm of the lapse, α. We not that the
948
+ 1-norm is not normalised by the number of points.
949
+ 10
950
+ 12
951
+ 14
952
+ 16
953
+ 18
954
+ 20
955
+ numer of points in each dimension n
956
+ 10
957
+ -4
958
+ 10
959
+ -3
960
+ 10
961
+ -2
962
+ 10
963
+ -1
964
+ 10
965
+ 0
966
+ ||h
967
+ (s)
968
+ n
969
+
970
+ h
971
+ (s)
972
+ 22
973
+ ||
974
+ h
975
+ BM
976
+ , 1-norm
977
+ h
978
+ BM
979
+ , maximum norm
980
+ h
981
+ DM
982
+ , 1-norm
983
+ h
984
+ DM
985
+ , maximum norm
986
+ FIG. 6.
987
+ Self-convergence of the specific enthalpy in dark halo
988
+ configurations.
989
+ Black: error norm of the baryonic specific
990
+ enthalpy h(BM), which is the inner fluid. Blue: error norm of
991
+ the specific enthalpy of DM, h(DM). We not that the 1-norm
992
+ is not normalised by the number of points.
993
+ the vanishing of its kinematic vorticity tensor [79]
994
+ ωαβ := P µ
995
+ α P ν
996
+ β ∇[µuν] = 0 ,
997
+ (21)
998
+ with P µ
999
+ α = δµ
1000
+ α + uµuα.
1001
+ This notion does not depend
1002
+ on the thermodynamic properties of the fluid and hence
1003
+ differences in the velocities can only be the result of the of
1004
+ the equations of motion, that are used in the derivation of
1005
+ the formlalism in Sec. II, i. e. the Euler equations [72, 80]
1006
+ u(s)µ∇µ(h(s)u(s)
1007
+ ν
1008
+ + ∇νh(s)) = 0 ,
1009
+ (22)
1010
+ which follow from ∇µT (s)
1011
+ µν = 0, and the continuity equa-
1012
+ tion
1013
+ ∇µ(ρ(s)
1014
+ 0 u(s)µ) = 0 .
1015
+ (23)
1016
+ If for example the DM would have the same four-velocity
1017
+ as the BM, it would still be irrotational, but might be
1018
+ incompatible with the laws of energy-momentum or par-
1019
+ ticle number conservation.
1020
+ In nature the disparity in the fluid velocities is affected
1021
+ by two counter-acting effects, particle scattering between
1022
+ BM and DM on the one hand and physics determining
1023
+ spin-down on the other hand.
1024
+ In our formulation the
1025
+ two fluids are modelled as non-interacting, but the BM-
1026
+ DM scattering cross-section might be non-zero in nature,
1027
+ which would drive the two fluids towards a common ve-
1028
+ locity. This process is counter-acted by effects driving the
1029
+ fluid into an irrotational state, as for example magnetic
1030
+ braking for BM [81–83]. It is unclear whether a similar
1031
+ effect exists for DM and whether it is dominant over the
1032
+ effect of BM-DM scattering. By assuming vanishing of
1033
+ the kinematic vorticity for the DM component, we as-
1034
+ sume that such an effect exists and it is also dominating
1035
+ over the scattering with BM.
1036
+ We find that both fluids move with basically the same
1037
+ velocity, with coinciding velocities in the star center,
1038
+
1039
+ 8
1040
+ 12
1041
+ 14
1042
+ 16
1043
+ 18
1044
+ 20
1045
+ x
1046
+ −0.15
1047
+ −0.10
1048
+ −0.05
1049
+ 0.00
1050
+ 0.05
1051
+ 0.10
1052
+ relative difference, V
1053
+ (s)x
1054
+ 1
1055
+
1056
+ V
1057
+ (DM)x
1058
+ /V
1059
+ (BM)x
1060
+ , dark halo
1061
+ 1
1062
+
1063
+ V
1064
+ (DM)x
1065
+ /V
1066
+ (BM)x
1067
+ , dark core
1068
+ V
1069
+ (BM)x
1070
+ , dark halo
1071
+ V
1072
+ (DM)x
1073
+ , dark core
1074
+ FIG. 7.
1075
+ Relative difference in the velocities for configurations
1076
+ with a separation of 32 M⊙. The difference is shown along a
1077
+ diagonal with the parametrization ri(s) = s(1, 1, 0)+ri
1078
+ c, going
1079
+ through the center of the star located at ri
1080
+ c = (16M⊙, 0, 0).
1081
+ V (BM)x (black, dash-dotted line) and V (DM)x (grey, dotted
1082
+ line) show the x-component of the velocity of the respective
1083
+ inner fluid.
1084
+ but increasing difference towards the surface of the in-
1085
+ ner fluid. We quantify this effect in terms of the residual
1086
+ three-velocity V (s)i, in which the orbital movement given
1087
+ by the Killing vector ξµ is split off,
1088
+ V (s)i = u(s)i/u(s)0 − ξi .
1089
+ (24)
1090
+ Fig. 7 shows the x-component of V (s)i and the relative
1091
+ difference of the fluid velocities for the region in which
1092
+ both fluids are present. We present results for configu-
1093
+ rations at a separation of 32 M⊙, a separation at which
1094
+ the DM halos in the dark halo configurations are already
1095
+ relatively close and deformed (Fig. 1). We find that dif-
1096
+ ferences in the two fluids are smaller for larger separation,
1097
+ which is intuitively understandable, because for large sep-
1098
+ arations the system goes to the limit of isolated NSs in
1099
+ which the fluid velocities coincide.
1100
+ The data in Fig. 7 is shown along a diagonal through
1101
+ the star parametrized in the following way:
1102
+ ri(s) =
1103
+ s(1, 1, 0)+ri
1104
+ c, where ri
1105
+ c is the center of the star. We choose
1106
+ to present the data along this diagonal because the differ-
1107
+ ence V (BM)i − V (DM)i has a quadrupolar structure with
1108
+ nodes going through ri
1109
+ c and being approximately parallel
1110
+ to the x and y axes. Hence the difference is basically zero
1111
+ on the x and y-axis, but very prominent along the spec-
1112
+ ified diagonal. The relative difference between the resid-
1113
+ ual velocities is below 0.2% near the center of the star
1114
+ and reaches up to 10% on the surfaces of the inner fluids.
1115
+ The difference between the velocities of the dark halo and
1116
+ dark core configurations is relatively small, which can be
1117
+ seen from the fact the curves of the velocities of the inner
1118
+ fluids lie on top of each other.
1119
+ D.
1120
+ Binding Energy
1121
+ NSs with a DM component are more tightly bound,
1122
+ because the DM component adds gravitating mass, but
1123
+ provides no additional repulsion to balance the gravita-
1124
+ tional pressure [8]. The gravitational binding energy of
1125
+ the particles is the difference of the ADM mass [27, 84, 85]
1126
+ and the sum of the rest masses m(s)
1127
+ 0i of the components.
1128
+ If all fluid particles would fall in from infinity, the true
1129
+ ADM mass would equal the total rest mass. However, the
1130
+ configurations that we construct do not contain GWs and
1131
+ therefore they do not model the energy lost in gravita-
1132
+ tional radiation. The difference in our ADM mass esti-
1133
+ mate and the total rest mass is, therefore, a measure of
1134
+ the particle binding energy:
1135
+ Ebind,p = MADM − m(BM)
1136
+ 01
1137
+ − m(DM)
1138
+ 01
1139
+ − m(BM)
1140
+ 02
1141
+ − m(DM)
1142
+ 02
1143
+ .
1144
+ (25)
1145
+ Fig. 8 shows the particle binding energy as a function of
1146
+ our estimate for the ADM angular momentum JADM. It
1147
+ can be seen that dark core configurations are more tightly
1148
+ bound than single fluid configurations. The dark halo
1149
+ configurations seemingly coincide with the single fluid
1150
+ case. This can be attributed to the relatively low DM
1151
+ fraction of only 0.5% in these configurations. All config-
1152
+ urations are more tightly bound for smaller JADM cor-
1153
+ responding to smaller stellar separations. This is due to
1154
+ the stronger orbital binding between the two stars.
1155
+ Most of the binding energy is contained in the indi-
1156
+ vidual stars and the contribution of the orbital binding
1157
+ energy is universal in all configurations. The orbital bind-
1158
+ ing energy Ebind,orb is the energy necessary for the two
1159
+ NSs to escape to infinity. It can be computed using the
1160
+ gravitational mass m(s)
1161
+ i
1162
+ of the components, by
1163
+ Ebind,orb = MADM −m(BM)
1164
+ 1
1165
+ −m(DM)
1166
+ 1
1167
+ −m(BM)
1168
+ 2
1169
+ −m(DM)
1170
+ 2
1171
+ .
1172
+ (26)
1173
+ The gravitational masses m(s)
1174
+ i
1175
+ are obtained by solving a
1176
+ TOV-like equation for isolated stars that have the same
1177
+ rest masses. The gravitational mass m(s)
1178
+ i
1179
+ is smaller than
1180
+ the rest mass m(s)
1181
+ i0 , because it accounts for the binding
1182
+ energy. Hence, Ebind,orb contains only contributions of
1183
+ the binding energy that are due to the mutual binding
1184
+ between the stars. Fig. 9 shows that the orbital binding
1185
+ energy is mostly independent of the DM configuration.
1186
+ The biggest effect is seen for the dark core configurations
1187
+ for which the magnitude of Ebind,orb is about 2% smaller
1188
+ than that of the other configurations.
1189
+ E.
1190
+ Deformation
1191
+ To quantify the deformation of the stars we compute
1192
+ the ratio of the diameters along the orbital radius and
1193
+ along the polar axes. The diameter along the orbital ra-
1194
+ dius is taken as ∆x, largest difference in the x-coordinates
1195
+ of two points on the fluid surface. The polar diameter
1196
+
1197
+ 9
1198
+ 6.00
1199
+ 6.25
1200
+ 6.50
1201
+ 6.75
1202
+ 7.00
1203
+ 7.25
1204
+ 7.50
1205
+ 7.75
1206
+ J
1207
+ ADM
1208
+ /M
1209
+ 2
1210
+
1211
+ −0.300
1212
+ −0.295
1213
+ −0.290
1214
+ −0.285
1215
+ −0.280
1216
+ −0.275
1217
+ −0.270
1218
+ particle binding energy E
1219
+ bind,
1220
+ p
1221
+ /M
1222
+
1223
+ single fluid
1224
+ dark halo
1225
+ dark core
1226
+ FIG. 8.
1227
+ Particle binding energy Ebind,p as a function of the
1228
+ ADM angular momentum.
1229
+ 6.00
1230
+ 6.25
1231
+ 6.50
1232
+ 6.75
1233
+ 7.00
1234
+ 7.25
1235
+ 7.50
1236
+ 7.75
1237
+ J
1238
+ ADM
1239
+ /M
1240
+ 2
1241
+
1242
+ −0.0300
1243
+ −0.0275
1244
+ −0.0250
1245
+ −0.0225
1246
+ −0.0200
1247
+ −0.0175
1248
+ −0.0150
1249
+ orbital binding energy E
1250
+ bind,
1251
+ orb
1252
+ /M
1253
+
1254
+ single fluid
1255
+ dark halo
1256
+ dark core
1257
+ FIG. 9.
1258
+ Orbital binding energy Ebind,orb as a function of the
1259
+ ADM angular momentum.
1260
+ ∆z, is the largest difference in the z-coordinate of two
1261
+ points on the fluid surface. The tidal force of the com-
1262
+ panion stretches the star in x-direction, whereas the poles
1263
+ are slightly flattened. This measure of deformation is of
1264
+ course coordinate-dependent, but it still provides some
1265
+ physical insights. Fig. 10 shows the deformation ∆x/∆z
1266
+ for each fluid surface. When the NSs are closer, the tidal
1267
+ forces on the companion are stronger and hence the de-
1268
+ formation is stronger. It can be observed that NSs with
1269
+ a DM core are systematically less deformed than their
1270
+ one-fluid counterparts.
1271
+ The strong deformation in the dark halo case can also
1272
+ be seen in Fig. 1, which shows a cut through the z = 0
1273
+ plane.
1274
+ For a separation of 32 M⊙ (47.3 km) the de-
1275
+ formation is clearly visible by eye. At a separation of
1276
+ 28 M⊙ (41.3 km) the deformation becomes already so
1277
+ strong that the surfaces of the NSs touch and mass shed-
1278
+ ding occurs.
1279
+ The closeness to mass shedding can be quantified in
1280
+ terms of the mass-shedding parameter χ, which was first
1281
+ introduced in [86] and which we define as
1282
+ χ(s) =
1283
+ ∂xh(s)|eq
1284
+ ∂zh(s)|pole,avg
1285
+ ,
1286
+ (27)
1287
+ 25
1288
+ 30
1289
+ 35
1290
+ 40
1291
+ 45
1292
+ 50
1293
+ 55
1294
+ 60
1295
+ 65
1296
+ separation D
1297
+ [km]
1298
+ 20
1299
+ 25
1300
+ 30
1301
+ 35
1302
+ 40
1303
+ 45
1304
+ separation D/M
1305
+
1306
+ 1.000
1307
+ 1.025
1308
+ 1.050
1309
+ 1.075
1310
+ 1.100
1311
+ 1.125
1312
+ 1.150
1313
+ deformation ratio ∆x/∆z
1314
+ BM, single fluid
1315
+ BM, dark halo
1316
+ BM, dark core
1317
+ DM, dark halo
1318
+ DM, dark core
1319
+ FIG. 10.
1320
+ Deformation ∆x/∆z of the fluid surfaces as func-
1321
+ tion of the NS centres. The deformation is computed as the
1322
+ ratio of the largest extents in x and z direction. Curves la-
1323
+ beled BM show the deformation of the surface of the baryonic
1324
+ fluid, whereas curves labeled DM show the deformation of the
1325
+ DM surface.
1326
+ where the label ”eq” denotes the point on the surface,
1327
+ which is facing towards the other companion and for
1328
+ which the x-coordinate is extremal.
1329
+ The label ”pole”
1330
+ denotes the surface points at which the z-coordinate is
1331
+ extremal and where in Eq. (27) the label ”avg” indicates
1332
+ that we have averaged over the values at the ”north and
1333
+ south pole”. Note that for non-spinning stars the ”north”
1334
+ and ”south pole” values only differ slightly due to round-
1335
+ off error. In the mass shedding limit χ(s) will tend to
1336
+ 0. We evaluate the χ(s) for each fluid component indi-
1337
+ vidually on the respective fluid surfaces. We show the
1338
+ resulting χ(s) as a function of the distance of the centres
1339
+ of the stars in Fig. 11. The DM fluid in the dark halo
1340
+ scenario is easily deformable, which leads to a relatively
1341
+ small mass shedding parameter of 0.9 already at a sep-
1342
+ aration of 44 M⊙. We find that a separation of 28 M⊙
1343
+ leads to a configuration with touching star surfaces, from
1344
+ which we conclude that mass shedding occurs somewhere
1345
+ at a separation between 28 and 29 M⊙, which means the
1346
+ system will transition relatively slowly to the mass shed-
1347
+ ding regime over a time where the two NSs decrease their
1348
+ separation by 16 M⊙. For the dark core configurations,
1349
+ on the other hand, the transition to mass shedding is
1350
+ rather sudden with χ reaching a value of 0.9 at sepa-
1351
+ ration of approximately 23 M⊙ and the mass shedding
1352
+ occurring for the baryonic fluid at a separation of 16 M⊙.
1353
+ V.
1354
+ CONCLUSION
1355
+ We have extended the SGRID code to construct
1356
+ constraint-solved, quasi-equlibrium configurations of bi-
1357
+ naries of NSs consisting of two non-interacting fluids.
1358
+ The second fluid represents DM that can comprise some
1359
+ part of the matter of NS. In this study we have used the
1360
+
1361
+ 10
1362
+ 25
1363
+ 30
1364
+ 35
1365
+ 40
1366
+ 45
1367
+ 50
1368
+ 55
1369
+ 60
1370
+ 65
1371
+ separation D
1372
+ [km]
1373
+ 20
1374
+ 25
1375
+ 30
1376
+ 35
1377
+ 40
1378
+ 45
1379
+ separation D/M
1380
+
1381
+ 0.4
1382
+ 0.6
1383
+ 0.8
1384
+ 1.0
1385
+ mass shedding parameter χ
1386
+ BM, single fluid
1387
+ BM, dark halo
1388
+ BM, dark core
1389
+ DM, dark halo
1390
+ DM, dark core
1391
+ FIG. 11.
1392
+ Mass shedding parameter χ as a function of the sep-
1393
+ aration of the NS. Curves labeled BM show the deformation
1394
+ of the surface of the baryonic fluid, whereas curves labeled
1395
+ DM show the deformation of the DM surface.
1396
+ EoS of a degenerate, relativistic Fermi gas with different
1397
+ particle masses to model the DM fluid.
1398
+ These quasi-
1399
+ equlibrium configurations can be used as initial data for
1400
+ NR inspiral simulations of DM admixed NS binaries. The
1401
+ BAM code can already evolve mirror DM [25] and could
1402
+ be easily extended to allow for general EoS for the DM
1403
+ fluid.
1404
+ Another possible application of the two fluid approach
1405
+ are superfluid NS cores. At sufficiently high density BM
1406
+ forms a state made of superfluid neutrons and supercon-
1407
+ ducting protons, which can be described in a two fluid
1408
+ approach.
1409
+ However, the two fluids still interact with
1410
+ each other due to the entrainment effect and the con-
1411
+ dition of beta-equilibrium [87]. Solutions of isolated NS
1412
+ with superfluid cores are constructed in [88, 89] taking
1413
+ into account the interaction of the fluids. For a study of
1414
+ superfluid and superconducting cores in binary NS the
1415
+ formalism in this work could be extended using a similar
1416
+ model for the interactions. In binary NS collisions the
1417
+ temperature will rise above the critical temperature for
1418
+ superfluidity and superconductivity, so that it becomes
1419
+ necessary to include even a third fluid representing the
1420
+ non-superfluid component.
1421
+ We have tested the convergence of the constructed con-
1422
+ figurations with respect to resolution. The Hamiltonian
1423
+ constraint converges polynomially with an order of ≈ 2.7.
1424
+ The lack of exponential convergence can be attributed to
1425
+ the presence of the non-smooth transition of the density
1426
+ at the surface of the inner fluid, which is not fitted to
1427
+ the boundaries of the spectral elements. Self-convergence
1428
+ tests for metric components and the specific enthalpies
1429
+ show that the solution improves with increasing reso-
1430
+ lution, but with a slightly broken convergence towards
1431
+ higher resolution, which we again attribute to the sur-
1432
+ face of the inner fluid. For future improvements to the
1433
+ code it is a worthwhile consideration to implement a new
1434
+ grid layout that allows fitting to the surface of a second
1435
+ fluid
1436
+ We have shown that the two fluids do not have the ex-
1437
+ act same velocities, but that the difference in the resid-
1438
+ ual velocities reaches up to 10% on the surface of the
1439
+ inner fluids. The difference in the velocity profiles will
1440
+ be even stronger if one assumes independent rotational
1441
+ states for the components. In this work we only inves-
1442
+ tigated only purely irrotational configurations, but our
1443
+ formalism, in principle, allows for to construct configu-
1444
+ rations with arbitrary spin for the individual stars and
1445
+ fluid components. This is relevant in particular for the
1446
+ DM component, which might only have insufficient mech-
1447
+ anisms to lose angular momentum and hence could be in
1448
+ a state of rapid rotation.
1449
+ The presence of DM affects the compactness and de-
1450
+ formability of NSs, which will change the merger dynam-
1451
+ ics. We have shown that the presence of DM can delay
1452
+ the point of mass-shedding to a later stage of the inspi-
1453
+ ral, i. e., towards closer separations. This is in accordance
1454
+ with the findings in numerical evolutions of two-fluid bi-
1455
+ nary mergers [25]. In the case of a DM halo, mass shed-
1456
+ ding could occur much earlier than for the baryonic com-
1457
+ ponent. However the matter contained in the DM halo
1458
+ is rather low and hence the impact of DM mass shedding
1459
+ on the dynamics of the BM is potentially small, never-
1460
+ theless, dynamical simulations are needed to verify this
1461
+ assumption.
1462
+ ACKNOWLEDGMENTS
1463
+ This work was supported by funding from the FCT
1464
+ – Funda¸c˜ao para a Ciˆencia e a Tecnologia, I.P., within
1465
+ the Project No.
1466
+ EXPL/FIS-AST/0735/2021.
1467
+ H.R.R.
1468
+ and V.S. also acknowledge the support from the project
1469
+ No. UIDB/04564/2020, and UIDP/04564/2020. W.T.
1470
+ acknowledges funding from the National Science Foun-
1471
+ dation under grant PHY-2136036.
1472
+ [1] G. Narain, J. Schaffner-Bielich, and I. N. Mishustin,
1473
+ Compact stars made of fermionic dark matter, Phys. Rev.
1474
+ D 74, 063003 (2006), arXiv:astro-ph/0605724 [astro-ph].
1475
+ [2] P. Ciarcelluti and F. Sandin, Have neutron stars a
1476
+ dark matter core?, Phys. Lett. B 695, 19 (2011),
1477
+ arXiv:1005.0857 [astro-ph.HE].
1478
+ [3] K. M. Zurek, Asymmetric Dark Matter: Theories, Sig-
1479
+ natures, and Constraints, Phys. Rept. 537, 91 (2014),
1480
+ arXiv:1308.0338 [hep-ph].
1481
+ [4] I. Goldman, R. N. Mohapatra, S. Nussinov, D. Rosen-
1482
+ baum, and V. Teplitz, Possible Implications of Asym-
1483
+ metric Fermionic Dark Matter for Neutron Stars, Phys.
1484
+
1485
+ 11
1486
+ Lett. B 725, 200 (2013), arXiv:1305.6908 [astro-ph.CO].
1487
+ [5] M. I. Gresham, H. K. Lou, and K. M. Zurek, Nuclear
1488
+ Structure of Bound States of Asymmetric Dark Matter,
1489
+ Phys. Rev. D 96, 096012 (2017), arXiv:1707.02313 [hep-
1490
+ ph].
1491
+ [6] J. Ellis, G. H¨utsi, K. Kannike, L. Marzola, M. Raidal, and
1492
+ V. Vaskonen, Dark matter effects on neutron star prop-
1493
+ erties, Phys. Rev. D 97, 123007 (2018), arXiv:1804.01418
1494
+ [astro-ph.CO].
1495
+ [7] A. E. Nelson, S. Reddy, and D. Zhou, Dark halos around
1496
+ neutron stars and gravitational waves, J. Cosm. Astrop.
1497
+ Phys. 2019, 012 (2019), arXiv:1803.03266 [hep-ph].
1498
+ [8] O. Ivanytskyi, V. Sagun, and I. Lopes, Neutron stars:
1499
+ New constraints on asymmetric dark matter, Phys. Rev.
1500
+ D 102, 063028 (2020).
1501
+ [9] H. C. Das, A. Kumar, B. Kumar, S. K. Biswal, T. Nakat-
1502
+ sukasa, A. Li, and S. K. Patra, Effects of dark matter on
1503
+ the nuclear and neutron star matter, Mon. Not. R. As-
1504
+ tron. Soc. 495, 4893 (2020), arXiv:2002.00594 [nucl-th].
1505
+ [10] A. Das, T. Malik, and A. C. Nayak, Dark matter ad-
1506
+ mixed neutron star properties in the light of gravitational
1507
+ wave observations: a two fluid approach, arxiv
1508
+ (2020),
1509
+ arXiv:2011.01318 [nucl-th].
1510
+ [11] V. Sagun,
1511
+ E. Giangrandi,
1512
+ O. Ivanytskyi,
1513
+ I. Lopes,
1514
+ and K. Bugaev, Constraints on the fermionic dark
1515
+ matter from observations of neutron stars, Proc. Sc.
1516
+ PANIC2021,
1517
+ 313
1518
+ (2022),
1519
+ arXiv:2111.13289
1520
+ [astro-
1521
+ ph.HE].
1522
+ [12] K.-L. Leung, M.-c. Chu, and L.-M. Lin, Tidal deforma-
1523
+ bility of dark matter admixed neutron stars, Phys. Rev.
1524
+ D 105, 123010 (2022).
1525
+ [13] D. R. Karkevandi, S. Shakeri, V. Sagun, and O. Ivanyt-
1526
+ skyi, Bosonic dark matter in neutron stars and its ef-
1527
+ fect on gravitational wave signal, Physical Review D 105,
1528
+ 10.1103/physrevd.105.023001 (2022).
1529
+ [14] V. Sagun, E. Giangrandi, O. Ivanytskyi, C. Providˆencia,
1530
+ and T. Dietrich, How does dark matter affect com-
1531
+ pact star properties and high density constraints of
1532
+ strongly interacting matter, in 15th Conference on
1533
+ Quark Confinement and the Hadron Spectrum (2022)
1534
+ arXiv:2211.10510 [astro-ph.HE].
1535
+ [15] F. Di Giovanni, N. Sanchis-Gual, P. Cerd´a-Dur´an, and
1536
+ J. A. Font, Can fermion-boson stars reconcile multimes-
1537
+ senger observations of compact stars?, Phys. Rev. D 105,
1538
+ 063005 (2022), arXiv:2110.11997 [gr-qc].
1539
+ [16] A. Maselli, P. Pnigouras, N. G. Nielsen, C. Kouvaris, and
1540
+ K. D. Kokkotas, Dark stars: gravitational and electro-
1541
+ magnetic observables, Phys. Rev. D 96, 023005 (2017),
1542
+ arXiv:1704.07286 [astro-ph.HE].
1543
+ [17] E. Giangrandi, V. Sagun, O. Ivanytskyi, C. Providˆencia,
1544
+ and T. Dietrich, The effects of self-interacting bosonic
1545
+ dark
1546
+ matter
1547
+ on
1548
+ neutron
1549
+ star
1550
+ properties
1551
+ (2022),
1552
+ arXiv:2209.10905 [astro-ph.HE].
1553
+ [18] N. Rutherford, G. Raaijmakers, C. Prescod-Weinstein,
1554
+ and A. Watts, Constraining bosonic asymmetric dark
1555
+ matter with neutron star mass-radius measurements,
1556
+ arXiv (2022), arXiv:2208.03282 [astro-ph.HE].
1557
+ [19] J. Bamber, J. C. Aurrekoetxea, K. Clough, and P. G. Fer-
1558
+ reira, Black hole merger simulations in wave dark matter
1559
+ environments (2022), arXiv:2210.09254 [gr-qc].
1560
+ [20] P. S. Cole, G. Bertone, A. Coogan, D. Gaggero, T. Kary-
1561
+ das, B. J. Kavanagh, T. F. M. Spieksma, and G. M.
1562
+ Tomaselli, Disks, spikes, and clouds: distinguishing en-
1563
+ vironmental effects on BBH gravitational waveforms
1564
+ (2022), arXiv:2211.01362 [gr-qc].
1565
+ [21] J. Ellis, A. Hektor, G. H¨utsi, K. Kannike, L. Marzola,
1566
+ M. Raidal, and V. Vaskonen, Search for Dark Matter Ef-
1567
+ fects on Gravitational Signals from Neutron Star Merg-
1568
+ ers, Phys. Lett. B 781, 607 (2018), arXiv:1710.05540
1569
+ [astro-ph.CO].
1570
+ [22] M. Bezares, D. Vigan`o, and C. Palenzuela, Gravitational
1571
+ wave signatures of dark matter cores in binary neutron
1572
+ star mergers by using numerical simulations, Phys. Rev.
1573
+ D 100, 044049 (2019), arXiv:1905.08551 [gr-qc].
1574
+ [23] A. Bauswein, G. Guo, J.-H. Lien, Y.-H. Lin, and M.-
1575
+ R. Wu, Compact Dark Objects in Neutron Star Mergers
1576
+ (2020), arXiv:2012.11908 [astro-ph.HE].
1577
+ [24] M.
1578
+ Hippert,
1579
+ E.
1580
+ Dillingham,
1581
+ H.
1582
+ Tan,
1583
+ D.
1584
+ Curtin,
1585
+ J. Noronha-Hostler, and N. Yunes, Dark Matter or Reg-
1586
+ ular Matter in Neutron Stars?
1587
+ How to tell the differ-
1588
+ ence from the coalescence of compact objects (2022),
1589
+ arXiv:2211.08590 [astro-ph.HE].
1590
+ [25] M. Emma,
1591
+ F. Schianchi,
1592
+ F. Pannarale,
1593
+ V. Sagun,
1594
+ and T. Dietrich, Numerical Simulations of Dark Mat-
1595
+ ter Admixed Neutron Star Binaries, arXiv
1596
+ (2022),
1597
+ arXiv:2206.10887 [gr-qc].
1598
+ [26] E. Gourgoulhon, 3+1 Formalism in General Relativity
1599
+ (Springer, Berlin, 2012) gr-qc/0703035.
1600
+ [27] T. W. Baumgarte and S. L. Shapiro, Numerical Relativ-
1601
+ ity: Solving Einstein’s Equations on the Computer (Cam-
1602
+ bridge University Press, Cambridge, 2010).
1603
+ [28] W. Tichy, The initial value problem as it relates to nu-
1604
+ merical relativity, Reports on Progress in Physics 80,
1605
+ 026901 (2017).
1606
+ [29] C. Kouvaris, The Dark Side of Neutron Stars, Adv.
1607
+ High Energy Phys. 2013, 856196 (2013), arXiv:1308.3222
1608
+ [astro-ph.HE].
1609
+ [30] N. F. Bell, G. Busoni, S. Robles, and M. Virgato, Im-
1610
+ proved Treatment of Dark Matter Capture in Neutron
1611
+ Stars, JCAP 09, 028, arXiv:2004.14888 [hep-ph].
1612
+ [31] N. F. Bell, G. Busoni, T. F. Motta, S. Robles, A. W.
1613
+ Thomas, and M. Virgato, Nucleon Structure and Strong
1614
+ Interactions in Dark Matter Capture in Neutron Stars,
1615
+ Phys. Rev. Lett. 127, 111803 (2021), arXiv:2012.08918
1616
+ [hep-ph].
1617
+ [32] C. Kouvaris and P. Tinyakov, Can neutron stars con-
1618
+ strain dark matter?, Phys. Rev. D 82, 063531 (2010),
1619
+ arXiv:1004.0586 [astro-ph.GA].
1620
+ [33] A. Del Popolo, M. Le Delliou, and M. Deliyergiyev, Neu-
1621
+ tron Stars and Dark Matter, Universe 6, 222 (2020).
1622
+ [34] G. Bertone and M. Fairbairn, Compact Stars as Dark
1623
+ Matter
1624
+ Probes,
1625
+ Phys.
1626
+ Rev.
1627
+ D
1628
+ 77,
1629
+ 043515
1630
+ (2008),
1631
+ arXiv:0709.1485 [astro-ph].
1632
+ [35] M. Baryakhtar et al., Dark Matter In Extreme As-
1633
+ trophysical Environments, in 2022 Snowmass Summer
1634
+ Study (2022) arXiv:2203.07984 [hep-ph].
1635
+ [36] C. Kouvaris, Wimp annihilation and cooling of neutron
1636
+ stars, Phys. Rev. D 77, 023006 (2008), arXiv:0708.2362
1637
+ [astro-ph].
1638
+ [37] A. de Lavallaz and M. Fairbairn, Neutron stars as
1639
+ dark matter probes, Phys. Rev. D 81, 123521 (2010),
1640
+ arXiv:1004.0629 [astro-ph.GA].
1641
+ [38] K. Hamaguchi, N. Nagata, and K. Yanagi, Dark Matter
1642
+ Heating vs. Rotochemical Heating in Old Neutron Stars,
1643
+ Phys. Lett. B 795, 484 (2019), arXiv:1905.02991 [hep-
1644
+ ph].
1645
+ [39] M. ´A. P´erez-Garc´ıa and J. Silk, Dark matter seeding and
1646
+
1647
+ 12
1648
+ the kinematics and rotation of neutron stars, Phys. Lett.
1649
+ B 711, 6 (2012), arXiv:1111.2275 [astro-ph.CO].
1650
+ [40] A. Sedrakian, Axion cooling of neutron stars, Phys. Rev.
1651
+ D 93, 065044 (2016).
1652
+ [41] A. Sedrakian, Axion cooling of neutron stars. II. Be-
1653
+ yond hadronic axions, Phys. Rev. D 99, 043011 (2019),
1654
+ arXiv:1810.00190 [astro-ph.HE].
1655
+ [42] T. Dietrich and K. Clough, Cooling binary neutron
1656
+ star remnants via nucleon-nucleon-axion bremsstrahlung,
1657
+ Phys. Rev. D 100, 083005 (2019), arXiv:1909.01278 [gr-
1658
+ qc].
1659
+ [43] M. Buschmann, C. Dessert, J. W. Foster, A. J. Long,
1660
+ and B. R. Safdi, Upper Limit on the QCD Axion Mass
1661
+ from Isolated Neutron Star Cooling, Phys. Rev. Lett.
1662
+ 128, 091102 (2022), arXiv:2111.09892 [hep-ph].
1663
+ [44] J. S. Read, B. D. Lackey, B. J. Owen, and J. L. Fried-
1664
+ man, Constraints on a phenomenologically parameterized
1665
+ neutron-star equation of state, Phys. Rev. D 79, 124032
1666
+ (2009), arXiv:0812.2163 [astro-ph].
1667
+ [45] F. Douchin and P. Haensel, A unified equation of state
1668
+ of dense matter and neutron star structure, Astron. As-
1669
+ trophys. 380, 151 (2001), arXiv:astro-ph/0111092.
1670
+ [46] J. Antoniadis, P. C. C. Freire, N. Wex, T. M. Tauris, R. S.
1671
+ Lynch, M. H. van Kerkwijk, M. Kramer, C. Bassa, V. S.
1672
+ Dhillon, T. Driebe, J. W. T. Hessels, V. M. Kaspi, V. I.
1673
+ Kondratiev, N. Langer, T. R. Marsh, M. A. McLaugh-
1674
+ lin, T. T. Pennucci, S. M. Ransom, I. H. Stairs, J. van
1675
+ Leeuwen, J. P. W. Verbiest, and D. G. Whelan, A Massive
1676
+ Pulsar in a Compact Relativistic Binary, Science 340,
1677
+ 448 (2013), arXiv:1304.6875 [astro-ph.HE].
1678
+ [47] E. Fonseca et al., Refined Mass and Geometric Measure-
1679
+ ments of the High-mass PSR J0740+6620, Astrophys. J.
1680
+ Lett. 915, L12 (2021), arXiv:2104.00880 [astro-ph.HE].
1681
+ [48] M. C. Miller, F. K. Lamb, A. J. Dittmann, S. Bog-
1682
+ danov, Z. Arzoumanian, K. C. Gendreau, S. Guillot,
1683
+ A. K. Harding, W. C. G. Ho, J. M. Lattimer, R. M. Lud-
1684
+ lam, S. Mahmoodifar, S. M. Morsink, P. S. Ray, T. E.
1685
+ Strohmayer, K. S. Wood, T. Enoto, R. Foster, T. Oka-
1686
+ jima, G. Prigozhin, and Y. Soong, Psr j0030+0451 mass
1687
+ and radius from nicer data and implications for the prop-
1688
+ erties of neutron star matter, The Astrophysical Journal
1689
+ 887, L24 (2019).
1690
+ [49] T. E. Riley et al., A NICER View of PSR J0030+0451:
1691
+ Millisecond Pulsar Parameter Estimation, Astrophys. J.
1692
+ Lett. 887, L21 (2019), arXiv:1912.05702 [astro-ph.HE].
1693
+ [50] G. Raaijmakers et al., Constraining the dense matter
1694
+ equation of state with joint analysis of NICER and
1695
+ LIGO/Virgo measurements, Astrophys. J. Lett. 893, L21
1696
+ (2020), arXiv:1912.11031 [astro-ph.HE].
1697
+ [51] M. C. Miller et al., The Radius of PSR J0740+6620 from
1698
+ NICER and XMM-Newton Data, Astrophys. J. Lett.
1699
+ 918, L28 (2021), arXiv:2105.06979 [astro-ph.HE].
1700
+ [52] T. E. Riley et al., A NICER View of the Massive Pul-
1701
+ sar PSR J0740+6620 Informed by Radio Timing and
1702
+ XMM-Newton Spectroscopy, Astrophys. J. Lett. 918,
1703
+ L27 (2021), arXiv:2105.06980 [astro-ph.HE].
1704
+ [53] B. P. Abbott et al. (LIGO Scientific, Virgo), Properties
1705
+ of the binary neutron star merger GW170817, Phys. Rev.
1706
+ X 9, 011001 (2019), arXiv:1805.11579 [gr-qc].
1707
+ [54] B. P. Abbott et al. (LIGO Scientific, Virgo), GW190425:
1708
+ Observation of a Compact Binary Coalescence with To-
1709
+ tal Mass ∼ 3.4M⊙, Astrophys. J. Lett. 892, L3 (2020),
1710
+ arXiv:2001.01761 [astro-ph.HE].
1711
+ [55] D. Clowe, M. Bradaˇc, A. H. Gonzalez, M. Markevitch,
1712
+ S. W. Randall, C. Jones, and D. Zaritsky, A direct em-
1713
+ pirical proof of the existence of dark matter, The Astro-
1714
+ physical Journal 648, L109 (2006).
1715
+ [56] S. W. Randall, M. Markevitch, D. Clowe, A. H. Gonza-
1716
+ lez, and M. Bradac, Constraints on the Self-Interaction
1717
+ Cross-Section of Dark Matter from Numerical Simula-
1718
+ tions of the Merging Galaxy Cluster 1E 0657-56, Astro-
1719
+ phys. J. 679, 1173 (2008), arXiv:0704.0261 [astro-ph].
1720
+ [57] J.
1721
+ Billard
1722
+ et
1723
+ al.,
1724
+ Direct
1725
+ detection
1726
+ of
1727
+ dark
1728
+ mat-
1729
+ ter—APPEC committee report*, Rept. Prog. Phys. 85,
1730
+ 056201 (2022), arXiv:2104.07634 [hep-ex].
1731
+ [58] V. Sagun, E. Giangrandi, O. Ivanytskyi, I. Lopes, and
1732
+ K. A. Bugaev, Constraints on the fermionic dark matter
1733
+ from observations of neutron stars, PoS PANIC2021,
1734
+ 313 (2022), arXiv:2111.13289 [astro-ph.HE].
1735
+ [59] W. Tichy, Constructing quasi-equilibrium initial data for
1736
+ binary neutron stars with arbitrary spins, Phys. Rev.
1737
+ D86, 064024 (2012), arXiv:1209.5336 [gr-qc].
1738
+ [60] W.
1739
+ Tichy,
1740
+ A.
1741
+ Rashti,
1742
+ T.
1743
+ Dietrich,
1744
+ R.
1745
+ Dudi,
1746
+ and
1747
+ B. Br¨ugmann, Constructing binary neutron star ini-
1748
+ tial data with high spins, high compactnesses, and
1749
+ high mass ratios, Physical Review D 100, 10.1103/phys-
1750
+ revd.100.124046 (2019), arXiv:1910.09690 [gr-qc].
1751
+ [61] E. Gourgoulhon, P. Grandcl´ement, J.-A. Marck, J. No-
1752
+ vak, and K. Taniguchi, http://www.lorene.obspm.fr.
1753
+ [62] E.
1754
+ Gourgoulhon,
1755
+ P.
1756
+ Grandcl´ement,
1757
+ K.
1758
+ Taniguchi,
1759
+ J. Marck, and S. Bonazzola, Quasiequilibrium sequences
1760
+ of synchronized and irrotational binary neutron stars in
1761
+ general relativity: Method and tests, Phys. Rev. D 63,
1762
+ 064029 (2001).
1763
+ [63] N. Tacik, F. Foucart, H. P. Pfeiffer, R. Haas, S. Ossokine,
1764
+ J. Kaplan, C. Muhlberger, M. D. Duez, L. E. Kidder,
1765
+ M. A. Scheel, and B. Szil´agyi, Binary Neutron Stars with
1766
+ Arbitrary Spins in Numerical Relativity, Phys. Rev. D92,
1767
+ 124012 (2015), arXiv:1508.06986 [gr-qc].
1768
+ [64] P. Grandcl´ement, S. Auliac, H. Roussille, and L. J. Pa-
1769
+ penfort, https://kadath.obspm.fr/fuka/.
1770
+ [65] L. J. Papenfort, S. D. Tootle, P. Grandcl´ement, E. R.
1771
+ Most, and L. Rezzolla, New public code for initial data
1772
+ of unequal-mass, spinning compact-object binaries, Phys.
1773
+ Rev. D 104, 024057 (2021), arXiv:2103.09911 [gr-qc].
1774
+ [66] A. Rashti, F. M. Fabbri, B. Br¨ugmann, S. V. Chaurasia,
1775
+ T. Dietrich, M. Ujevic, and W. Tichy, New pseudospec-
1776
+ tral code for the construction of initial data, Phys. Rev.
1777
+ D 105, 104027 (2022), arXiv:2109.14511 [gr-qc].
1778
+ [67] K. Ury¯u and A. Tsokaros, New code for equilibriums
1779
+ and quasiequilibrium initial data of compact objects,
1780
+ Phys.Rev. D85, 064014 (2012), arXiv:1108.3065 [gr-qc].
1781
+ [68] K. Ury¯u, S. Yoshida, E. Gourgoulhon, C. Markakis,
1782
+ K. Fujisawa, A. Tsokaros, K. Taniguchi, and Y. Eriguchi,
1783
+ New code for equilibriums and quasiequilibrium initial
1784
+ data of compact objects. iv. rotating relativistic stars
1785
+ with mixed poloidal and toroidal magnetic fields, Phys.
1786
+ Rev. D 100, 123019 (2019), arXiv:1906.10393 [gr-qc].
1787
+ [69] D. Wittman, N. Golovich, and W. A. Dawson, The Mis-
1788
+ measure of Mergers: Revised Limits on Self-interacting
1789
+ Dark Matter in Merging Galaxy Clusters, Astrophys. J.
1790
+ 869, 104 (2018), arXiv:1701.05877 [astro-ph.CO].
1791
+ [70] G. B. Cook, Initial data for numerical relativity, Liv-
1792
+ ing Reviews in Relativity 3, 10.12942/lrr-2000-5 (2000),
1793
+ arXiv:gr-qc/0007085 [gr-qc].
1794
+ [71] W. Tichy, The initial value problem as it relates to nu-
1795
+ merical relativity, Rept. Prog. Phys. 80, 026901 (2017),
1796
+ arXiv:1610.03805 [gr-qc].
1797
+
1798
+ 13
1799
+ [72] W. Tichy, Initial data for binary neutron stars with
1800
+ arbitrary
1801
+ spins,
1802
+ Phys.
1803
+ Rev.
1804
+ D
1805
+ 84,
1806
+ 024041
1807
+ (2011),
1808
+ arXiv:1107.1440 [gr-qc].
1809
+ [73] G. Sch¨afer and N. Wex, Phys. Lett. A 174, 196 (1993).
1810
+ [74] W. Tichy, B. Br¨ugmann, M. Campanelli, and P. Diener,
1811
+ Binary black hole initial data for numerical general rela-
1812
+ tivity based on post-Newtonian data, Phys. Rev. D 67,
1813
+ 064008 (2003), gr-qc/0207011.
1814
+ [75] W. Tichy and B. Br¨ugmann, Quasi-equilibrium binary
1815
+ black hole sequences for puncture data derived from he-
1816
+ lical killing vector conditions, Phys. Rev. D 69, 024006
1817
+ (2004), gr-qc/0307027.
1818
+ [76] F. Swesty, Thermodynamically consistent interpolation
1819
+ for equation of state tables, Journal of Computational
1820
+ Physics 127, 118 (1996).
1821
+ [77] A. Cumming, E. F. Brown, F. J. Fattoyev, C. J.
1822
+ Horowitz, D. Page, and S. Reddy, Lower limit on the
1823
+ heat capacity of the neutron star core, Phys. Rev. C 95,
1824
+ 025806 (2017), arXiv:1608.07532 [astro-ph.HE].
1825
+ [78] D. Page, U. Geppert, and F. Weber, The Cooling of com-
1826
+ pact stars, Nucl. Phys. A 777, 497 (2006), arXiv:astro-
1827
+ ph/0508056.
1828
+ [79] S. W. Hawking and G. F. R. Ellis, The Large Scale Struc-
1829
+ ture of Space-Time (Cambridge University Press, Cam-
1830
+ bridge (UK), 1973).
1831
+ [80] M. Shibata, A relativistic formalism for computation
1832
+ of irrotational binary stars in quasi equilibrium states,
1833
+ Phys. Rev. D 58, 024012 (1998), arXiv:gr-qc/9803085.
1834
+ [81] J. P. Ostriker and J. E. Gunn, On the nature of pulsars.
1835
+ i. theory., Astrophys. J., 157:
1836
+ 1395-1417(Sept. 1969).
1837
+ 10.1086/150160 (1969).
1838
+ [82] M. Colpi, A. Possenti, S. Popov, and F. Pizzolato, Spin
1839
+ and magnetism in old neutron stars, in Physics of Neu-
1840
+ tron Star Interiors, edited by D. Blaschke, A. Sedrakian,
1841
+ and N. K. Glendenning (Springer Berlin Heidelberg,
1842
+ Berlin, Heidelberg, 2001) pp. 440–466, astro-ph/0012394.
1843
+ [83] A. Rogers and S. Safi-Harb, On the diversity of com-
1844
+ pact objects within supernova remnants – II. energy-loss
1845
+ mechanisms, Monthly Notices of the Royal Astronomical
1846
+ Society 465, 383 (2016), arXiv:1610.04685 [astro-ph.HE].
1847
+ [84] R. Arnowitt, S. Deser, and C. W. Misner, The dynamics
1848
+ of general relativity, in Gravitation: An Introduction to
1849
+ Current Research, edited by L. Witten (Wiley, New York,
1850
+ 1962) pp. 227–265.
1851
+ [85] E. Gourgoulhon, 3+1 formalism and bases of numerical
1852
+ relativity (2007), arXiv:gr-qc/0703035.
1853
+ [86] E. Gourgoulhon, P. Grandcl´ement, K. Taniguchi, J.-A.
1854
+ Marck, and S. Bonazzola, Quasiequilibrium sequences
1855
+ of synchronized and irrotational binary neutron stars in
1856
+ general relativity: Method and tests, Phys. Rev. D 63,
1857
+ 064029 (2001), gr-qc/0007028.
1858
+ [87] G. L. Comer and R. Joynt, Relativistic mean field model
1859
+ for entrainment in general relativistic superfluid neutron
1860
+ stars, Phys. Rev. D 68, 023002 (2003), gr-qc/0212083.
1861
+ [88] R. Prix, J. Novak, and G. L. Comer, Relativistic numeri-
1862
+ cal models for stationary superfluid neutron stars, Phys.
1863
+ Rev. D 71, 043005 (2005), gr-qc/0410023.
1864
+ [89] N. Chamel, Two-fluid models of superfluid neutron star
1865
+ cores, Monthly Notices of the Royal Astronomical Society
1866
+ 388, 737 (2008), arXiv:0805.1007 [astro-ph].
1867
+
EdE1T4oBgHgl3EQf-gYV/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
F9E0T4oBgHgl3EQfRAC9/vector_store/index.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8543255a1afc371582fee66fa20be6a3a0ac7c99e8ccb693d3d6546e1b90a3f2
3
+ size 200007
FtAyT4oBgHgl3EQf5Poe/vector_store/index.faiss ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:828c7a99b4c6477c4e38755cd28db14f852bfe659e3410bb93aeeedcc02e8b9b
3
+ size 2359341
FtAyT4oBgHgl3EQf5Poe/vector_store/index.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:167201d1716483ab82f04f6dd7b25991a118343fdf22ea315d3cd6beb9def9ec
3
+ size 85733
G9AyT4oBgHgl3EQfS_dT/content/2301.00096v1.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8b85a1218bb19dd460e659ff61492d2b22efb9029fcc7398c8cd5e0f8960ce9
3
+ size 558581
HNAyT4oBgHgl3EQfrvkK/vector_store/index.faiss ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1da31b3bf6a7e89a47397ac4155b65d254620959296b5c3bdf37a743394305fe
3
+ size 4653101
INAzT4oBgHgl3EQfxv6_/content/tmp_files/2301.01744v1.pdf.txt ADDED
The diff for this file is too large to render. See raw diff
 
INAzT4oBgHgl3EQfxv6_/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
KdE2T4oBgHgl3EQfAgZ0/content/tmp_files/2301.03592v1.pdf.txt ADDED
@@ -0,0 +1,1427 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ EFFICIENT INTRA-RACK RESOURCE DISAGGREGATION FOR
2
+ HPC USING CO-PACKAGED DWDM PHOTONICS
3
+ A PREPRINT
4
+ George Michelogiannakis
5
+ Lawrence Berkeley National Laboratory
6
7
+ Yehia Arafa
8
+ Qualcomm Technologies, Inc
9
10
+ Brandon Cook
11
+ Lawrence Berkeley National Laboratory
12
13
+ Liang Yuan Dai
14
+ Columbia University
15
16
+ Abdel-Hameed Badawy
17
+ New Mexico State University
18
19
+ Madeleine Glick
20
+ Columbia University
21
22
+ Yuyang Wang
23
+ Columbia University
24
25
+ Keren Bergman
26
+ Columbia University
27
28
+ John Shalf
29
+ Lawrence Berkeley National Laboratory
30
31
+ January 11, 2023
32
+ ABSTRACT
33
+ The diversity of workload requirements and increasing hardware heterogeneity in emerging high
34
+ performance computing (HPC) systems motivate resource disaggregation. Disaggregation separates
35
+ servers into their constituent compute and memory resources so that they can be allocated as required
36
+ to each workload. Previous work has shown the potential of intra-rack resource disaggregation,
37
+ but it is not clear how to realize these gains and cost-effectively meet the stringent bandwidth
38
+ and latency requirements of HPC applications. To that end, we describe how modern photonic
39
+ components can be co-designed with modern HPC racks to implement flexible intra-rack resource
40
+ disaggregation and fully meet the high escape bandwidth of contemporary multi chip module (MCM)
41
+ packages and all chip types in modern HPC racks with negligible power overhead. We show how to
42
+ use distributed indirect routing to meet these demands without the need for significant complexity
43
+ for reconfiguration that spatial optical switches require. We then show that intra-rack resource
44
+ disaggregation implemented using emerging photonics and parallel optical wavelength-selective
45
+ switches satisfies bit error rate (BER) and bandwidth constraints and provides an average application
46
+ speedup of 23.9% for 31 out-of-order CPU and 61.5% for 27 GPU benchmarks compared to a similar
47
+ system that instead uses modern electronic switches for disaggregation, due to their higher latency.
48
+ Using observed resource usage from a production system, we estimate that an iso-performance
49
+ intra-rack disaggregated HPC system using photonics would require 4× fewer memory modules and
50
+ 2× fewer NICs than a non-disaggregated baseline.
51
+ Keywords Photonics · disaggregation · AWGR · spatial
52
+ arXiv:2301.03592v1 [cs.DC] 9 Jan 2023
53
+
54
+ arXiv Template
55
+ A PREPRINT
56
+ 1
57
+ Introduction
58
+ Leading high performance computing (HPC) systems are steadily embracing heterogeneity of compute and memory
59
+ resources as a means to preserve performance scaling and reduce system power Liu et al. [2012], Top [2018], Ujaldón
60
+ [2016]. This trend is already apparent with the integration of GPUs Mittal and Vetter [2015], Tiwari et al. [2015],
61
+ Gao and Zhang [2016] and is expected to continue with fixed-function or reconfigurable accelerators such as field
62
+ programmable gate arrays (FPGAs), Milojicic [2020], Asaadi and Chapman [2017], Segal et al. [2014], Hogervorst
63
+ et al. [2021], Lant et al. [2020], Dimond et al. [2011], Ramirez-Gargallo et al. [2019], emerging customized accelerators,
64
+ and heterogeneous memory Venkata et al. [2017]. In addition, key HPC workloads show considerable diversity in
65
+ computational and memory access patterns Michelogiannakis et al. [2022], Rodrigo et al. [2016].
66
+ This expectation of resource heterogeneity, workload diversity, and today’s method of allocating resources to applications
67
+ in units of statically-configured nodes where every node is identical and unused resources are left to idle, raises the
68
+ concern of resource underutilization (referred to as “marooned resources”). These marooned resources increase both
69
+ capital and operational costs without improving performance. This has led to the emergence of resource disaggregation.
70
+ Disaggregation refers to decomposing servers into their constituent compute and memory resources so that these can be
71
+ allocated as required according to the needs of each workload. Hyperscale datacenters have readily embraced resource
72
+ disaggregation and have demonstrated that it significantly improves utilization of GPUs and memory Guleria et al.
73
+ [2019a], Peng et al. [2020], Taylor [2015], Li et al. [2022], Koh et al. [2019], Papaioannou et al. [2016], Gonzalez et al.
74
+ [2020], Cheng et al. [2019a], Guleria et al. [2019b].
75
+ Although file storage is routinely disaggregated in modern systems Per, Michelogiannakis et al. [2022], Petersen and
76
+ Bent [2017], HPC has been slow to embrace disaggregation of compute and memory resources Glick et al. [2020], Guo
77
+ et al. [2021] due to the sensitivity of HPC workloads to bandwidth and latency that cannot be met by current PCIe/CXL
78
+ or Ethernet link technologies used in contemporary disaggregated architectures. Studies showed that disaggregation only
79
+ among resources in the same rack (i.e., intra-rack resource disaggregation) in HPC could reduce resources by 5.36% to
80
+ 69.01% while avoiding the overhead of full-system disaggregation Michelogiannakis et al. [2022], but the impact of
81
+ increased memory latency and specific architectural trade offs have not been explored. Thus, although disaggregation
82
+ using electronic networks have been demonstrated in hyperscale datacenters Lin et al. [2020], Papaioannou et al. [2016],
83
+ Call et al. [2020], minimizing adverse effects to and addressing the stringent bandwidth density and latency demands of
84
+ HPC workloads requires a thorough investigation.
85
+ In this work, our contributions are as follows. Firstly, we describe how to use emerging photonic links and switches to
86
+ design modern and practical resource-disaggregated HPC racks based on an existing GPU-accelerated HPE/Cray EX
87
+ supercomputer Per. Secondly, we show how state of the art commercially available photonics and advanced packaging
88
+ multi chip modules (MCMs) meet bit error rate (BER) requirements, impose a negligible power overhead, and deliver
89
+ sufficient bandwidth to satisfy the escape bandwidth of all chips in modern HPC racks. Thirdly, we show how to use
90
+ distributed indirect routing and arrayed waveguide grating routers (AWGRs) Liu et al. [2020], Zhang et al. [2019] to
91
+ satisfy all bandwidth requirements without the overhead and latency for reconfiguration that spatial Seok et al. [2019a],
92
+ Ding et al. [2016] and wave-selective Huang et al. [2020] switches require. Moreover, having demonstrated negligible
93
+ adverse impact to all other metrics, we show that intra-rack disaggregation using emerging photonics provides an
94
+ average application speedup of 23.9% for 31 out-of-order (OOO) CPU and 61.5% for 27 GPU benchmarks compared to
95
+ a similar system that instead uses state of the art electronic switches, which also increase power overhead by three orders
96
+ of magnitude. Finally, based on observed resource usage, we estimate that a system based on racks using photonics
97
+ for resource disaggregation can have 43% fewer overall chips compared to a non-disaggregated system with the same
98
+ computational throughput.
99
+ 2
100
+ Related Work
101
+ Hyperscale datacenters predominantly focus on full-system resource disaggregation where applications can allocate
102
+ fine-grain resources of different types, today typically graphical processing units (GPUs) Guleria et al. [2019a] and
103
+ memory Peng et al. [2020], Lim et al. [2009], Koh et al. [2019], Gonzalez et al. [2020], which are located anywhere in
104
+ the system or within a group of racks Lin et al. [2020], Papaioannou et al. [2016], Call et al. [2020]. In such a system,
105
+ resources of the same type are typically placed in the same rack.
106
+ However, full-system, flexible, and fine-grain disaggregation introduces significant overhead because of the higher
107
+ latency and lower bandwidth density of contemporary hardware that is used to implement resource disaggregation –
108
+ typically PCIe, 100Gig Ethernet, and eventually compute express link (CXL) Van Doren [2019] over electronic links.
109
+ This overhead does not simply increase power and procurement cost, but rather adds potentially substantial latency
110
+ between key resources such as central processing units (CPUs) and memory that traditionally exhibit latency-sensitive
111
+ 2
112
+
113
+ arXiv Template
114
+ A PREPRINT
115
+ communication. The aforementioned studies quote a several orders of magnitude increase in network and memory
116
+ latency due to full-system resource disaggregation to improve resource utilization by 35% at most Zervas et al. [2018].
117
+ Another study found that application performance degradation depends on both network bandwidth and latency, but
118
+ can still reach up to 40% even with high bandwidth, low-latency networks Gao et al. [2016]. Work on SPEC and
119
+ commercial benchmarks also found an up to 27% application slowdown due to the additional memory latency Abali
120
+ et al. [2015]. A study on Microsoft’s Azure found a range of performance slowdowns up to 30% from an extra 65 ns to
121
+ access main memory Li et al. [2022]. Software defined networks (SDNs) based on electrical networks fare no better in
122
+ terms of overhead Gao et al. [2016], Han et al. [2013], Call et al. [2020].
123
+ Hybrid full-system photonic–electronic approaches have also been proposed that rely on circuit switching Zervas et al.
124
+ [2018] for reconfiguration. As a result, a few studies call for intra-rack disaggregation in datacenters Taylor [2015],
125
+ Lim et al. [2009], Guleria et al. [2019b]. Even the low latency and high bandwidth density of modern photonics cannot
126
+ fully satisfy the bandwidth, energy, and latency requirements of full system disaggregation. This makes system-wide
127
+ disaggregation impractical in many cases Lin et al. [2020], Zervas et al. [2018], Cheng et al. [2019a], Cheng et al.
128
+ [2018].
129
+ Recent full system approaches in high performance computing (HPC) rely on optics to connect CPUs and memory, and
130
+ electronic switches for hard disk drives (HDDs) to increase resource CPU utilization by 36.6% and memory 21.5% Guo
131
+ et al. [2021]. In contrast, another study confirms that production HPC systems can reduce resources from 5.36%
132
+ to 69.01% with intra-rack disaggregation and still satisfy the worst-case average rack utilization Michelogiannakis
133
+ et al. [2022]. Similar to datacenters, intra-rack disaggregation in HPC promises the lowest overhead and impact to
134
+ applications Glick et al. [2020], Taylor [2015], Guleria et al. [2019b].
135
+ Related work has research other aspects necessary to make resource disaggregation practical in a system, such as
136
+ job scheduling Fan et al. [2019], Agosta et al. [2018], Amaral et al. [2021], Domeniconi et al. [2019], how the
137
+ operating system (OS) and runtime should adapt Maccabe [2017], Hwu et al. [2015], Shan et al. [2018], programming
138
+ of code portability in heterogeneous systems Gioiosa et al. [2020], Agosta et al. [2018], partitioning of application
139
+ data Khaleghzadeh et al. [2020], fault tolerance Hussain [2020], how to fairly compare the performance of different het-
140
+ erogeneous systems Jamieson et al. [2018], and the impact of heterogeneous resources to application performance Tang
141
+ et al. [2017], Lastovetsky [2015], Venkata et al. [2017]. These are important but out of scope topics for our study.
142
+ 2.1
143
+ Under-utilization in Production Systems
144
+ We use NERSC’s Cori system as an exemplar production HPC system, while recognizing workload requirements
145
+ on other systems may differ. In NERSC’s Cori, before Perlmutter came online and thus Cori was runnign the full
146
+ NERSC workload, three quarters of the time Haswell nodes use less than 17.4% of memory capacity (50.1% for
147
+ KNL nodes) and less than 0.46 GB/s of memory bandwidth Michelogiannakis et al. [2022]. These observations are
148
+ similar to observations collected on LANL clusters Peng et al. [2020] and Alibaba machines that execute batch jobs.
149
+ Likewise, half of the time Cori nodes use no more than half of their compute cores and three quarters of the time 1.25%
150
+ of available network interface controller (NIC) bandwidth. Similarly, in Lawrence Livermore National Laboratory
151
+ (LANL) clusters, approximately 75% of the time, no more than 20% of memory capacity is used Peng et al. [2020].
152
+ Alibaba’s published data Guo et al. [2019] show that memory is underutilized similarly to Cori for machines that
153
+ execute batch jobs. Data from Google systems shows that memory and disk capacity of tasks is spread over three orders
154
+ of magnitude and typically underutilized Han et al. [2013]. Azure reports 25% of memory under-utilization Li et al.
155
+ [2022]. Datacenters have also reported 28% to 55% CPU idle in the case of Google trace data Patel et al. [2015] and
156
+ 20% to 50% most of the time in Alibaba Guo et al. [2019]. Early studies also suggest GPU under-utilization Li et al.
157
+ [2015], Jeon et al. [2019], Li et al. [2011].
158
+ 3
159
+ Photonics for Resource Disaggregation
160
+ Here we walk through available optical link and switch technologies and argue that photonics today meet the strict
161
+ performance and error rate requirements to efficiently implement intra-rack resource disaggregation in HPC.
162
+ 3.1
163
+ Memory Technologies and Requirements
164
+ IO systems in HPC are already largely disaggregated over conventional system-scale interconnects since the underlying
165
+ technologies (disk or SSD) are relatively high latency and lower bandwidth Michelogiannakis et al. [2022], Terzenidis
166
+ et al. [2018]. By contrast, memory technologies (particularly high bandwidth memorys (HBMs) needed by GPUs) are
167
+ much higher bandwidth and much less tolerant of latency and require much lower bit error rates (BERs). Given that
168
+ memory disaggregation imposes the most challenging constraints among other resources in today’s compute nodes, we
169
+ 3
170
+
171
+ arXiv Template
172
+ A PREPRINT
173
+ Figure 1: Logical schematic of a DWDM link using ring resonator technology and a comb-laser source. Each ring is
174
+ tuned to a different frequency of light and can be used to modulate that specific wavelength of light (a channel). Comb
175
+ laser sources provide a comb of frequencies of light to provide those wavelengths for encoding. All of the encoded
176
+ optical channels share the same optical fiber and are decoded using the rings on the receiving side to route channels to
177
+ the photodetectors.
178
+ Aggregated comb laser sources
179
+ Active photonic interposer
180
+ (a) Active Optical MCM
181
+ (b) Blade
182
+ Optical Circuit Switches
183
+ Optical Fiber
184
+ (c) Rack/Pod
185
+ Figure 2: Overall physical structure of Rack/Pod scale resource disaggregation from photonically connected MCMs
186
+ to the Rack/Pod scale pooling of disaggregated resources. The conversion from CXL-over-fiber to HBM or NVM
187
+ electrical protocol is implemented in the active interposer for the photonics MCM.
188
+ will use DDR and HBM memory technology to set our performance target. A typical DDR4 memory has a response
189
+ latency of approximately 90 ns and for HBM the average response latency is 90-140 ns Wang et al. [2020]. Still, any
190
+ added latency between the CPU and memory from resource disaggregation may penalize application performance as
191
+ we quantify later. Server-class memories typically require BERs of less than 10−18 to achieve tolerable failures in
192
+ time (FIT) rates with conventional single-error-correct/double-error-detect (SEC-DED) protection Meza et al. [2015],
193
+ Sridharan et al. [2015]. Forward error correction (FEC) can reduce the BER, but with additional latency Luyi et al.
194
+ [2012].
195
+ 3.2
196
+ Optical Link Technologies
197
+ We consider a range of photonic link technologies that include conventional 100 Gbps Ethernet physical interfaces that
198
+ represent the current baseline link technology for memory disaggregation. We also introduce a range of cutting-edge
199
+ dense wavelength division multiplexing (DWDM) link technologies that are either demonstrated as research prototypes
200
+ or are commercially available. The photonic components all come from existing commercial technologies (100 Gbps,
201
+ 400 Gbps, Ayar TeraPhy) and some research prototypes from DARPA PIPES (the 1-2 Tb link technologies). These
202
+ higher performance link technologies must be co-packaged in order to achieve their bandwidth density. These link
203
+ technologies are summarized in Table 1. The technology for the optical links is depicted in Figure 1. Delivering
204
+ multiple channels of laser light to the package has been challenging to scale cost-effectively if each "color" of light
205
+ were to require a separate laser source. This concern was alleviated by the emergence of quantum dot and soliton comb
206
+ laser sources shown in Figure 4 that can produce hundreds of usable light frequencies with wall-plug efficiencies of up
207
+ to 41% Kim et al. [2019a].
208
+ 4
209
+
210
+ 7NNN00NNN:.MMCustomAccelGPuCPUDisaggregatedresourceRack/PodarXiv Template
211
+ A PREPRINT
212
+ Figure 3: Copackaged optics are required for DWDM link technologies to achieve the kind of bandwidth density
213
+ required to operate at native memory bandwidths.
214
+ 3.3
215
+ Active Photonic MCMs
216
+ Many CPUs and GPUs do not have the necessary off-chip bandwidth for full utilization of their compute resources
217
+ because operating their I/O pins at higher bandwidth incurs a power cost Chen et al. [2017], Jouppi et al. [2017].
218
+ Using emerging high-speed optical links directly to the MCM, illustrated in Figure 3 provides to the order of 10×
219
+ gains in escape bandwidth Glick et al. [2020], Wade [2019], Bergman et al. [2018], Maniotis et al. [2021]. This is
220
+ a necessary property to enable efficient resource disaggregation as well as handle changing bandwidth requirements
221
+ of key applications such as machine learning that drastically shifts bandwidth between inter-GPUs and off-chip from
222
+ inference to training.
223
+ MCMs with integrated photonics have been demonstrated in both 2.5D and 3D interposer platforms Glick et al. [2020],
224
+ Minkenberg et al. [2021], Sutono et al. [1998], Abrams et al. [2020]. They can use different die-to-die link standards
225
+ such as UCIe. Active interposer platforms combine the photonic integrated circuit (PIC) and interposer into a single
226
+ integrated substrate. The active interposer allows photonic components to be fabricated and directly integrated with
227
+ through silicon vias (TSVs) and additional metal redistribution layers. Electronic circuits are flip-chipped on top of
228
+ active interposers using copper pillars Dittrich et al. [2017]. Further work has embedded photonic switch fabrics within
229
+ MCM platforms with a crosstalk suppression and extinction ratio of >50dB and on-chip loss as low <1.8dB Glick et al.
230
+ 5
231
+
232
+ 2.5DIntegratedTransceiver
233
+ ComputeNode
234
+ Optical
235
+ Fibers
236
+ FCBumps
237
+ PIC
238
+ FC Bumps
239
+ EIC
240
+ Interposer
241
+ ASIC/FPGA
242
+ PCBEIO
243
+ PIC
244
+ 25umpitchpad
245
+ array to be bumpec
246
+ 25umpitchpad
247
+ arraytobeb
248
+ MCM
249
+ C3
250
+ ElectricalloonPcBto
251
+ be wirebondedto PiC
252
+ EIC
253
+ PICarXiv Template
254
+ A PREPRINT
255
+ Figure 4: Comb laser sources provide the many discrete optical frequencies for the DWDM link with up to 41%
256
+ experimentally measured conversion efficiency.
257
+ [2020]. This was further scaled up to support more than 100 ports with microring resonators using a scalable switch
258
+ fabric that combined switching in the space domain with wavelength-selectivity to define fine-grained connectivity for
259
+ node disaggregation Huang et al. [2020], Glick et al. [2020].
260
+ 3.3.1
261
+ Link Protocol
262
+ We adopt CXL as our link protocol Van Doren [2019]. CXL is an overlay on the PCIe-Gen6 physical layer, it includes
263
+ guaranteed ordering of events and is a broadly adopted industry standard with published specifications. However, our
264
+ study does not rely on any features of any particular protocol so alternatives such as UCIe also apply.
265
+ 3.3.2
266
+ Link Propagation and Encoding/Decoding Latency
267
+ The target reach for an intra-rack disaggregation solution is approximately 1-4 meters. Given the speed of light c
268
+ and light propagating through optical material that has an index of refraction that is near r1.5, the effective latency
269
+ of propagating through an optical fiber at nominally 0.75c is approximately 5 ns per meter. Therefore, rack-scale
270
+ disaggregation adds 5-15 ns to our latency budget, less than 20% of the typical DRAM latency. The link latency for
271
+ SERDES and photonic ring modulation is negligible. Intra-rack fiber lengths up to 4 meters require no intervening
272
+ Electrical Optical (OEO) conversions.
273
+ 3.3.3
274
+ Bit Error Rates and FEC
275
+ To achieve 10−18 BER required for memory technologies, FEC Luyi et al. [2012] will likely be required. Using the
276
+ lightweight FEC scheme that is proposed for CXL Van Doren [2019] and PCIe Gen6 Sharma [2020] as an example,
277
+ the all-inclusive latency for FEC can be as low as 2 ns. Therefore, for 200 Gbps, serialization delay is 10 ns and the
278
+ FEC calculations add 2-3 ns. At 400 Gbps and above, the net latency for FEC would be 5 ns plus 2-3 ns. Of note, this
279
+ approach to achieving these BER targets is achievable with less than a 0.1% bandwidth loss.
280
+ 6
281
+
282
+ 100um-6 dBm
283
+ 10
284
+ -10 dBm
285
+ S-band
286
+ C-band
287
+ L-band
288
+ 0
289
+ Bm)
290
+ -10
291
+ NO
292
+ -20
293
+ -30
294
+ -40
295
+ 1,520
296
+ 1,540
297
+ 1,560
298
+ 1,580
299
+ 1,600
300
+ Wavelength (nm)arXiv Template
301
+ A PREPRINT
302
+ BW
303
+ (Gbps)
304
+ Energy
305
+ (pJ/bit)
306
+ Link
307
+ Gbps ×
308
+ Chan-
309
+ nels
310
+ #Links
311
+ (2
312
+ TB/s
313
+ es-
314
+ cape)
315
+ Agg.
316
+ Ws (2
317
+ TB/s
318
+ es-
319
+ cape)
320
+ Ref.
321
+ 100
322
+ 30
323
+ 25 × 4
324
+ 160
325
+ 480
326
+ Fathololoumi
327
+ et
328
+ al.
329
+ [2021],
330
+ Agrell
331
+ et
332
+ al.
333
+ [2016]
334
+ 400
335
+ 30
336
+ 100 × 4
337
+ 40
338
+ 197
339
+ Wei
340
+ et
341
+ al.
342
+ [2015]
343
+ 768
344
+ < 1
345
+ 32 × 24
346
+ 21
347
+ 14.4
348
+ Wade
349
+ [2019]
350
+ 1,024
351
+ 0.45
352
+ 16 × 64
353
+ 16
354
+ 7.2
355
+ Kim
356
+ et
357
+ al.
358
+ [2019b]
359
+ 2,048
360
+ 0.3
361
+ 16 × 128
362
+ 8
363
+ 4.8
364
+ Kim
365
+ et
366
+ al.
367
+ [2019b]
368
+ Table 1: A range of WDM photonic link technologies.
369
+ In terms of impact on BER, this PCIe/CXL-like correction scheme corrects all single bursts of up to 16 bits. Double
370
+ bursts will likely be mis-corrected, but the chance of a bad flit decreases quadratically (e.g., a flit BER of 10−6 becomes
371
+ 10−12 as you need two error bursts per flit to fail). Each flit is protected with a strong 64-flit CRC such that the flit FIT
372
+ rate (CRC escapes) is significantly less than one part per billion. Lastly, the FEC escapes become link retransmissions
373
+ and the ASIC-to-ASIC connection sees close to zero errors. As a result, emerging memory fabric protocols such as
374
+ CXL, which could be run over our evaluated physical links, are capable of achieving a BER rate that meets the stringent
375
+ memory system requirements and minimizes performance loss due to retransmission.
376
+ 3.4
377
+ Optical Circuit Switch Technologies
378
+ Motivated by minimizing latency, our vision for a disaggregated rack is to have photonically-enabled MCMs that are
379
+ connected via an optical circuit switch, as shown in Figure 2. Compute and memory chips would be in the center of the
380
+ MCM and the edge of the MCM would contain co-packaged optical silicon in-package photonics (SiPs). Switches with
381
+ all-optical paths include spatial- and wave-selective approaches, shown in Table 2.
382
+ 3.4.1
383
+ Spatial Optical Switches
384
+ In recent years, the primary switching cells investigated are microelectromechanical systems (MEMS) actuated
385
+ couplers, Mach-Zehnder interferometers (MZIs), and microring resonators (MRRs). Taking after their free-space
386
+ counterpart, photonic MEMS-actuated switches are broadband spatial switches that have demonstrated radix scaling up
387
+ to 240×240 Seok et al. [2019b]. However, MEMS switching cells generally require high driving voltages (greater than
388
+ 20 V), which make them less attractive for co-integration with electronic drivers, but typically offer low inter-channel
389
+ cross-talk and low optical losses. Spatial switches can also use mirrors Cal, photonic integrated circuits Ding et al.
390
+ [2016], or tiled planar silicon photonics Seok et al. [2019a]. MZI switches are more co-integration friendly compared
391
+ to MEMS but have only been shown to scale up to 32×32 Ikeda et al. [2020]. This limit can be seen as a consequence
392
+ of the higher insertion-loss scaling resulting from cascaded MZI cells, as well as the susceptibility of popular MZI
393
+ topologies to first-order crosstalk.
394
+ The challenge for scaling-up the spatial approach is the quantization of package and MCM escape bandwidth and
395
+ reduced configuration options. For example, at 768 Gbps (the Ayar TeraPhy Wade [2019]), the number of fibers escaping
396
+ the package is 21 fibers, which means the package can be connected only up to 21 different potential destinations using
397
+ a spatial switch.
398
+ 7
399
+
400
+ arXiv Template
401
+ A PREPRINT
402
+ Switch
403
+ Type
404
+ Radix
405
+ Wave-
406
+ lengths
407
+ per
408
+ port
409
+ B/W
410
+ per
411
+ channel
412
+ (wave-
413
+ length)
414
+ Insertion
415
+ Loss
416
+ Crosstalk
417
+ Mach-
418
+ Zehnder
419
+ based Ikeda
420
+ et
421
+ al.
422
+ [2020]
423
+ 32×32
424
+ 1
425
+ 439
426
+ Gbps
427
+ 12.8
428
+ dB
429
+ -26.6
430
+ dB
431
+ MEMS-
432
+ actuated Seok
433
+ et
434
+ al.
435
+ [2019b]
436
+ 240×240 1
437
+
438
+ 9.8 dB
439
+ -70 dB
440
+ Microring
441
+ res-
442
+ onator Khope
443
+ et
444
+ al.
445
+ [2017],
446
+ Cheng
447
+ et
448
+ al.
449
+ [2019b]
450
+ 8×8
451
+ (128×128)
452
+ 8
453
+ (128)
454
+ 100
455
+ Gbps
456
+ (42
457
+ Gbps)
458
+ 5dB
459
+ (10dB)
460
+ (-35
461
+ dB)
462
+ Casc.
463
+ AW-
464
+ GRs Sato
465
+ [2018]
466
+ 370×370 370
467
+ 25
468
+ Gbps
469
+ 15 dB
470
+ -35 dB
471
+ Table 2: High-radix CMOS-compatible photonic switches.
472
+ 3.4.2
473
+ Wavelength Selective Optical Switches and AWGRs
474
+ The inherent wavelength-selectivity of MRR switching cells allows for the straightforward implementation of
475
+ wavelength-selective switching (WSS) topologies. This enables one to establish all-to-all networks by leveraging
476
+ wavelength-division multiplexing (WDM). Currently, MRR-based switches with the largest radix include the 8×8
477
+ crossbar Khope et al. [2017] and switch-and-select Nikolova et al. [2017], but have been experimentally emulated to
478
+ include a 16×16 Clos Dai et al. [2020]. The metrics in Dai et al. [2020] can be seen to correlate very closely with the
479
+ scaling proposed in Cheng et al. [2019b], making a practical case for the 128×128 shown in Table 2.
480
+ All-to-all networks via WDM signals can also be achieved by arrayed waveguide grating routers (AWGRs) Liu et al.
481
+ [2020], Zhang et al. [2019], Proietti et al. [2013], Lea [2015], Terzenidis et al. [2018]. As AWGRs are passive optical
482
+ elements, no reconfiguration is possible within the routing fabric itself. Instead, fast wavelength-tunable lasers must
483
+ be leveraged at the transmitter of every node if it wishes to address a different destination since AWGRs shuffle the
484
+ light frequencies such that one lambda goes to each endpoint from each source. AWGRs enable us to implement an
485
+ N×N all-to-all topology using just O(N) fibers (each carrying N frequencies of light) whereas an implementation
486
+ using copper would require N2 wires. Although the cost of fast wavelength-tunable lasers is still an ongoing research
487
+ topic Dhoore, Sören and Roelkens, Günther and Morthier, Geert [2019], AWGRs are mature, commercially available,
488
+ and well established in literature FSp.
489
+ In AWGRs, only a limited number of ports can be practically supported due to the walk-off of passband center
490
+ frequencies from the carrier wavelength grid and the worse crosstalk associated with a larger number of ports (N).
491
+ A feasible implementation of AWGR-based optical switches with large N has been demonstrated utilizing cascaded
492
+ small-size AWGRs Sato [2018]. Specifically, N M × M AWGRs (front-AWGRs) are interconnected with M N × N
493
+ AWGRs (rear-AWGRs) to effectively act as an MN × MN AWGR. Each output port of a front-AWGR is connected
494
+ to an input port of a rear-AWGR, where the interconnection pattern can be optimized with knowledge of port-specific
495
+ insertion losses to minimize the worst-case insertion loss of the aggregated AWGR. Further up-scaling of the switch
496
+ radix can be achieved by interconnecting small K × K delivery-coupling switchs (DC-switchs) with multiple copies
497
+ of the MN × MN AWGRs, yielding a KMN × KMN switching capability. This architecture has been verified by
498
+ hardware prototypes of 270 × 270 and 1440 × 1440 Sato et al. [2013], Ueda et al. [2016], showing ∼15 dB insertion
499
+ loss and below −35 dB crosstalk suppression. In order to accommodate the 350 MCMs of our rack, a reasonable
500
+ 8
501
+
502
+ arXiv Template
503
+ A PREPRINT
504
+ 1
505
+ 5
506
+ 3
507
+ 2
508
+ 8
509
+ 4
510
+ 6
511
+ 7
512
+ Figure 5: With an AWGR, endpoint 1 has one wavelength directly connecting it to endpoint 3. If it desires more
513
+ bandwidth, it can route through another intermediate endpoint (indirect routing) chosen in a Valiant fashion Liu et al.
514
+ [2020], Teh et al. [2020]. Here, the link from 1 to 7 is available (green) but the link from 7 to 3 is not (red). The chosen
515
+ path is from 1 to 6 to 3 because both links are available.
516
+ configuration is KMN = 3 × 12 × 11 = 396. This results in 370 ports and 370 wavelengths per port (Table 2). Since
517
+ AWGRs typically have a 25 GHz optical bandwidth if the wavelength grid is 50 GHz, with PAM4 we assume 25 Gbps
518
+ per wavelength Dai et al. [2020], Bhoja [2017].
519
+ Wave-selective switches Huang et al. [2020], Marom et al. [2017] can steer any subset of wavelengths to a given
520
+ destination, not just all (spatial) or one (AWGR). Dynamic programming methods can avoid sending the same frequency
521
+ of light from two different sources to the same destination. Since this is a relatively new technology, we constructed a
522
+ model shown in Table 2 that projects the performance of a larger radix switch that is comprised of smaller demonstrated
523
+ building blocks.
524
+ 3.4.3
525
+ Reconfiguration Time
526
+ Spatial and wave-selective switches typically require centralized scheduling Teh et al. [2020] to reach a steady globally
527
+ optimal solution. The reconfiguration time can range from tens of nanoseconds to tens of milliseconds. In production
528
+ HPC systems, multi-node jobs start every few seconds and last from minutes to hours Michelogiannakis et al. [2019,
529
+ 2022]. Also, job resource usage and communication becomes predictable early, does not change fast, and typically
530
+ remains predictable throughout a job’s execution time Michelogiannakis et al. [2022, 2019], Shalf et al. [2005], Vetter
531
+ and Mueller [2002]. Therefore, even milliseconds of reconfiguration time is ample.
532
+ 4
533
+ Control Logic
534
+ Here we describe how we can perform indirect routing to increase point-to-point bandwidth using only per-source logic.
535
+ 4.1
536
+ Indirect routing in AWGRs
537
+ AWGRs dedicate exactly one wavelength between any source–destination pair. If a source–destination pair requests
538
+ more bandwidth than what a single wavelength can satisfy, sources can use indirect routing an example of which is
539
+ shown in Figure 5. Sources can split traffic to N intermediate destinations in parallel in order to use the bandwidth of
540
+ 9
541
+
542
+ arXiv Template
543
+ A PREPRINT
544
+ N wavelengths. This does not consume additional power in the photonic components assuming lasers are constantly
545
+ powered. Sources consider indirect paths only if the direct (single-hop) bandwidth to their desired destination does
546
+ not suffice. A source considers indirect destinations for which the direct bandwidth from the source is available and
547
+ whose wavelengths from the intermediate hop to the desired final destination is available. Among potentially multiple
548
+ candidates, sources choose one in a Valiant fashion Liu et al. [2020], Teh et al. [2020], Domke et al. [2019]. This
549
+ is done on a per-flow basis in order to avoid out of order packet delivery. This routing logic can be modelled as an
550
+ allocator problem and implemented with a low latency and area penalty Ma et al. [2014], Becker and Dally [2009].
551
+ Indirect routing relies on sources knowing which other sources attached to the same AWGR are utilizing their local
552
+ wavelengths in order to identify a productive intermediate destination. For instance, in Figure 5 endpoint 1 should
553
+ know whether the wavelengths from 7 to 3 and 6 to 7 are occupied. For that, we rely on piggybacking where traffic
554
+ between a source and a destination periodically includes the state of the sources’s wavelengths as a way to broadcast
555
+ local state to the rest of the endpoints attached to the same AWGR Jiang et al. [2009]. In the case of a N×N AWGR,
556
+ each source uses N bits to encode which of its N local wavelengths it is using with one-hot encoding. Even if we
557
+ piggyback this information multiple times a second, the bandwidth impact is negligible. For instance, if we multiplex
558
+ multiple flows into a wavelength and therefore denote 8 bits per wavelength, the status vector per source becomes
559
+ 256 × 8 = 2048bits = 256bytes. If, due to stale information, sources pick an intermediate destination whose
560
+ wavelength direct to the final destination is not available, the intermediate destination performs indirect routing through
561
+ a second intermediate destination, and so on. If no data is exchanged between a pair, thus presenting no opportunity for
562
+ piggybacking, that pair can exchange a separate control message with the same information.
563
+ 4.2
564
+ Spatial and Wave-Selective Switches
565
+ Spatial and wave-selective switches can use indirect routing in tandem with reconfiguration. Indirect routing reduces
566
+ the need for reconfiguration, but intermediate hops should be chosen among hops that already have a direct connection
567
+ with the final destination; otherwise, the intermediate hop itself may trigger a reconfiguration. The synergy between
568
+ indirect routing and switch reconfiguration was explored in Teh et al. [2020].
569
+ 5
570
+ Disaggregated Rack Design
571
+ For the rest of our study, we will model an HPC rack based on a GPU-accelerated HPE/Cray EX Supercomputer Per
572
+ where a rack contains 128 GPU-accelerated nodes. Each node of our model system contains an AMD Milan CPU
573
+ that has eight memory controllers each supporting a 3200MHz DDR4 module. Therefore, each CPU has 256 GB of
574
+ memory with a maximum bandwidth of 204.8 GBps. A compute node also has four NVIDIA Ampere A100 GPUs.
575
+ Each GPU supports 12 third generation NVLink links each supporting 25 GBps per direction. Each GPU also has 40
576
+ GB of co-located HBM with a bandwidth of 1555.2 GBps. Each node also has four 31.5 GBps PCI Gen4 links to
577
+ connect each GPU to the CPU. The CPU also connects to four Slingshot 11 NICs with 200 Gbps per direction De Sensi
578
+ et al. [2020a]. Note that our photonic disaggregation hardware is orthogonal to and thus does not impair past work
579
+ related to disaggregation such as runtimes, OS support, endpoint sharing management, and security.
580
+ 5.1
581
+ MCMs and Escape Bandwidth
582
+ We organize chips within each rack into an MCMs package. For simplicity, we restrict all MCMs to have the same
583
+ escape bandwidth and we place chips of only the same type in MCMs. We then make conservative assumptions for next
584
+ generation photonics that are entering the market today based on our analysis of Section 3. In particular, each MCM has
585
+ 32 optical fibers attached to it, a conservative assumption compared to the five arrays of 24 fibers demonstrated Hosseini
586
+ et al. [2021]. Each fiber supports 64 wavelengths (channels) of 25 Gbps each for a 6400 GBps escape bandwidth per
587
+ MCM. We vary the number of chips per MCM such that each chip enjoys the same escape bandwidth as in our baseline
588
+ rack Per. Therefore, our photonic architecture does not restrict chip escape bandwidth. Table 3 shows the number of
589
+ chips per MCM and the total number of MCMs containing chips of that type to satisfy chip escape bandwidth. Each
590
+ MCM contains a controller chip that interfaces the native protocol of the disaggregated resource to the CXL protocol
591
+ over the photonic links. CXL’s overhead and its associated FEC is included in our model of the overall architecture.
592
+ 5.2
593
+ Optical Switches
594
+ The radix and wavelengths per port of optical switches dictate number of MCMs we can fully connect optically with
595
+ a single switch as well as the amount of direct (single-hop) bandwidth. From Section 3.4, we pick state-of-the-art
596
+ representatives of wave-selective, cascaded AWGRs, and spatial optical switches. Their parameters are shown in
597
+ Table 4. Even though spatial Seok et al. [2019b] and wave-selective switches Huang et al. [2020] are capable of 100
598
+ 10
599
+
600
+ arXiv Template
601
+ A PREPRINT
602
+ Chip type
603
+ Chips per MCM
604
+ # MCMs per rack
605
+ CPU
606
+ 14
607
+ 10
608
+ GPU
609
+ 3
610
+ 171
611
+ NIC
612
+ 203
613
+ 3
614
+ HBM
615
+ 4
616
+ 128
617
+ DDR4
618
+ 27
619
+ 38
620
+ Total
621
+ 350
622
+ Table 3: The number of chips ((CPU, GPU, NIC, HBM, or DDR4 module) per MCM and MCMs in a rack assuming
623
+ 32 fibers per MCM, 64 wavelengths of 25 Gbps per fiber. The target BER to and from memory is 10−18 (Section 3.1).
624
+ Switch type
625
+ State of the art
626
+ Switch radix
627
+ Cascaded AWGRs Sato [2018]
628
+ 370
629
+ Spatial Seok et al. [2019b]
630
+ 240
631
+ Wave-Selective Huang et al. [2020]
632
+ 256
633
+ Gbps per wavelength
634
+ All switches
635
+ 25
636
+ Wavelengths per port
637
+ Cascaded AWGRs Sato [2018]
638
+ 370
639
+ Spatial Seok et al. [2019b]
640
+ 240
641
+ Wave-Selective Huang et al. [2020]
642
+ 256
643
+ Table 4: Switch configuration for our study.
644
+ Gbps per wavelength, most links available widely today do not support that (Table 1). In addition, we show that we can
645
+ still satisfy bandwidth demands with the conservative assumption of 25 Gbps per wavelength.
646
+ To connect our 350 MCMs using 370×370 AWGRs, we can combine MCM fibers in five groups of six and connect each
647
+ group to one port of five parallel AWGRs. However, this would require each AWGR port to handle 384 wavelengths.
648
+ To respect the per port 370 wavelength limitation of our AWGR configuration but still satisfy the full escape bandwidth
649
+ of MCMs, we combine the remaining 14 wavelengths along with the remaining two fibers per MCM (128 + 14 = 142
650
+ wavelengths total) that were left unconnected into an extra parallel AWGR, for a total of six parallel AWGRs. We then
651
+ connect MCM fibers to AWGRs in a staggered manner such that each MCM connects to each other MCM using at least
652
+ five 25 Gbps direct-path wavelengths, for a direct MCM–MCM bandwidth of 25 × 5 = 125 Gbps.
653
+ For simplicity, because of their relative small difference and because wave-selective switches can also achieve configu-
654
+ rations that spatial switches can, we treat both wave-selective and spatial switches as 256 ports with 256 wavelengths
655
+ per port. Each MCM can connect to 2048
656
+ 256 = 8 parallel switches. However, because the radix of optical switches is lower
657
+ than the number of MCMs, we instantiate 11 optical switches and connect MCMs in a staggered manner such that
658
+ optical switch with an index I connects to MCMs that have an index starting from (32 × I) mod 350 until (I + 255)
659
+ mod 350. This way, a small number of optical switch ports are left unconnected in order to not exceed the 32 fibers
660
+ per MCM. Similar to AWGRs, these ports can support future larger racks. If the switches configure appropriately,
661
+ each MCM has at least three direct paths to any other MCM. Each path has 256 wavelengths, thus the direct MCM
662
+ bandwidth is 256 × 3 × 25 = 2304 Gbps.
663
+ 6
664
+ Evaluation
665
+ Having previously evaluated in Section 3.3.3 that photonic switches satisfy BER requirements, in this Section we
666
+ analyze the impact of photonic-based intra-rack resource disaggregation to bandwidth, latency, and power. We then
667
+ compare against electronic switches and estimate system-wide savings.
668
+ 6.1
669
+ Bandwidth Evaluation
670
+ We distinguish two test cases based on Section 5.2: (A) Six parallel AWGRs and (B) 11 parallel wave-selective switches.
671
+ 6.1.1
672
+ Available Bandwidth
673
+ Using indirect routing and switch reconfiguration, any one particular MCM can use its full escape bandwidth to reach a
674
+ single destination MCM. In test case (A), all wavelengths escaping an MCM can reach the same destination MCM
675
+ using indirect routing. In test case (B), 768 wavelengths can be configured to route directly to a destination MCM
676
+ 11
677
+
678
+ arXiv Template
679
+ A PREPRINT
680
+ and the other 2048 − 768 = 1280 wavelengths can be configured to route indirectly through intermediate MCMs.
681
+ This assumes that other MCMs will not contend for bandwidth that may disrupt indirect routing or complicate switch
682
+ reconfiguration. While the direct (single-hop) bandwidth between cases (A) and (B) has a large difference, case (A)
683
+ always provides that direct bandwidth between MCMs whereas a spatial or wave-selective switch requires a scheduler
684
+ and leaves the majority of input–output combinations unconnected at any one time, thus also has to use indirect routing
685
+ to compensate.
686
+ Based on system profiling data of a production open-science HPC system Michelogiannakis et al. [2022], the 125
687
+ Gbps direct bandwidth between MCMs in test case (A) suffices over 99.5% of the time between CPUs and main
688
+ memory (DDR4) and virtually all the time between memory and NICs. In addition, the bandwidth of a single AWGR
689
+ wavelength of 25 Gbps suffices 97% of the time between CPUs and memory as well as between memory and NICs.
690
+ This means that with a 97% probability, four of the five wavelengths between a memory and CPUs or NICs and
691
+ memory pair are available to use for indirect routing in case the direct 125 Gbps bandwidth does not suffice between
692
+ another memory–CPU or NIC–memory pair. Therefore, the probability at any one time that the direct bandwidth does
693
+ not suffice for a number of CPU–memory and NIC–memory pairs large enough such that they cannot find unused
694
+ bandwidth in other pairs to use for indirect routing is multiple orders of magnitude less than 0.1% and thus negligible.
695
+ To further reduce the probability, congested pairs can use direct paths from CPUs to CPUs that communicate minimally
696
+ and NICs to other NICs that do not communicate at all Michelogiannakis et al. [2022]. Therefore, test case (A) satisfies
697
+ bandwidth between CPUs, NICs, and main memory (DDR4).
698
+ Figure 6: Average and maximum slowdown for each suite and input set size. The slowdown is for an additional 35ns of
699
+ latency between the LLC and main memory from the additional photonic components. Left: in-order pipeline compute
700
+ cores. Right: Out of order (OOO) compute cores.
701
+ For GPUs, in test case (A) with indirect routing a single GPU can use a total of 125 × 512 = 8000 GBps to access any
702
+ one HBM or more in case a GPU is allocated more than one HBMs. This well satisfies the 1555.2 GBps that NVIDIA
703
+ Ampere A100 GPUs in our model rack Per access HBMs with today and leaves 8000 − 1555.2 = 6444.8 GBps unused
704
+ per GPU. In addition, in the worst case, an MCM containing three GPUs will communicate at full bandwidth (12
705
+ NVLink links of 25 GBps per each of the three GPU equals 900 GBps) to other MCMs containing GPUs. Here, if
706
+ all GPUs in the rack acts similarly, we cannot rely on indirect routing from a GPU through an intermediate GPU to
707
+ reach a destination GPU. The direct 125 Gbps bandwidth between GPU MCMs do not suffice. Therefore, each GPU
708
+ can use the 6444.8 GBps of unused bandwidth to and from HBMs for indirect routing to well cover the 900 GBps
709
+ bandwidth that would otherwise use NVLink GPU–GPU links. This leaves 6444.8 − 900 = 5544.8 GBps per GPU that
710
+ can support direct HBM–HBM communication such as due to GPUDirect RDMA, indirect routing for other MCMs, or
711
+ simply increase available bandwidth to memory. Of note, our analysis does not use direct optical paths from GPUs to
712
+ main memory (DDR4). Future protocols may use for these paths or they can be used to provide even more indirect
713
+ routing bandwidth.
714
+ Our analysis shows that test case (A) with AWGRs more than satisfies bandwidth demands and avoids the need for a
715
+ scheduler to reconfigure spatial and wave-selective switches that would otherwise add overhead and reduce reaction
716
+ time.
717
+ 6.2
718
+ Latency Evaluation
719
+ Intra-rack resource disaggregation based on modern photonics increases the latency significantly less than full system
720
+ disaggregation. For intra-rack disaggregation we assume an additional latency between MCMs of 35 ns. That additional
721
+ latency covers 15 ns for electrical–optical–electrical conversion and 4 meters of photonic propagation at 5 ns per meter,
722
+ which covers round-trip distance of typical two-meter tall racks (Section 3.3.2). The small impact of distance to latency
723
+ 12
724
+
725
+ Average
726
+ Maximum
727
+ 100
728
+ 1
729
+ (%)
730
+ 75
731
+ umopmos
732
+ 50
733
+ Percentage
734
+ 25
735
+ 0
736
+ Parsec
737
+ Parsec
738
+ Parsec
739
+ NAS A
740
+ NAS B
741
+ NAS C
742
+ Rodinia
743
+ small
744
+ medium
745
+ largeAverage
746
+ Maximum
747
+ 125
748
+ 100
749
+ (%)
750
+ slowdown
751
+ 75
752
+ Percentage s
753
+ 50
754
+ 25
755
+ 0
756
+ Parsec
757
+ Parsec
758
+ Parsec
759
+ NAS A
760
+ NAS B
761
+ NAS C
762
+ Rodinia
763
+ small
764
+ medium
765
+ largearXiv Template
766
+ A PREPRINT
767
+ with photonics practically makes MCMs in a rack equi-distant, thus mitigating a traditional queuing delay versus
768
+ locality tradeoff in job scheduling Jeon et al. [2019]. Indirect routing would increase latency by a few extra ns, but the
769
+ probability of routing indirectly is low. Because 35 ns is orders of magnitude lower than system-wide network latency,
770
+ we do not consider the effect of the additional 35 ns to inter-rack communication through NICs.
771
+ 6.2.1
772
+ CPU Evaluation
773
+ We experimentally quantify the impact to application performance with in-order pipeline and out-of-order (OOO)
774
+ compute cores. In-order cores provide insight of the impact of memory latency when the compute core does not mask
775
+ latency, whereas OOO cores are representative of modern cores. We use full system simulation in Gem5 Binkert et al.
776
+ [2011] of x86 compute cores running an Ubuntu 18.4 guest OS. We configure the cache hierarchy to match the CPUs
777
+ of our model HPC rack Per. We calculate the slowdown of application execution time when we add 35 ns of latency
778
+ between the LLC and main memory, compared to a baseline system with no additional latency to memory. Latency is
779
+ the only potential source of application slowdown since our architecture satisfies the full escape bandwidth for each
780
+ chip.
781
+ We evaluate the impact in three benchmark suites: PARSEC 3.1 Bienia et al. [2008], NAS parallel benchmarks
782
+ 3.4.1 Bailey et al. [1992], and Rodinia Che et al. [2009]. For PARSEC we evaluate small, medium, and large input sets.
783
+ For NAS, we evaluate input sizes “A”, “B”, and “C”. For Rodinia we use the single default input set. These benchmark
784
+ suites have been widely used and contain a large variety of computation kernels that are representative of key HPC
785
+ applications such as stencils, graph processing, linear algebra, computational mathematics, grid, sorting, and many
786
+ others that have been observed to be important workloads in NERSC’s systems ?. Overall, we use 58 benchmarks to
787
+ provide a wide representation. We use a single compute core to better focus on the effect of the additional latency to
788
+ memory.
789
+ Figure 6 shows slowdown percentages for benchmarks across our three suites for an in-order core on the left and OOO
790
+ core on the right. As shown, NAS benchmarks are negligibly affected by the increased latency. Rodinia benchmarks
791
+ have an average slowdown of 15% with in-order cores and 13% for OOO cores. However, a single benchmark (NW)
792
+ has a slowdown of approximately 76% for in-order cores. The largest slowdown for the rest of Rodinia benchmarks
793
+ across in-order and OOO cores is 12%. Finally, PARSEC benchmarks are impacted the most, but the average slowdown
794
+ remains below 25% except for large inputs using OOO cores. OOO cores typically tolerate memory access latency
795
+ better, but they also produce more memory accesses per unit of time compared to in-order cores.
796
+ Figure 7 shows slowdown for individual PARSEC benchmarks for large inputs and in-order cores. As shown, only
797
+ three benchmarks exceed a 25% slowdown, while eight benchmarks have a slowdown of no more approximately 3.5%.
798
+ Therefore, our experiments show that while some benchmarks (three in PARSEC and one in Rodinia) experience
799
+ important slowdowns, the majority of benchmarks are impacted minimally even without mitigation strategies. This is
800
+ the case with all of the NAS benchmarks we used, eight PARSEC, and all but one Rodinia benchmarks. For benchmarks
801
+ that are more affected, there is a range of hardware and software techniques Mutlu et al. [2006], Parcerisa and Gonzalez
802
+ [2001], Mowry et al. [1998], Nekkalapu et al. [2008] to increase memory tolerance that we can apply to further reduce
803
+ application slowdown.
804
+ 6.2.2
805
+ Recovering Performance
806
+ To gauge the effectiveness of strategies to recover application performance, we test the impact of the following remedies
807
+ applied one at a time: (i) 256 miss status handling registers (MSHRs) in the LLC, (ii) doubling the LLC size with the
808
+ default number of 16 MSHRs, and (iii) default LLC configuration but a strided prefetcher with a larger stride than the
809
+ default four. Figure 8 shows the slowdown percentage that we were able to recover for PARSEC benchmarks through
810
+ these three techniques at the best, average, and worst case. This figure is the only one that includes these remedies in
811
+ this results of this section. As shown, about a 20% performance loss for small and large inputs is recovered by average.
812
+ The most effective remedy is doubling the LLC size. The reason for the smaller speedup for medium is that due the
813
+ particular LLC size, memory access patterns, and input sizes in PARSEC, medium experienced a smaller benefit from a
814
+ larger LLC. These findings motivate future work to mitigate the latency impact of the disaggregation hardware, similar
815
+ to mitigating the increased latency to access emerging memory technologies Mittal and Vetter [2016].
816
+ 6.2.3
817
+ Sensitivity to Latency
818
+ To show the sensitivity of application performance to the amount of additional latency, Figure 9 shows application
819
+ slowdown for 25 ns, 30 ns, and 35 ns for in-order cores (OOO cores show comparable trends). As shown, reducing
820
+ the additional latency to 25 ns from 35 ns reduces application slowdown by as much as half. This motivates latency
821
+ improvements in photonic components or shorter rack distances.
822
+ 13
823
+
824
+ arXiv Template
825
+ A PREPRINT
826
+ Figure 7: Results for individual PARSEC benchmarks with large inputs.
827
+ Figure 8: Percentage of slowdown that we can recover with LLC modifications.
828
+ 14
829
+
830
+ Parsec slowdown: Large inputs. Single in-order core
831
+ 100
832
+ 75
833
+ (%) easad
834
+ 50
835
+ 25Best case
836
+ Average
837
+ Worst case
838
+ (%)
839
+ 50
840
+ recovered
841
+ 40
842
+ 30
843
+ slowdown
844
+ 20
845
+ Percentage
846
+ 10
847
+ 0
848
+ Parsec small
849
+ Parsec medium
850
+ Parsec largearXiv Template
851
+ A PREPRINT
852
+ Figure 9: Percentage slowdown for 25ns, 30ns, and 35ns of additional LLC–memory latency for in order cores.
853
+ The overall average slowdown across all benchmarks is approximately 13% for both in-order cores and OOO cores
854
+ without architectural remedies, for large PARSEC inputs, and “B” size NAS inputs. This considerably less than
855
+ slowdowns quoted in past work for full-system disaggregation, furthering the case for intra-rack disaggregation.
856
+ 6.2.4
857
+ GPU Evaluation
858
+ To evaluate the impact of the additional latency between GPUs and HBMs or DDR4 main memory, we extend the
859
+ publicly available version of PPT-GPU Arafa et al. [2021] toolkit to account for the additional latency between the
860
+ main memory of the GPU and the LLC. In our evaluation, we modeled one NVIDIA A100 GPU Choquette and Gandhi
861
+ [2020] running a total of 27 applications that have a total of 2133 kernels from different benchmark suites. We run 13
862
+ applications from Rodinia Che et al. [2009] and 10 applications from Polybench Grauer-Gray et al. [2012]. Polybench
863
+ applications are linear algebra applications that stress the GPU cache and main memory. Furthermore, we run AlexNet,
864
+ CifarNet, GRU, and LSTM from the Tango deep network Karki et al. [2019] benchmark suite. We use the default input
865
+ sizes and configuration that came with the benchmarks, detailed in Arafa et al. [2021]. We run applications using the
866
+ “SASS” model, where we extract memory and instruction traces for each application.
867
+ Figure 10 shows the effect of different latencies on the performance of our GPU benchmarks. We compare performance
868
+ in terms of the total predicted cycles. As shown, the highest average slowdown is 24% for Polybench. The overall
869
+ average slowdowns across the 27 applications is only 8%, 10%, and 12% for the 25 ns, 30 ns, and 35 ns additional
870
+ latency, respectively. For these benchmarks, doubling the LLC size recovers an average of 8% of the performance loss.
871
+ 6.2.5
872
+ CPU–GPU Comparison
873
+ We illustrate the difference in memory latency tolerance of in-order CPUs, OOO CPUs, and GPUs in Figure 11 for the
874
+ intersection of Rodinia benchmarks that correctly ran on both CPU and GPU with their default input sets. As shown,
875
+ GPUs tolerate the additional 35 ns latency significantly better with a maximum slowdown of 3.3%. This is promising
876
+ for resource disaggregation given the steady growth of GPU presence in HPC systems.
877
+ 6.3
878
+ Power Overhead
879
+ We calculate the per-rack power overhead of our photonic solution for 350 MCMs with 2048 escape wavelengths from
880
+ each MCM and 25 Gbps per wavelength. If we use a DFB laser array demonstrated in Rahimi et al. [2022] with a 11%
881
+ wall plug efficiency (WPE) at 10 dDm, a total of 256 × 256 such lasers consumes 64.5 kW. For the components and
882
+ distances in our study, the required optical power per wavelength is 10 dBm. Furthermore, 350×2048 of the modulators
883
+ 15
884
+
885
+ 25ns
886
+ 30ns
887
+ 35ns
888
+ 100
889
+ 75
890
+ 50
891
+ 25
892
+ 0
893
+ Parsec
894
+ par
895
+ ROarXiv Template
896
+ A PREPRINT
897
+ Figure 10: Percentage slowdown for 25ns, 30ns, and 35ns of additional LLC–memory latency for different GPU
898
+ benchmark suites.
899
+ Figure 11: Percentage slowdown for CPU and GPU Rodinia benchmarks.
900
+ 16
901
+
902
+ 25ns
903
+ 30ns
904
+ 35ns
905
+ 80.00%
906
+ %
907
+ 60.00%
908
+ S
909
+ 40.00%
910
+ 20.00%
911
+ 0.00%
912
+ Rodinia-avg Rodinia-max PBench-avg PBench-max Tango-avg
913
+ Tango-max35ns in-order CPU
914
+ 35nsO0OCPU
915
+ 35ns GPU
916
+ 80
917
+ (%)
918
+ 60
919
+ slowdown (
920
+ 40
921
+ ercentage
922
+ 20
923
+ ParXiv Template
924
+ A PREPRINT
925
+ and receivers of Sun et al. [2020] that consume 0.8 and 2.12 pJ/bit at 25 Gbps respectively result in a total additional
926
+ power of 52.5 kW. Finally, the switches of Table 2 consume no more than 1 kW at the worst case. In summary, the total
927
+ power overhead taking into account parallel switches is no more than 150 kW. Our analysis assumes the components
928
+ are constantly on. Considering that the maximum power consumption of a single A100 GPU is a few hundreds of Ws
929
+ and our modelled rack contains 512 such GPUs, the power overhead for our photonic solution is negligible.
930
+ 6.4
931
+ Comparison With Electronic Switches
932
+ Electronic SERDES signalling rate per wire is only 112 Gbps for a short reach. Also, typical CXL or PCIe signaling
933
+ rates top out at 35 GHz/wire. In fact, as SERDES rates increase, the distance that those signals can reach reduces down
934
+ to even a few millimeters due to the resistance and capacitance of copper wires. Photonics break the reach limitations
935
+ of copper and with co-packaging can achieve 4 Tbps per mm of shoreline on the chip die.
936
+ Focusing on electronic switches, Rosetta De Sensi et al. [2020b] and Infiniband Katebzadeh et al. [2020] have a
937
+ measured per hop latency of no less than approximately 200 ns. Emerging PCIe Gen5 switches add just 10 ns per
938
+ hop Vasa et al. [2020], but only support 100 lanes per switch. To fully connect our disaggregated rack, we consider a
939
+ two-level tree network with four hops (the top level is composed of an internal two-hop subnetwork). These four hops
940
+ will be in addition to the 35 ns we previously evaluated for FEC and propagation (propagation delay is comparable
941
+ between copper and photonic for rack distances), since our photonic solution uses switches with negligible traversal
942
+ latency. Therefore, the additional latency for disaggregation in the PCIe case becomes 85 ns compared to 35 ns for
943
+ our photonic architecture. Finally, we also consider the latency through one hop of an Anton 3 network, which is
944
+ approximately 90 ns by average Shim et al. [2022], though scaling up to match our rack size would require multiple
945
+ hops. These latencies represent the best case for electronic packet switches because scheduler decisions or congestion
946
+ can cause higher worst-case (tail) latencies that may further penalize application performance. This assumes that we
947
+ connect only one lane per endpoint which carries 32 Gbps for PCIe Gen5 and 29 Gbps for Anton 3. This is multiple
948
+ times less than the per-chip bandwidth of photonics our photonic architecture.
949
+ Figure 12 shows the speedup of a system that implements intra-rack disaggregation with emerging photonics with an
950
+ additional 35 ns latency to and from DDR4 and HBM memory compared to a similar system that uses modern electronic
951
+ switches instead. 85 ns is the lowest case for electronic switches and corresponds to a four-hop PCIe Gen5 network or a
952
+ single-hop Anton 3 network. As shown, for CPU benchmarks if we only take into account “medium” from PARSEC to
953
+ avoid counting PARSEC benchmarks three times, the average speedup for in-order CPUs is 12.7% and the maximum
954
+ 76%. For OOO compute cores, the average is 23.9% and maximum 78.3%. For GPUs, the average and maximum are
955
+ both 61.5%. This analysis clearly shows the adverse impact of the additional latency of electronic switches and further
956
+ motivates the use of photonics for intra-rack resource disaggregation. Furthermore, the four electronic switches of this
957
+ analysis consume at least many tens of Watts of power, which is multiple orders of magnitude higher than our photonic
958
+ solution.
959
+ 6.5
960
+ Iso-Performance Comparison
961
+ Based on our performance evaluations, in order to preserve system-wide average computational throughput as our
962
+ baseline GPU-accelerated HPE/Cray EX system Per, our photonically-disaggregated system requires 13% more CPUs
963
+ and 8% more GPUs. However, intra-rack resource disaggregation allows our rack to have an average 4× fewer memory
964
+ modules and 2× fewer NICs Michelogiannakis et al. [2022]. Combining the two effects, our disaggregated rack has
965
+ 1082 total modules compared to 1920 in the baseline system, a 43% reduction. Alternatively, we can preserve all
966
+ rack resources and instead add 128 of a combination of CPUs and GPUs (with their HBMs), which is only a 7% chip
967
+ increase across the rack. Doing so doubles computational throughput.
968
+ 7
969
+ Conclusion
970
+ We have designed a resource disaggregated HPC rack that uses modern photonic links and switches to meet BER and
971
+ bandwidth requirements of HPC applications, has a negligible power impact, uses distributed indirect routing instead of
972
+ complex switch reconfiguration, and provides a 23.9% for CPUs or 61.5% for GPUs speedup compared to a similar
973
+ disaggregated rack implemented with modern electronic switches. Our architecture enables a disaggregated system to
974
+ preserve its performance but use 43% fewer overall chips.
975
+ 17
976
+
977
+ arXiv Template
978
+ A PREPRINT
979
+ Figure 12: Speedup of a system that uses emerging photonics to implement intra-rack resource disaggregation that adds
980
+ 35 ns of additional latency to and from memory compared to a similar system that uses modern electronic switches and
981
+ adds 85 ns of memory latency instead.
982
+ References
983
+ B. Liu, D. Zydek, H. Selvaraj, and L. Gewali. Accelerating high performance computing applications: Using cpus,
984
+ gpus, hybrid CPU/GPU, and fpgas. In 2012 13th International Conference on Parallel and Distributed Computing,
985
+ Applications and Technologies, pages 337–342, 2012.
986
+ The top500 HPC list, November 2018. URL https://www.top500.org/green500/lists/2018/11/.
987
+ M. Ujaldón. Hpc accelerators with 3d memory. In 2016 IEEE Intl Conference on Computational Science and
988
+ Engineering (CSE) and IEEE Intl Conference on Embedded and Ubiquitous Computing (EUC) and 15th Intl
989
+ Symposium on Distributed Computing and Applications for Business Engineering (DCABES), pages 320–328, 2016.
990
+ doi:10.1109/CSE-EUC-DCABES.2016.203.
991
+ Sparsh Mittal and Jeffrey S. Vetter. A survey of cpu-gpu heterogeneous computing techniques. ACM Comput. Surv., 47
992
+ (4), July 2015. ISSN 0360-0300. doi:10.1145/2788396.
993
+ D. Tiwari, S. Gupta, J. Rogers, D. Maxwell, P. Rech, S. Vazhkudai, D. Oliveira, D. Londo, N. DeBardeleben, P. Navaux,
994
+ L. Carro, and A. Bland. Understanding gpu errors on large-scale hpc systems and the implications for system design
995
+ and operation. In 2015 IEEE 21st International Symposium on High Performance Computer Architecture (HPCA),
996
+ pages 331–342, 2015. doi:10.1109/HPCA.2015.7056044.
997
+ Y. Gao and P. Zhang. A survey of homogeneous and heterogeneous system architectures in high performance
998
+ computing.
999
+ In 2016 IEEE International Conference on Smart Cloud (SmartCloud), pages 170–175, 2016.
1000
+ doi:10.1109/SmartCloud.2016.36.
1001
+ D. Milojicic. Accelerators for artificial intelligence and high-performance computing. Computer, 53(2):14–22, 2020.
1002
+ H. Asaadi and B. Chapman. Comparative study of deep learning framework in hpc environments. In 2017 New York
1003
+ Scientific Data Summit (NYSDS), pages 1–7, 2017. doi:10.1109/NYSDS.2017.8085040.
1004
+ O. Segal, N. Nasiri, M. Margala, and W. Vanderbauwhede. High level programming of FPGAs for HPC and data
1005
+ centric applications. In 2014 IEEE High Performance Extreme Computing Conference (HPEC), pages 1–3, 2014.
1006
+ doi:10.1109/HPEC.2014.7040979.
1007
+ 18
1008
+
1009
+ CPUtiming
1010
+ CPUOOO
1011
+ GPU
1012
+ 80
1013
+ Photonic speedup over electronic (%)
1014
+ 60
1015
+ 40
1016
+ 20
1017
+ xew
1018
+ avg
1019
+ avg
1020
+ small
1021
+ avg
1022
+ max
1023
+ max
1024
+ avg
1025
+ avg
1026
+ large
1027
+ Prsc
1028
+ Prsc
1029
+ Prsc
1030
+ PrscarXiv Template
1031
+ A PREPRINT
1032
+ Tom Hogervorst, Tong Dong Qiu, Giacomo Marchiori, Alf Birger, Markus Blatt, and Razvan Nane. Hardware
1033
+ acceleration of HPC computational flow dynamics using HBM-enabled FPGAs, 2021.
1034
+ J. Lant, J. Navaridas, M. Luján, and J. Goodacre. Toward fpga-based HPC: Advancing interconnect technologies. IEEE
1035
+ Micro, 40(1):25–34, 2020. doi:10.1109/MM.2019.2950655.
1036
+ Rob Dimond, Sebastien Racanière, and Oliver Pell. Accelerating large-scale hpc applications using fpgas. In 2011
1037
+ IEEE 20th Symposium on Computer Arithmetic, pages 191–192, 2011. doi:10.1109/ARITH.2011.34.
1038
+ G. Ramirez-Gargallo, M. Garcia-Gasulla, and F. Mantovani. Tensorflow on state-of-the-art HPC clusters: A machine
1039
+ learning use case. In 2019 19th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing
1040
+ (CCGRID), pages 526–533, 2019. doi:10.1109/CCGRID.2019.00067.
1041
+ M. G. Venkata, F. Aderholdt, and Z. Parchman. SharP: Towards programming extreme-scale systems with hierarchical
1042
+ heterogeneous memory. In 2017 46th International Conference on Parallel Processing Workshops (ICPPW), pages
1043
+ 145–154, 2017. doi:10.1109/ICPPW.2017.32.
1044
+ George Michelogiannakis, Benjamin Klenk, Brandon Cook, Min Yee Teh, Madeleine Glick, Larry Dennison, Keren
1045
+ Bergman, and John Shalf. A case for intra-rack resource disaggregation in HPC. ACM Trans. Archit. Code Optim.,
1046
+ jan 2022. ISSN 1544-3566. doi:10.1145/3514245. Just Accepted.
1047
+ G. P. Rodrigo, P. Östberg, E. Elmroth, K. Antypas, R. Gerber, and L. Ramakrishnan. Towards understanding job
1048
+ heterogeneity in hpc: A nersc case study. In 2016 16th IEEE/ACM International Symposium on Cluster, Cloud and
1049
+ Grid Computing (CCGrid), pages 521–526, 2016.
1050
+ A. Guleria, J. Lakshmi, and C. Padala.
1051
+ Quadd:
1052
+ Quantifying accelerator disaggregated datacenter effi-
1053
+ ciency. In 2019 IEEE 12th International Conference on Cloud Computing (CLOUD), pages 349–357, 2019a.
1054
+ doi:10.1109/CLOUD.2019.00064.
1055
+ I. Peng, R. Pearce, and M. Gokhale. On the memory underutilization: Exploring disaggregated memory on HPC
1056
+ systems. In 2020 IEEE 32nd International Symposium on Computer Architecture and High Performance Computing
1057
+ (SBAC-PAD), pages 183–190, 2020.
1058
+ J. Taylor. Facebook’s data center infrastructure: Open compute, disaggregated rack, and beyond. In 2015 Optical Fiber
1059
+ Communications Conference and Exhibition (OFC), pages 1–1, 2015.
1060
+ Huaicheng Li, Daniel S. Berger, Stanko Novakovic, Lisa Hsu, Dan Ernst, Pantea Zardoshti, Monish Shah, Ishwar
1061
+ Agarwal, Mark D. Hill, Marcus Fontoura, and Ricardo Bianchini. First-generation memory disaggregation for cloud
1062
+ platforms. arXiv, 2022. doi:10.48550/ARXIV.2203.00241.
1063
+ K. Koh, K. Kim, S. Jeon, and J. Huh. Disaggregated cloud memory with elastic block management. IEEE Transactions
1064
+ on Computers, 68(1):39–52, 2019.
1065
+ A. D. Papaioannou, R. Nejabati, and D. Simeonidou. The benefits of a disaggregated data centre: A resource
1066
+ allocation approach.
1067
+ In 2016 IEEE Global Communications Conference (GLOBECOM), pages 1–7, 2016.
1068
+ doi:10.1109/GLOCOM.2016.7842314.
1069
+ J. Gonzalez, A. Gazman, M. Hattink, M. G. Palma, M. Bahadori, R. Rubio-Noriega, L. Orosa, M. Glick, O. Mutlu,
1070
+ K. Bergman, and R. Azevedo. Optically connected memory for disaggregated data centers. In 2020 IEEE 32nd
1071
+ International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD), pages 43–50,
1072
+ 2020. doi:10.1109/SBAC-PAD49847.2020.00017.
1073
+ Qixiang Cheng, Keren Bergman, Yishen Huang, Hao Yang, Meisam Bahadori, Nathan Abrams, Xiang Meng,
1074
+ Madeleine Glick, Yang Liu, and Michael Hochberg. Silicon photonic switch topologies and routing strategies
1075
+ for disaggregated data centers. IEEE Journal of Selected Topics in Quantum Electronics, PP:1–1, 12 2019a.
1076
+ doi:10.1109/JSTQE.2019.2960950.
1077
+ A. Guleria, J. Lakshmi, and C. Padala. Emf: Disaggregated gpus in datacenters for efficiency, modularity and flexibility.
1078
+ In 2019 IEEE International Conference on Cloud Computing in Emerging Markets (CCEM), pages 1–8, 2019b.
1079
+ doi:10.1109/CCEM48484.2019.000-5.
1080
+ NERSC’s Perlmutter configuration. URL https://docs.nersc.gov/systems/perlmutter/.
1081
+ Torben Kling Petersen and John Bent.
1082
+ Hybrid flash arrays for hpc storage systems:
1083
+ An alternative to
1084
+ burst buffers.
1085
+ In 2017 IEEE High Performance Extreme Computing Conference (HPEC), pages 1–7, 2017.
1086
+ doi:10.1109/HPEC.2017.8091092.
1087
+ M. Glick, N. C. Abrams, Q. Cheng, M. Y. Teh, Y. H. Hung, O. Jimenez, S. Liu, Y. Okawachi, X. Meng, L. Johansson,
1088
+ M. Ghobadi, L. Dennison, G. Michelogiannakis, J. Shalf, A. Liu, J. Bowers, A. Gaeta, M. Lipson, and K. Bergman.
1089
+ PINE: Photonic integrated networked energy efficient datacenters (enlitened program) [invited]. IEEE/OSA Journal
1090
+ of Optical Communications and Networking, 12(12):443–456, 2020. doi:10.1364/JOCN.402788.
1091
+ 19
1092
+
1093
+ arXiv Template
1094
+ A PREPRINT
1095
+ Xiaotao Guo, Xuwei Xue, Bitao Pan, Fulong Yan, Georgios Exarchakos, and Nicola Calabretta. Fosda: A hybrid
1096
+ disaggregated hpc architecture based on distributed nanoseconds optical switches. In Future computing, 2021. ISBN
1097
+ 978-1-61208-846-4.
1098
+ R. Lin, Y. Cheng, M. D. Andrade, L. Wosinska, and J. Chen. Disaggregated data centers: Challenges and trade-offs.
1099
+ IEEE Communications Magazine, 58(2), 2020.
1100
+ Aaron Call, Jordà Polo, David Carrera, Francesc Guim, and Sujoy Sen. Disaggregating non-volatile memory for
1101
+ throughput-oriented genomics workloads. CoRR, abs/2007.02813, 2020.
1102
+ Gengchen Liu, Roberto Proietti, Marjan Fariborz, Pouya Fotouhi, Xian Xiao, and S. J. Ben Yoo. Architecture
1103
+ and performance studies of 3d-hyper-flex-lion for reconfigurable all-to-all hpc networks. In Proceedings of the
1104
+ International Conference for High Performance Computing, Networking, Storage and Analysis, SC ’20. IEEE Press,
1105
+ 2020. ISBN 9781728199986.
1106
+ Y. Zhang, X. Xiao, K. Zhang, S. Li, A. Samanta, Y. Zhang, K. Shang, R. Proietti, K. Okamoto, and S. J. B. Yoo.
1107
+ Foundry-enabled scalable all-to-all optical interconnects using silicon nitride arrayed waveguide router interposers
1108
+ and silicon photonic transceivers. IEEE Journal of Selected Topics in Quantum Electronics, 25(5):1–9, 2019.
1109
+ doi:10.1109/JSTQE.2019.2910415.
1110
+ Tae Joon Seok, Kyungmok Kwon, Johannes Henriksson, Jianheng Luo, and Ming C. Wu. Wafer-scale silicon
1111
+ photonic switches beyond die size limit. Optica, 6(4):490–494, Apr 2019a. doi:10.1364/OPTICA.6.000490. URL
1112
+ http://opg.optica.org/optica/abstract.cfm?URI=optica-6-4-490.
1113
+ Yunhong Ding, Valerija Kamchevska, Kjeld Dalgaard, Feihong Ye, Rameez Asif, Simon Gross, Michael Withford,
1114
+ Michael Galili, Toshio Morioka, and L.K. Oxenlowe. Reconfigurable sdm switching using novel silicon photonic
1115
+ integrated circuit. Scientific Reports, 6, 08 2016. doi:10.1038/srep39058.
1116
+ Yishen Huang, Qixiang Cheng, Anthony Rizzo, and Keren Bergman. Push—pull microring-assisted space-and-
1117
+ wavelength selective switch. Opt. Lett., 45(10):2696–2699, May 2020. doi:10.1364/OL.392482.
1118
+ Kevin Lim, Jichuan Chang, Trevor Mudge, Parthasarathy Ranganathan, Steven K. Reinhardt, and Thomas F. Wenisch.
1119
+ Disaggregated memory for expansion and sharing in blade servers. SIGARCH Comput. Archit. News, 37(3):267–278,
1120
+ jun 2009. ISSN 0163-5964. doi:10.1145/1555815.1555789.
1121
+ S. Van Doren. Abstract - hoti 2019: Compute express link. In 2019 IEEE Symposium on High-Performance Interconnects
1122
+ (HOTI), pages 18–18, 2019. doi:10.1109/HOTI.2019.00017.
1123
+ G. Zervas, H. Yuan, A. Saljoghei, Q. Chen, and V. Mishra. Optically disaggregated data centers with minimal remote
1124
+ memory latency: Technologies, architectures, and resource allocation [invited]. IEEE/OSA Journal of Optical
1125
+ Communications and Networking, 10(2):A270–A285, 2018. doi:10.1364/JOCN.10.00A270.
1126
+ Peter X. Gao, Akshay Narayan, Sagar Karandikar, Joao Carreira, Sangjin Han, Rachit Agarwal, Sylvia Ratnasamy, and
1127
+ Scott Shenker. Network requirements for resource disaggregation. In Proceedings of the 12th USENIX Conference
1128
+ on Operating Systems Design and Implementation, OSDI’16. USENIX Association, 2016. ISBN 9781931971331.
1129
+ Bulent Abali, Richard J. Eickemeyer, Hubertus Franke, Chung-Sheng Li, and Marc A. Taubenblatt. Disaggregated and
1130
+ optically interconnected memory: when will it be cost effective?, 2015.
1131
+ Sangjin Han, Norbert Egi, Aurojit Panda, Sylvia Ratnasamy, Guangyu Shi, and Scott Shenker. Network support for
1132
+ resource disaggregation in next-generation datacenters. In Proceedings of the Twelfth ACM Workshop on Hot Topics
1133
+ in Networks, HotNets-XII, 2013. ISBN 9781450325967. doi:10.1145/2535771.2535778.
1134
+ Y. Cheng, M. De Andrade, L. Wosinska, and J. Chen. Resource disaggregation versus integrated servers in data centers:
1135
+ Impact of internal transmission capacity limitation. In 2018 European Conference on Optical Communication
1136
+ (ECOC), pages 1–3, 2018. doi:10.1109/ECOC.2018.8535214.
1137
+ Yuping Fan, Zhiling Lan, Paul Rich, William E. Allcock, Michael E. Papka, Brian Austin, and David Paul. Scheduling
1138
+ beyond CPUs for HPC. Proceedings of the 28th International Symposium on High-Performance Parallel and
1139
+ Distributed Computing, Jun 2019.
1140
+ Giovanni Agosta, William Fornaciari, Giuseppe Massari, Anna Pupykina, Federico Reghenzani, and Michele Zanella.
1141
+ Managing heterogeneous resources in HPC systems. PARMA-DITAM ’18, page 7–12, New York, NY, USA, 2018.
1142
+ Association for Computing Machinery. ISBN 9781450364447.
1143
+ Marcelo Amaral, Jorda Polo, David Carrera, Nelson Gonzalez, Chih-Chieh Yang, Alessandro Morari, Bruce D’Amora,
1144
+ Alaa Youssef, and Malgorzata Steinder. DRMaestro: orchestrating disaggregated resources on virtualized data-centers.
1145
+ Journal of Cloud Computing, 10, 03 2021. doi:10.1186/s13677-021-00238-6.
1146
+ G. Domeniconi, E. Lee, Vanamala Venkataswamy, and Swaroopa Dola. CuSH: Cognitive scheduler for heterogeneous
1147
+ high performance computing system. 2019.
1148
+ 20
1149
+
1150
+ arXiv Template
1151
+ A PREPRINT
1152
+ Arthur B. Maccabe. Operating and runtime systems challenges for hpc systems. In Proceedings of the 7th Inter-
1153
+ national Workshop on Runtime and Operating Systems for Supercomputers ROSS 2017, ROSS ’17, 2017. ISBN
1154
+ 9781450350860. doi:10.1145/3095770.3095771.
1155
+ W. Hwu, L. Chang, H. Kim, A. Dakkak, and I. El Hajj.
1156
+ Transitioning hpc software to exascale heteroge-
1157
+ neous computing. In 2015 Computational Electromagnetics International Workshop (CEM), pages 1–2, 2015.
1158
+ doi:10.1109/CEM.2015.7237412.
1159
+ Yizhou Shan, Yutong Huang, Yilun Chen, and Yiying Zhang. Legoos: A disseminated, distributed os for hardware
1160
+ resource disaggregation.
1161
+ In Proceedings of the 13th USENIX Conference on Operating Systems Design and
1162
+ Implementation, OSDI’18, page 69–87. USENIX Association, 2018. ISBN 9781931971478.
1163
+ Roberto Gioiosa, Ozcelik Burcu Mutlu, Seyong Lee, S. Jeffrey Vetter, Giulio Picierro, and Marco Cesati. The minos
1164
+ computing library - efficient parallel programming for extremely heterogeneous systems. GPGPU@PPoPP, pages
1165
+ 1–10, 2020.
1166
+ H. Khaleghzadeh, R. R. Manumachu, and A. Lastovetsky. A hierarchical data-partitioning algorithm for performance
1167
+ optimization of data-parallel applications on heterogeneous multi-accelerator numa nodes. IEEE Access, 8:7861–
1168
+ 7876, 2020.
1169
+ Zaeem Hussain. Heterogeneity aware fault tolerance for extreme scale computing. August 2020. URL http:
1170
+ //d-scholarship.pitt.edu/39456/.
1171
+ P. Jamieson, A. Sanaullah, and M. Herbordt. Benchmarking heterogeneous hpc systems including reconfigurable fabrics:
1172
+ Community aspirations for ideal comparisons. In 2018 IEEE High Performance extreme Computing Conference
1173
+ (HPEC), pages 1–6, 2018.
1174
+ K. Tang, D. Tiwari, S. Gupta, S. S. Vazhkudai, and X. He. Effective running of end-to-end hpc workflows on emerging
1175
+ heterogeneous architectures. In 2017 IEEE International Conference on Cluster Computing (CLUSTER), pages
1176
+ 344–348, 2017. doi:10.1109/CLUSTER.2017.22.
1177
+ Alexey Lastovetsky. Heterogeneous parallel computing: from clusters of workstations to hierarchical hybrid platforms.
1178
+ Supercomputing Frontiers and Innovations, 1(3), 2015. ISSN 2313-8734.
1179
+ Jing Guo, Zihao Chang, Sa Wang, Haiyang Ding, Yihui Feng, Liang Mao, and Yungang Bao. Who limits the resource
1180
+ efficiency of my datacenter: An analysis of alibaba datacenter traces. In Proceedings of the International Symposium
1181
+ on Quality of Service, IWQoS ’19, 2019. ISBN 9781450367783. doi:10.1145/3326285.3329074.
1182
+ J. Patel, V. Jindal, I. Yen, F. Bastani, J. Xu, and P. Garraghan. Workload estimation for improving resource management
1183
+ decisions in the cloud. In 2015 IEEE Twelfth International Symposium on Autonomous Decentralized Systems, pages
1184
+ 25–32, 2015. doi:10.1109/ISADS.2015.17.
1185
+ Teng Li, Vikram K. Narayana, and Tarek A. El-Ghazawi. Efficient resource sharing through GPU virtualization on
1186
+ accelerated high performance computing systems. CoRR, abs/1511.07658, 2015.
1187
+ Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, unjie Qian, Wencong Xiao, and Fan Yang. Analysis of
1188
+ large-scale multi-tenant gpu clusters for dnn training workloads. In Proceedings of the 2019 USENIX Conference on
1189
+ Usenix Annual Technical Conference, USENIX ATC ’19, page 947–960, USA, 2019. USENIX Association. ISBN
1190
+ 9781939133038.
1191
+ Teng Li, Vikram K. Narayana, Esam El-Araby, and Tarek El-Ghazawi. Gpu resource sharing and virtualization on high
1192
+ performance computing systems. In 2011 International Conference on Parallel Processing, pages 733–742, 2011.
1193
+ doi:10.1109/ICPP.2011.88.
1194
+ N. Terzenidis, M. Moralis-Pegios, G. Mourgias-Alexandris, T. Alexoudi, K. Vyrsokinos, and N. Pleros. High-port and
1195
+ low-latency optical switches for disaggregated data centers: The hipolaos switch architecture. IEEE/OSA Journal of
1196
+ Optical Communications and Networking, 10(7):102–116, 2018. doi:10.1364/JOCN.10.00B102.
1197
+ Zeke Wang, Hongjing Huang, Jie Zhang, and Gustavo Alonso. Benchmarking high bandwidth memory on FPGAs,
1198
+ 2020.
1199
+ Justin Meza, Qiang Wu, Sanjeev Kumar, and Onur Mutlu. Revisiting memory errors in large-scale production data
1200
+ centers: Analysis and modeling of new trends from the field. In 45th Annual IEEE/IFIP International Conference
1201
+ on Dependable Systems and Networks, DSN 2015, Rio de Janeiro, Brazil, June 22-25, 2015, pages 415–426. IEEE
1202
+ Computer Society, 2015. doi:10.1109/DSN.2015.57.
1203
+ Vilas Sridharan, Nathan DeBardeleben, Sean Blanchard, Kurt B. Ferreira, Jon Stearley, John Shalf, and Sudhanva
1204
+ Gurumurthi. Memory errors in modern systems: The good, the bad, and the ugly. In Özcan Özturk, Kemal Ebcioglu,
1205
+ and Sandhya Dwarkadas, editors, Proceedings of the Twentieth International Conference on Architectural Support
1206
+ for Programming Languages and Operating Systems, ASPLOS ’15, Istanbul, Turkey, March 14-18, 2015, pages
1207
+ 297–310. ACM, 2015. doi:10.1145/2694344.2694348.
1208
+ 21
1209
+
1210
+ arXiv Template
1211
+ A PREPRINT
1212
+ S. Luyi, F. Jinyi, and Y. Xiaohua. Forward error correction. In 2012 Fourth International Conference on Computational
1213
+ and Information Sciences, pages 37–40, 2012. doi:10.1109/ICCIS.2012.158.
1214
+ Bok Young Kim, Yoshitomo Okawachi, Jae K. Jang, Mengjie Yu, Xingchen Ji, Yun Zhao, Chaitanya Joshi, Michal
1215
+ Lipson, and Alexander L. Gaeta. Turn-key, high-efficiency kerr comb source. Opt. Lett., 44(18):4475–4478, Sep
1216
+ 2019a. doi:10.1364/OL.44.004475.
1217
+ Shaoming Chen, Samuel Irving, Lu Peng, Yue Hu, Ying Zhang, and Ashok Srivastava. Using switchable pins to
1218
+ increase off-chip bandwidth in chip-multiprocessors. IEEE Trans. Parallel Distrib. Syst., 28(1):274–289, jan 2017.
1219
+ ISSN 1045-9219. doi:10.1109/TPDS.2016.2546246.
1220
+ Norman P. Jouppi, Cliff Young, Nishant Patil, David Patterson, Gaurav Agrawal, Raminder Bajwa, Sarah Bates, Suresh
1221
+ Bhatia, Nan Boden, Al Borchers, Rick Boyle, Pierre-luc Cantin, Clifford Chao, Chris Clark, Jeremy Coriell, Mike
1222
+ Daley, Matt Dau, Jeffrey Dean, Ben Gelb, Tara Vazir Ghaemmaghami, Rajendra Gottipati, William Gulland, Robert
1223
+ Hagmann, C. Richard Ho, Doug Hogberg, John Hu, Robert Hundt, Dan Hurt, Julian Ibarz, Aaron Jaffey, Alek
1224
+ Jaworski, Alexander Kaplan, Harshit Khaitan, Daniel Killebrew, Andy Koch, Naveen Kumar, Steve Lacy, James
1225
+ Laudon, James Law, Diemthu Le, Chris Leary, Zhuyuan Liu, Kyle Lucke, Alan Lundin, Gordon MacKean, Adriana
1226
+ Maggiore, Maire Mahony, Kieran Miller, Rahul Nagarajan, Ravi Narayanaswami, Ray Ni, Kathy Nix, Thomas Norrie,
1227
+ Mark Omernick, Narayana Penukonda, Andy Phelps, Jonathan Ross, Matt Ross, Amir Salek, Emad Samadiani,
1228
+ Chris Severn, Gregory Sizikov, Matthew Snelham, Jed Souter, Dan Steinberg, Andy Swing, Mercedes Tan, Gregory
1229
+ Thorson, Bo Tian, Horia Toma, Erick Tuttle, Vijay Vasudevan, Richard Walter, Walter Wang, Eric Wilcox, and
1230
+ Doe Hyun Yoon. In-datacenter performance analysis of a tensor processing unit. In Proceedings of the 44th Annual
1231
+ International Symposium on Computer Architecture, ISCA ’17, pages 1–12, New York, NY, USA, 2017. ACM. ISBN
1232
+ 978-1-4503-4892-8. doi:10.1145/3079856.3080246.
1233
+ M. Wade. TeraPHY: A chiplet technology for low-power, high-bandwidth in-package optical i/o. In 2019 IEEE Hot
1234
+ Chips 31 Symposium (HCS), pages i–xlviii, 2019. doi:10.1109/HOTCHIPS.2019.8875658.
1235
+ Keren Bergman, John Shalf, George Michelogiannakis, Sebastien Rumley, Larry Dennison, and Monia Ghobadi. Pine:
1236
+ An energy efficient flexibly interconnected photonic data center architecture for extreme scalability. In 2018 IEEE
1237
+ Optical Interconnects Conference (OI), pages 25–26, 2018. doi:10.1109/OIC.2018.8422036.
1238
+ Pavlos Maniotis, Laurent Schares, Marc A. Taubenblatt, and Daniel M. Kuchta. Co-packaged optics for HPC and data
1239
+ center networks. In Henning Schröder and Ray T. Chen, editors, Optical Interconnects XXI, volume 11692, pages 17
1240
+ – 22. International Society for Optics and Photonics, SPIE, 2021. doi:10.1117/12.2579066.
1241
+ Cyriel Minkenberg, Rajagopal Krishnaswamy, Aaron Zilkie, and David Nelson. Co-packaged datacenter optics:
1242
+ Opportunities and challenges. IET Optoelectronics, 15(2):77–91, 2021. doi:https://doi.org/10.1049/ote2.12020.
1243
+ A. Sutono, A. Pham, J. Laskar, and W.R. Smith. Investigations of multi-layer ceramic-based mcm technology. In IEEE
1244
+ 7th Topical Meeting on Electrical Performance of Electronic Packaging (Cat. No.98TH8370), pages 83–86, 1998.
1245
+ doi:10.1109/EPEP.1998.733854.
1246
+ Nathan C. Abrams, Qixiang Cheng, Madeleine Glick, Moises Jezzini, Padraic Morrissey, Peter O’Brien, and Keren
1247
+ Bergman. Silicon photonic 2.5d multi-chip module transceiver for high-performance data centers. Journal of
1248
+ Lightwave Technology, 38(13):3346–3357, 2020. doi:10.1109/JLT.2020.2967235.
1249
+ Michael Dittrich, Andy Heinig, Fabian Hopsch, and Robert Trieb. Heterogeneous interposer based integration of chips
1250
+ with copper pillars and c4 balls to achieve high speed interfaces for adc application. In 2017 IEEE 67th Electronic
1251
+ Components and Technology Conference (ECTC), pages 643–648, 2017. doi:10.1109/ECTC.2017.221.
1252
+ S. Fathololoumi, D. Hui, S. Jadhav, J. Chen, K. Nguyen, M. N. Sakib, Z. Li, H. Mahalingam, S. Amiralizadeh,
1253
+ N. N. Tang, H. Potluri, M. Montazeri, H. Frish, R. A. Defrees, C. Seibert, A. Krichevsky, J. K. Doylend, J. Heck,
1254
+ R. Venables, A. Dahal, A. Awujoola, A. Vardapetyan, G. Kaur, M. Cen, V. Kulkarni, S. S. Islam, R. L. Spreitzer,
1255
+ S. Garag, A. C. Alduino, R. Chiou, L. Kamyab, S. Gupta, B. Xie, R. S. Appleton, S. Hollingsworth, S. McCargar,
1256
+ Y. Akulova, K. M. Brown, R. Jones, D. Zhu, T. Liljeberg, and L. Liao. 1.6 tbps silicon photonics integrated circuit
1257
+ and 800 gbps photonic engine for switch co-packaging demonstration. Journal of Lightwave Technology, 39(4):
1258
+ 1155–1161, 2021. doi:10.1109/JLT.2020.3039218.
1259
+ Erik Agrell, Magnus Karlsson, A R Chraplyvy, David J Richardson, Peter M Krummrich, Peter Winzer, Kim Roberts,
1260
+ Johannes Karl Fischer, Seb J Savory, Benjamin J Eggleton, Marco Secondini, Frank R Kschischang, Andrew Lord,
1261
+ Josep Prat, Ioannis Tomkos, John E Bowers, Sudha Srinivasan, Mae Brandt-Pearce, and Nicolas Gisin. Roadmap of
1262
+ optical communications. Journal of Optics, 18(6):063002, may 2016. doi:10.1088/2040-8978/18/6/063002.
1263
+ J. Wei, Q. Cheng, R. V. Penty, I. H. White, and D. G. Cunningham. 400 gigabit ethernet using advanced modulation
1264
+ formats: Performance, complexity, and power dissipation. IEEE Communications Magazine, 53(2):182–189, 2015.
1265
+ doi:10.1109/MCOM.2015.7045407.
1266
+ 22
1267
+
1268
+ arXiv Template
1269
+ A PREPRINT
1270
+ Bok Young Kim, Yoshitomo Okawachi, Jae K. Jang, Mengjie Yu, Xingchen Ji, Yun Zhao, Chaitanya Joshi, Michal
1271
+ Lipson, and Alexander L. Gaeta. Turn-key, high-efficiency kerr comb source. Opt. Lett., 44(18):4475–4478, Sep
1272
+ 2019b. doi:10.1364/OL.44.004475.
1273
+ D. D. Sharma. Pci express® 6.0 specification at 64.0 gt/s with pam-4 signaling: a low latency, high bandwidth, high
1274
+ reliability and cost-effective interconnect. In 2020 IEEE Symposium on High-Performance Interconnects (HOTI),
1275
+ pages 1–8, 2020. doi:10.1109/HOTI51249.2020.00016.
1276
+ Kazuhiro Ikeda, Keijiro Suzuki, Ryotaro Konoike, Shu Namiki, and Hitoshi Kawashima. Large-scale silicon pho-
1277
+ tonics switch based on 45-nm cmos technology. Optics Communications, 466:125677, 2020. ISSN 0030-4018.
1278
+ doi:https://doi.org/10.1016/j.optcom.2020.125677.
1279
+ Tae Joon Seok, Kyungmok Kwon, Johannes Henriksson, Jianheng Luo, and Ming C. Wu. 240×240 wafer-scale silicon
1280
+ photonic switches. In 2019 Optical Fiber Communications Conference and Exhibition (OFC), pages 1–3, 2019b.
1281
+ Akhilesh S. P. Khope, Andrew M. Netherton, Takako Hirokawa, Nicolas Volet, Eric J. Stanton, Clint Schow, Roger
1282
+ Helkey, Adel A. M. Saleh, John E. Bowers, and Rod C. Alferness. Elastic wdm optoelectronic crossbar switch with
1283
+ on-chip wavelength control. In Advanced Photonics 2017 (IPR, NOMA, Sensors, Networks, SPPCom, PS), page
1284
+ PTh1D.3. Optica Publishing Group, 2017. doi:10.1364/PS.2017.PTh1D.3.
1285
+ Qixiang Cheng, Meisam Bahadori, Yu-Han Hung, Yishen Huang, Nathan Abrams, and Keren Bergman. Scalable
1286
+ microring-based silicon clos switch fabric with switch-and-select stages. IEEE Journal of Selected Topics in Quantum
1287
+ Electronics, 25(5):1–11, 2019b. doi:10.1109/JSTQE.2019.2911421.
1288
+ Ken-ichi Sato. Realization and Application of Large-Scale Fast Optical Circuit Switch for Data Center Networking.
1289
+ Journal of Lightwave Technology, 36(7):1411–1419, April 2018. ISSN 1558-2213. doi:10.1109/JLT.2018.2801308.
1290
+ S series optical circuit switch. URL https://www.calient.net/products/s-series-photonic-switch/.
1291
+ Dessislava Nikolova, David M Calhoun, Yang Liu, Sébastien Rumley, Ari Novack, Tom Baehr-Jones, Michael Hochberg,
1292
+ and Keren Bergman. Modular architecture for fully non-blocking silicon photonic switch fabric. Microsystems &
1293
+ nanoengineering, 3(1):1–9, 2017.
1294
+ Liang Yuan Dai, Yu-Han Hung, Qixiang Cheng, and Keren Bergman. Experimental demonstration of pam-4 transmission
1295
+ through microring silicon photonic clos switch fabric. In 2020 Optical Fiber Communications Conference and
1296
+ Exhibition (OFC), pages 1–3, 2020.
1297
+ R. Proietti, Y. Yin, R. Yu, C. J. Nitta, V. Akella, C. Mineo, and S. J. B. Yoo. Scalable optical interconnect architecture
1298
+ using awgr-based tonak lion switch with limited number of wavelengths. Journal of Lightwave Technology, 31(24):
1299
+ 4087–4097, 2013. doi:10.1109/JLT.2013.2285883.
1300
+ C. Lea. A scalable awgr-based optical switch. Journal of Lightwave Technology, 33(22):4612–4621, Nov 2015. ISSN
1301
+ 0733-8724. doi:10.1109/JLT.2015.2479296.
1302
+ Dhoore, Sören and Roelkens, Günther and Morthier, Geert. Fast wavelength-tunable lasers on silicon. IEEE JOURNAL
1303
+ OF SELECTED TOPICS IN QUANTUM ELECTRONICS, 25(6):1500908:1–1500908:8, 2019. ISSN 1077-260X.
1304
+ Optical circuit switch product availability. https://www.fs.com/products/66601.html.
1305
+ Ken-ichi Sato, Hiroshi Hasegawa, Tomonobu Niwa, and Toshio Watanabe. A large-scale wavelength routing optical
1306
+ switch for data center networks. IEEE Communications Magazine, 51(9):46–52, September 2013. ISSN 1558-1896.
1307
+ doi:10.1109/MCOM.2013.6588649.
1308
+ Koh Ueda, Yojiro Mori, Hiroshi Hasegawa, Hiroyuki Matsuura, Kiyo Ishii, Haruhiko Kuwatsuka, Shu Namiki, Toshio
1309
+ Watanabe, and Ken-ichi Sato. Demonstration of 1,440×1,440 fast optical circuit switch for datacenter networking.
1310
+ In 2016 21st OptoElectronics and Communications Conference (OECC) Held Jointly with 2016 International
1311
+ Conference on Photonics in Switching (PS), pages 1–3, July 2016.
1312
+ Sudeep Bhoja. Pam4 signaling for intra-data center and data center to data center connectivity (dci). In 2017 Optical
1313
+ Fiber Communications Conference and Exhibition (OFC), pages 1–54, 2017.
1314
+ D. M. Marom, P. D. Colbourne, A. D’errico, N. K. Fontaine, Y. Ikuma, R. Proietti, L. Zong, J. M. Rivas-Moscoso,
1315
+ and I. Tomkos. Survey of photonic switching architectures and technologies in support of spatially and spectrally
1316
+ flexible optical networking [invited]. IEEE/OSA Journal of Optical Communications and Networking, 9(1):1–26,
1317
+ 2017. doi:10.1364/JOCN.9.000001.
1318
+ Min Yee Teh, Yu-Han Hung, George Michelogiannakis, Shijia Yan, Madeleine Glick, John Shalf, and Keren
1319
+ Bergman.
1320
+ Tago: Rethinking routing design in high performance reconfigurable networks.
1321
+ In SC20: Inter-
1322
+ national Conference for High Performance Computing, Networking, Storage and Analysis, pages 1–16, 2020.
1323
+ doi:10.1109/SC41405.2020.00029.
1324
+ 23
1325
+
1326
+ arXiv Template
1327
+ A PREPRINT
1328
+ George Michelogiannakis, Yiwen Shen, Min Yee Teh, Xiang Meng, Benjamin Aivazi, Taylor Groves, John Shalf,
1329
+ Madeleine Glick, Manya Ghobadi, Larry Dennison, and Keren Bergman. Bandwidth steering in HPC using silicon
1330
+ nanophotonics. In Proceedings of the International Conference for High Performance Computing, Networking,
1331
+ Storage and Analysis, SC ’19, 2019. ISBN 9781450362290. doi:10.1145/3295500.3356145.
1332
+ John Shalf, Shoaib Kamil, Leonid Oliker, and David Skinner. Analyzing ultra-scale application communication
1333
+ requirements for a reconfigurable hybrid interconnect. In Proceedings of the ACM/IEEE SC2005 Conference on
1334
+ High Performance Networking and Computing, November 12-18, 2005, Seattle, WA, USA, CD-Rom, page 17, 2005.
1335
+ doi:10.1109/SC.2005.12.
1336
+ Jeffrey S. Vetter and Frank Mueller. Communication characteristics of large-scale scientific applications for contempo-
1337
+ rary cluster architectures. In 16th International Parallel and Distributed Processing Symposium (IPDPS 2002), 15-19
1338
+ April 2002, Fort Lauderdale, FL, USA, CD-ROM/Abstracts Proceedings, 2002. doi:10.1109/IPDPS.2002.1015504.
1339
+ Jens Domke, Satoshi Matsuoka, Ivan R. Ivanov, Yuki Tsushima, Tomoya Yuki, Akihiro Nomura, Shin’ichi Miura, Nie
1340
+ McDonald, Dennis L. Floyd, and Nicolas Dubé. Hyperx topology: First at-scale implementation and comparison to
1341
+ the fat-tree. In Proceedings of the International Conference for High Performance Computing, Networking, Storage
1342
+ and Analysis, SC ’19, 2019. ISBN 9781450362290. doi:10.1145/3295500.3356140.
1343
+ Sheng Ma, Zhiying Wang, Natalie Enright Jerger, Li Shen, and Nong Xiao. Novel flow control for fully adaptive
1344
+ routing in cache-coherent nocs. IEEE Transactions on Parallel and Distributed Systems, 25(9):2397–2407, 2014.
1345
+ doi:10.1109/TPDS.2013.166.
1346
+ Daniel U. Becker and William J. Dally.
1347
+ Allocator implementations for network-on-chip routers.
1348
+ In Proceed-
1349
+ ings of the Conference on High Performance Computing Networking, Storage and Analysis, pages 1–12, 2009.
1350
+ doi:10.1145/1654059.1654112.
1351
+ Nan Jiang, John Kim, and William J. Dally. Indirect adaptive routing on large scale interconnection networks. In
1352
+ Proceedings of the 36th Annual International Symposium on Computer Architecture, ISCA ’09, pages 220–231, 2009.
1353
+ ISBN 978-1-60558-526-0. doi:10.1145/1555754.1555783.
1354
+ Daniele De Sensi, Salvatore Di Girolamo, Kim H. McMahon, Duncan Roweth, and Torsten Hoefler. An in-depth analysis
1355
+ of the slingshot interconnect. In SC20: International Conference for High Performance Computing, Networking,
1356
+ Storage and Analysis, pages 1–14, 2020a. doi:10.1109/SC41405.2020.00039.
1357
+ Kaveh Hosseini, Edwin Kok, Sergey Y. Shumarayev, Chia-Pin Chiu, Arnab Sarkar, Asako Toda, Yanjing Ke, Allen
1358
+ Chan, Daniel Jeong, Mason Zhang, Sangeeta Raman, Thungoc Tran, Kumar Abhishek Singh, Pavan Bhargava,
1359
+ Chong Zhang, Haiwei Lu, Ravi Mahajan, Xiaoqian Li, Nitin Deshpande, Conor O’Keeffe, Tim Tri Hoang, Uma
1360
+ Krishnamoorthy, Chen Sun, Roy Meade, Vladimir Stojanovic, and Mark Wade. 8 tbps co-packaged FPGA and
1361
+ silicon photonics optical IO. In 2021 Optical Fiber Communications Conference and Exhibition (OFC), pages 1–3,
1362
+ 2021.
1363
+ Nathan Binkert, Bradford Beckmann, Gabriel Black, Steven K. Reinhardt, Ali Saidi, Arkaprava Basu, Joel Hestness,
1364
+ Derek R. Hower, Tushar Krishna, Somayeh Sardashti, Rathijit Sen, Korey Sewell, Muhammad Shoaib, Nilay Vaish,
1365
+ Mark D. Hill, and David A. Wood. The gem5 simulator. SIGARCH Comput. Archit. News, 39(2):1–7, aug 2011.
1366
+ ISSN 0163-5964. doi:10.1145/2024716.2024718.
1367
+ Christian Bienia, Sanjeev Kumar, Jaswinder Pal Singh, and Kai Li. The parsec benchmark suite: Characterization and
1368
+ architectural implications. In 2008 International Conference on Parallel Architectures and Compilation Techniques
1369
+ (PACT), pages 72–81, 2008.
1370
+ D.H. Bailey, E. Barszcz, L. Dagum, and H.D. Simon.
1371
+ Nas parallel benchmark results.
1372
+ In Supercom-
1373
+ puting ’92:Proceedings of the 1992 ACM/IEEE Conference on Supercomputing, pages 386–393, 1992.
1374
+ doi:10.1109/SUPERC.1992.236665.
1375
+ Shuai Che, Michael Boyer, Jiayuan Meng, David Tarjan, Jeremy W. Sheaffer, Sang-Ha Lee, and Kevin Skadron.
1376
+ Rodinia: A benchmark suite for heterogeneous computing. In 2009 IEEE International Symposium on Workload
1377
+ Characterization (IISWC), pages 44–54, 2009. doi:10.1109/IISWC.2009.5306797.
1378
+ O. Mutlu, Hyesoon Kim, and Y.N. Patt. Efficient runahead execution: Power-efficient memory latency tolerance. IEEE
1379
+ Micro, 26(1):10–20, 2006. doi:10.1109/MM.2006.10.
1380
+ J.-M. Parcerisa and A. Gonzalez. Improving latency tolerance of multithreading through decoupling. IEEE Transactions
1381
+ on Computers, 50(10):1084–1094, 2001. doi:10.1109/12.956093.
1382
+ T.C. Mowry, C.Q.C. Chan, and A.K.W. Lo. Comparative evaluation of latency tolerance techniques for software
1383
+ distributed shared memory. In Proceedings 1998 Fourth International Symposium on High-Performance Computer
1384
+ Architecture, pages 300–311, 1998. doi:10.1109/HPCA.1998.650569.
1385
+ 24
1386
+
1387
+ arXiv Template
1388
+ A PREPRINT
1389
+ Satyanarayana Nekkalapu, Haitham Akkary, Komal Jothi, Renjith Retnamma, and Xiaoyu Song.
1390
+ A simple la-
1391
+ tency tolerant processor. In 2008 IEEE International Conference on Computer Design, pages 384–389, 2008.
1392
+ doi:10.1109/ICCD.2008.4751889.
1393
+ Sparsh Mittal and Jeffrey S. Vetter. A survey of software techniques for using non-volatile memories for storage
1394
+ and main memory systems. IEEE Trans. Parallel Distrib. Syst., 27(5):1537–1550, may 2016. ISSN 1045-9219.
1395
+ doi:10.1109/TPDS.2015.2442980.
1396
+ Yehia Arafa, Abdel-Hameed Badawy, Ammar ElWazir, Atanu Barai, Ali Eker, Gopinath Chennupati, Nandakishore
1397
+ Santhi, and Stephan Eidenbenz. Hybrid, scalable, trace-driven performance modeling of GPGPUs. In Proceedings of
1398
+ the International Conference for High Performance Computing, Networking, Storage and Analysis, SC ’21, 2021.
1399
+ ISBN 9781450384421. doi:10.1145/3458817.3476221.
1400
+ Jack Choquette and Wish Gandhi. NVIDIA A100 GPU: Performance & innovation for GPU computing. In 2020 IEEE
1401
+ Hot Chips 32 Symposium (HCS), pages 1–43, 2020. doi:10.1109/HCS49909.2020.9220622.
1402
+ S. Grauer-Gray, L. Xu, R. Searles, S. Ayalasomayajula, and J. Cavazos. Auto-tuning a high-level language targeted to
1403
+ gpu codes. In Innovative Parallel Computing, InPar ’12, page 1–10, 2012. doi:10.1109/InPar.2012.6339595.
1404
+ A. Karki, C. P. Keshava, S. M. Shivakumar, J. Skow, G. M. Hegde, and H. Jeon. Detailed characterization of deep
1405
+ neural networks on gpus and fpgas. In Proceedings of the 12th Workshop on General Purpose Processing Using
1406
+ GPUs, GPGPU ’19, page 12–21, 2019. doi:https://doi.org/10.1145/3300053.3319418.
1407
+ Javad Rahimi, Joris Van Kerrebrouck, Bahawal Haq, Johan Bauwelinck, Gunther Roelkens, and Geert Morthier.
1408
+ Demonstration of a high-efficiency short-cavity iii-v-on-si c-band dfb laser diode. IEEE Journal of Selected Topics in
1409
+ Quantum Electronics, 28(3):1–6, 2022. doi:10.1109/JSTQE.2021.3122552.
1410
+ Chen Sun, Daniel Jeong, Mason Zhang, Woorham Bae, Chong Zhang, Pavan Bhargava, Derek Van Orden, Shahab
1411
+ Ardalan, Chandarasekaran Ramamurthy, Erik Anderson, Austin Katzin, Haiwei Lu, Sidney Buchbinder, Behrooz
1412
+ Beheshtian, Anatoly Khilo, Michael Rust, Chen Li, Forrest Sedgwick, John Fini, Roy Meade, Vladimir Stojanovi´c,
1413
+ and Mark Wade. TeraPHY: An o-band wdm electro-optic platform for low power, terabit/s optical i/o. In 2020 IEEE
1414
+ Symposium on VLSI Technology, pages 1–2, 2020. doi:10.1109/VLSITechnology18217.2020.9265012.
1415
+ Daniele De Sensi, Salvatore Di Girolamo, Kim H. McMahon, Duncan Roweth, and Torsten Hoefler. An in-depth analysis
1416
+ of the slingshot interconnect. In Proceedings of the International Conference for High Performance Computing,
1417
+ Networking, Storage and Analysis, SC ’20. IEEE Press, 2020b. ISBN 9781728199986.
1418
+ M. R. Siavash Katebzadeh, Paolo Costa, and Boris Grot. Evaluation of an infiniband switch: Choose latency or
1419
+ bandwidth, but not both. In 2020 IEEE International Symposium on Performance Analysis of Systems and Software
1420
+ (ISPASS), pages 180–191, 2020. doi:10.1109/ISPASS48437.2020.00033.
1421
+ Mallikarjun Vasa, Chun-Lin Liao, Sanjay Kumar, Ching-Huei Chen, and Bhyrav Mutnury. Pcie gen-5 design challenges
1422
+ of high-speed servers. In 2020 IEEE 29th Conference on Electrical Performance of Electronic Packaging and Systems
1423
+ (EPEPS), pages 1–3, 2020. doi:10.1109/EPEPS48591.2020.9231458.
1424
+ Keun Sup Shim, Brian Greskamp, Brian Towles, Bruce Edwards, J. P. Grossman, and David E. Shaw. The specialized
1425
+ high-performance network on anton 3, 2022.
1426
+ 25
1427
+
KdE2T4oBgHgl3EQfAgZ0/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
KtAyT4oBgHgl3EQf6Prq/content/2301.00820v1.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7f56af75982630d72e7f2594d26939fd545ed996c93953319ec03452a9dfdee
3
+ size 2576273
KtAyT4oBgHgl3EQfsfmL/vector_store/index.faiss ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e13d85cfb6af075f56332045f61d465bc850be12f35903bf43f79dccd12bef5e
3
+ size 5636141
KtE1T4oBgHgl3EQfsgWq/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,540 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf,len=539
2
+ page_content='Leukemia Detection Based on Microscopic Blood Smear Images Using Deep Neural Networks Abdelmageed Ahmed dept.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
3
+ page_content=' Engineering Electrical and Computer Engineering University of Ottawa Cairo, Egypt ahass202@uottawa.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
4
+ page_content='ca Ahmed Kamal dept.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
5
+ page_content=' biomedical engineering department Minai university Minya, Egypt ahmd654@yahoo.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
6
+ page_content='com Alaa Nagy dept.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
7
+ page_content=' Engineering Electrical and Computer Engineering University of Ottawa Cairo, Egypt aelba046@uottawa.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
8
+ page_content='ca Daila Farghl dept.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
9
+ page_content=' biomedical engineering department Minai university Minya, Egypt dolly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
10
+ page_content='mostafa93@yahoo.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
11
+ page_content='com Abstract— In this paper we discuss a new method for detecting leukemia in microscopic blood smear images using deep neural networks to diagnose leukemia early in blood.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
12
+ page_content=' leukemia is considered one of the most dangerous mortality causes for a human being, the traditional process of diagnosis of leukemia in blood is complex, costly, and time- consuming, so patients could not receive medical treatment on time;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
13
+ page_content=' Computer vision classification technique using deep learning can overcome the problems of traditional analysis of blood smears, our system for leukemia detection provides 97.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
14
+ page_content='3 % accuracy in classifying samples as cancerous or normal samples by taking a shot of blood smear and passing it as an input to the system that will check whether it contains cancer or not.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
15
+ page_content=' In case of containing cancer cells, then the hematological expert passes the sample to a more complex device such as flow cytometry to generate complete information about the progress of cancer in the blood.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
16
+ page_content=' Keywords— Leukemia cells, leukemia detection, deep neural networks, deep learning.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
17
+ page_content=' I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
18
+ page_content=' INTRODUCTION Leukemia is a type of cancer affecting blood;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
19
+ page_content=' if it is detected late, it will result in death.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
20
+ page_content=' Leukemia develops when the bone marrow produces an excessive number of aberrant white blood cells.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
21
+ page_content=' The normal of the blood system will be disrupted when aberrant white blood cells are in excess.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
22
+ page_content=' Hematologists can identify abnormal blood when they draw a blood sample and study it[1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
23
+ page_content=' However, hematologists will inspect microscopic images visually, and the process is time- consuming and tiring [1 - 3].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
24
+ page_content=' Moreover, the process requires human experts and is prone to errors due to emotional disturbance and human physical capability, which has its limitations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
25
+ page_content=' Moreover, it is not easy to get consistent results from visual inspection [1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
26
+ page_content=' Visual inspection can only give qualitative results for further research [1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
27
+ page_content=' Studies indicate that the majority of modern methods.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
28
+ page_content=' Use all blood-related data, such as the number of red blood cells, hemoglobin level, hematocrit level, mean corpuscular volume, and much more, as the criterion for categorizing disorders like cancer, thalassemia, Etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
29
+ page_content=' Expensive testing and equipment labs are required to know all information about blood.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
30
+ page_content=' An automatic image processing system is urgently needed and can overcome related constraints in visual inspection.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
31
+ page_content=' The system to be developed will be based on microscopic images to recognize leukemia cells in blood smears.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
32
+ page_content=' The early and fast identification of the leukemia type greatly aids in providing the appropriate treatment for a particular type of leukemia [4].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
33
+ page_content=' The currently used diagnostic methods rely on analyzing immuno- phenotyping, fluorescence in situ hybridization, cytogenetic analysis, and cytochemistry.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
34
+ page_content=' Sophisticated and expensive laboratories are required in order to run the diagnostic methods, and it has been reported to provide a high ratio of misidentification;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
35
+ page_content=' with this system, more images can be processed, reduce analyzing time, exclude the influence of subjective factors, and increase the accuracy of identification process at the same time.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
36
+ page_content=' In machine learning, the inspection and classification of leukemia will be based on the texture, shape, size, color, and statistical analysis of white blood cells.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
37
+ page_content=" In contrast, deep learning makes it much more profound and gets the whole image's exclusive features." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
38
+ page_content=' This project is applied to increase efficiency globally and can simultaneously benefit and be a massive contribution to the medical and pattern recognition field.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
39
+ page_content=' The main objective is to enhance algorithms that can extract data from human blood where human blood is the primary source to detect diseases at an earlier stage and can prevent it quickly [5].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
40
+ page_content=' This system should be robust towards diversity among individuals, sample collection protocols, time, Etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
41
+ page_content=' This automated system can produce lab results quickly, easily, and efficiently.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
42
+ page_content=' II.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
43
+ page_content=' DATASET Images that were used in this project were downloaded from the internet and are available in ALL IDB[6], ASH Image Bank Hematology [7], Stock photo, vectors and Royalty-free Images[8], Shutter stock[9], Atlas of Hematology [10], Atlas of blood smear analysis[11], Blue Histology and American Society of Hematology [12], This dataset is composed of 630 images, contains 480 cancer images and 150 normal images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
44
+ page_content=' III.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
45
+ page_content=' METHODOLOGY A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
46
+ page_content=" Data Preprocessing 1) Remove duplication As the dataset is collected from various resources, had found that there are some repetitions, some images contain a watermark, and other contains websites' logo totally about 43 images, so now the data set has become 587 images." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
47
+ page_content=' 2) Resizing of images As the dataset has a different distribution of size, and for training the CNN model, it was needed to make all images in the dataset has the same size, so we applied a resizing technique and make all image 256 x 256 pixels to reduce the training time.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
48
+ page_content=' as shown in figure [5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
49
+ page_content='1] 3) Filtering images Before the processing stage, we need to remove noise and enhance line structures in images [13], and this is available by applying a median filter (3 x3) and sharpening the image (3 x3) ,as shown in Fig[1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
50
+ page_content=" Figure 1:(a) original image,(b) image resized by 256*256 and filtered by median and sharpen filters 4) Data augmentation Image data augmentation is a method for artificially increasing the size of a training dataset by producing altered copies of the dataset's images [14]." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
51
+ page_content=' The capacity of fit models to generalize what they have learned to new pictures may be improved by training deep-learning neural network models on more data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
52
+ page_content=' Additionally, augmentation techniques can provide variants of the images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
53
+ page_content=' Through the ImageDataGenerator class, the Keras deep learning neural network framework can fit models by adding picture data [15].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
54
+ page_content=' There are many different types of augmentation techniques, some of them as: a) Flipping An image flip means reversing the rows or columns of pixels in the case of a vertical or horizontal flip [9].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
55
+ page_content=' b) Horizontal and Vertical Shift Augmentation A shift to an image means moving all pixels of the image in one direction, such as horizontally or vertically, while keeping the image dimensions the same;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
56
+ page_content=' this means that some of the pixels will be clipped off the image, and there will be a region of the image where new pixel values will have to be specified [16].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
57
+ page_content=' c) Random Zoom Augmentation A zoom augmentation randomly enlarges the image and either interpolate or adds new pixel values around the image [16].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
58
+ page_content=' d) Shearing Shearing will automatically crop the correct area from the sheared image so that we have an image with no black space or padding [16].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
59
+ page_content=' e) Interpolation (Nearest) A technique for creating new data points within the range of a discrete set of existing data points is interpolation [17].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
60
+ page_content=' Nearest neighbor interpolation is the most straightforward approach to interpolation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
61
+ page_content=' Rather than calculate an average value by some weighting criteria or generate an intermediate value based on complicated rules, this method simply determines the "nearest" neighboring pixel and assumes its intensity value of it [13].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
62
+ page_content=' And Fig[2] indicates a sample image with its augmented one.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
63
+ page_content=' (a) (b) Fig 2: (a) original image and (b) augmented image.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
64
+ page_content=' B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
65
+ page_content=' Processing stage After augmentation processes, our data become 1550 images for cancer and 1480 for normal.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
66
+ page_content=' To fit data to models, we divided it through coding into three data sets: training set, validation set, and test set by ratios 60%, 20%, and 20%, respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
67
+ page_content=' Then the next stage is to train the model that can be able to classify the images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
68
+ page_content=' Our optimizing parameters are accuracy and validation accuracy: to get the best of them as possible, we trained three networks with different architectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
69
+ page_content=' a) BasicCNN model In this model, the input images were (RGB) color images with a resolution of 128x128 pixels.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
70
+ page_content=' It consists of 3 convolutional layers with max pooling layers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
71
+ page_content=' A rectified linear unit follows each convolutional layer (relu).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
72
+ page_content=' We used a constant filter size (3x3), and the number of (a) (b)Filters (128), the stride of ones (equal 1), and fully connected layers trained for two categories classification using the sigmoid activation function.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
73
+ page_content=' Where we classified the data set into leukemia cells or normal cells, this architect achieved 90.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
74
+ page_content='99% accuracy and 84.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
75
+ page_content='97 % validation accuracy after 17 epochs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
76
+ page_content=' Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
77
+ page_content='3: Indicates the block diagram of the basic CNN model b) Alexnet architecture In this study, we deployed the pre-trained AlexNet to detect ALL and classify its subtypes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
78
+ page_content=' This architecture was proposed by Krizhevsky et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
79
+ page_content=', nine who deployed this architecture for the ImageNet Large Scale Visual Recognition Challenge 2012,20 and won the challenge in the first place.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
80
+ page_content=' Input images were Red, Green, and Blue (RGB) color images with a resolution of 227 x 227 pixels.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
81
+ page_content=' It consists of 5 convolutional layers with three max polling layers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
82
+ page_content=' Each convolutional layer in AlexNet architecture is followed by a rectified linear unit (ReLU).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
83
+ page_content=' All the parameters, including the filter size, the number of filters, and the stride for each layer, are illustrated in Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
84
+ page_content='4;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
85
+ page_content=' we replaced the SoftMax layer with a sigmoid layer as we want to classify the input image into only two types of this architect achieved 55.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
86
+ page_content='35% accuracy and 49.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
87
+ page_content='76 % validation accuracy after 12 epochs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
88
+ page_content=' Figure 4: AlexNet architecture for acute lymphoblastic leukemia subtype classification.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
89
+ page_content=' Last 2 layers are newly added.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
90
+ page_content=' c) Modification of model used in published paper This used a retrained model that had been used in a published paper [20], shown in figure 5, and we changed the values of the hyperparameter to become as shown in figure 6;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
91
+ page_content=' This network contains five layers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
92
+ page_content=' The first three layers perform feature extraction, and the other two layers (fully connected and SoftMax) classify the extracted features.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
93
+ page_content=' The input image has a size of 128x128x3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
94
+ page_content=' In convolution layer 1, we used a constant filter size of 5x5 and a total of 16 different filters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
95
+ page_content=' The stride is one, and no zero-padding was applied.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
96
+ page_content=' The second and third convolution layers have the same structure as the first one but a different number of filters, 32 and 64, respectively.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
97
+ page_content=' We used a pooling layer with filter size two and stride 2 to decrease the volume spatially.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
98
+ page_content=" During the model we learned, the mini-batch's chosen size was 128, and ReLu was used as the activation function." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
99
+ page_content=' This architect gives: accuracy = 97.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
100
+ page_content='73 % validation accuracy = 94.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
101
+ page_content='64 % Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
102
+ page_content=' 5: The original architecture of CNN in the mentioned paper.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
103
+ page_content=' Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
104
+ page_content=' 6: Architecture of CNN after changes in hyperparameter IV.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
105
+ page_content=' EXPREMENTL RESULT Our experiments were conducted on Python 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
106
+ page_content='7 with 3030 images, 60% (1818 images) of them for training, 20% (606 images) for validation, and the remaining 20% (606 images) for testing our model.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
107
+ page_content=' In order to evaluate each model and clarify the best one, we compare them by some statistically measured parameters: A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
108
+ page_content=' Accuracy Train accuracy For the basic CNN model, train accuracy comes to 90.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
109
+ page_content='99% after 17 epochs;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
110
+ page_content=' our leukemia classifier is doing an excellent classification, as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
111
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
112
+ page_content='1a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
113
+ page_content=' For AlexNet architecture, the accuracy achieved its maximum accuracy of 56% after 11 epochs;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
114
+ page_content=' that means our model is terrible on leukemia classification as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
115
+ page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
116
+ page_content='1 b, but the Modification of the model used in Thanh et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
117
+ page_content=' paper [18] achieved the maximum accuracy over all models 97.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
118
+ page_content='73 % after ten epochs as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
119
+ page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
120
+ page_content='1c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
121
+ page_content=' 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
122
+ page_content='1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
123
+ page_content='3 FC Max Max Max Conv layer Conv layer Conv layer Input image pooling pooling sigmoid pooling 128*128*3 63*63*128 61*61*128 30*30*128 28*28*12814*14*128 126*126*128 No padding No paddingFullyConnected Layer Fully Connected 4096 Layer Follo wedbyRelu 1024 L1 L2 L3 256 Norn Convolution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
124
+ page_content='.ReLu.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
125
+ page_content='.MaxPolling.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
126
+ page_content='.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
127
+ page_content='soon ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
128
+ page_content='384 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
129
+ page_content='Convolution5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
130
+ page_content='256 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
131
+ page_content='Convolution2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
132
+ page_content='ImageSize=13*13 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
133
+ page_content='Image Size = 13*13 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
134
+ page_content='InputImageSize ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
135
+ page_content='4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
136
+ page_content='Convolution ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
137
+ page_content='Convolution ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
138
+ page_content='Fully Connected ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
139
+ page_content='227x227 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
140
+ page_content='96 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
141
+ page_content='Filtersize=3*3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
142
+ page_content='Filter size= 3*3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
143
+ page_content='4096 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
144
+ page_content='Layer ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
145
+ page_content='Image Size=27*27 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
146
+ page_content='No of filters=384 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
147
+ page_content='Nooffilters=256 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
148
+ page_content='FullyConnected ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
149
+ page_content='Foilowedby Softmax ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
150
+ page_content='Convolution1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
151
+ page_content='Convolution ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
152
+ page_content='Stride=1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
153
+ page_content='Stride=1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
154
+ page_content='Filter size= 5*5 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
155
+ page_content='Maxpooling ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
156
+ page_content='Layer ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
157
+ page_content='Image Size = 55*55 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
158
+ page_content='No ofilters= 256 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
159
+ page_content='Filtersize=3*3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
160
+ page_content='Stride= 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
161
+ page_content='Stride=2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
162
+ page_content='Convolution ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
163
+ page_content='Filter size=11*11 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
164
+ page_content='Max pooling ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
165
+ page_content='Nooffilters=96 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
166
+ page_content='Filter size=3*3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
167
+ page_content='Stride=2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
168
+ page_content='Stride= 4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
169
+ page_content='MaxpoolingInput ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
170
+ page_content='Conv layer 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
171
+ page_content='Max-Pooing 1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
172
+ page_content='Conylayer2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
173
+ page_content='Max-Pooling 2Comv layer 3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
174
+ page_content='sdauragayuto ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
175
+ page_content='Com2Feane maps ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
176
+ page_content='Peo2Featuremap ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
177
+ page_content='Cows:Fetmas ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
178
+ page_content='Pooit:Feanremaps ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
179
+ page_content='FC Sohmax ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
180
+ page_content='100x100x3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
181
+ page_content='96x96x16 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
182
+ page_content='48x48x16 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
183
+ page_content='46x46x32 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
184
+ page_content='23x23x32 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
185
+ page_content='21x21x64FC ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
186
+ page_content='sigmoid ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
187
+ page_content='Max ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
188
+ page_content='Max ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
189
+ page_content='OH ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
190
+ page_content='Convlayer ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
191
+ page_content='Inputimage ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
192
+ page_content='Convlayer ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
193
+ page_content='pooling ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
194
+ page_content='Convlayer ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
195
+ page_content='pooling ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
196
+ page_content='128x128x3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
197
+ page_content='62x62x16 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
198
+ page_content='29x29x32 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
199
+ page_content='25x25x64 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
200
+ page_content='64 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
201
+ page_content='58x58x32 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
202
+ page_content='124x124x16 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
203
+ page_content='Stride=2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
204
+ page_content='Stride=1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
205
+ page_content='Stride=2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
206
+ page_content='Stride=1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
207
+ page_content='Stride=1 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
208
+ page_content='No padding ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
209
+ page_content='NopaddingValidation accuracy ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
210
+ page_content='Basic CNN Model validation accuracy reaches 85% after ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
211
+ page_content='17 epochs,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
212
+ page_content=' as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
213
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
214
+ page_content='1a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
215
+ page_content=' Therefore, we expect our model to perform with ~85% accuracy on new data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
216
+ page_content=' For AlexNet architecture, the accuracy achieved its maximum accuracy of 53.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
217
+ page_content='6% after 11 epochs;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
218
+ page_content=' that means our model is terrible on leukemia classification, as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
219
+ page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
220
+ page_content='1 b;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
221
+ page_content=' This means that we expect our model to perform with ~53.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
222
+ page_content='6% accuracy on new data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
223
+ page_content=' Nevertheless, in Modification of the model used in Thanh et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
224
+ page_content=' paper [94] achieved the maximum validation accuracy over all models at 94.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
225
+ page_content='3 % after ten epochs, as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
226
+ page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
227
+ page_content='1c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
228
+ page_content=' Therefore, we expect our model to perform with ~94.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
229
+ page_content='3 % accuracy on new data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
230
+ page_content=' We notice that our train metric increases as epochs increase while the validation accuracy metric decreases.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
231
+ page_content=' That means that our model fits the training set better but slightly loses its ability to predict new data, indicating that our models are beginning to overfit.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
232
+ page_content=' Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
233
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
234
+ page_content='1a, curve of val acc & train acc for basic CNN model Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
235
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
236
+ page_content='2b, curve of val acc & train acc for AlexNet architecture Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
237
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
238
+ page_content='2c, curve of validation accuracy & train accuracy for Modification of model used in Thanh et al paper [18] B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
239
+ page_content=" Confusion Matrix A classification problem's predicted outcomes are compiled in a confusion matrix." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
240
+ page_content=' The count values describe the number of accurate and inaccurate predictions for each class.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
241
+ page_content=' Because it is feasible to see the relationships between the classifier outputs and the real ones, this is a great alternative for reporting results in M-class classification issues.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
242
+ page_content=' For the basic CNN model, the number of leukemia images that are predicted as leukemia is 372, the number of leukemia images that are predicted as normal is 8, the number of normal images predicted as normal is 269, and the number of normal images that are predicted as leukemia is 51, as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
243
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
244
+ page_content='3 a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
245
+ page_content=' These accuracies show that this model is good at predicting leukemia images but bad at predicting normal images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
246
+ page_content=' For AlexNet architecture, the number of leukemia images that are predicted as leukemia is 0, the number of leukemia images that are predicted as normal is 380, the number of normal images predicted as normal is 157, and the number of normal images that are predicted as leukemia is 163, as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
247
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
248
+ page_content='3 b.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
249
+ page_content=' These accuracies show that this model is terrible at predicting normal images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
250
+ page_content=' The number of leukemia images predicted as leukemia for the modified model used in the published paper [18] is 369, the number of leukemia images predicted as normal is 11, the number of normal images predicted as normal is 301, and the number of normal images predicted as leukemia is 19;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
251
+ page_content=' as shown in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
252
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
253
+ page_content='3 c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
254
+ page_content=' These accuracies show that this model has done a great job of predicting normal images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
255
+ page_content=' Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
256
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
257
+ page_content='3a, Confusion matrix of basic CNN model Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
258
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
259
+ page_content='3b, Confusion matrix of AlexNet Architecture trainaccvsval acc 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
260
+ page_content='9 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
261
+ page_content='8 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
262
+ page_content='6 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
263
+ page_content='5 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
264
+ page_content='4 Tain 2 8 numof Epochstrain acc vs val acc 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
265
+ page_content='552 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
266
+ page_content='550 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
267
+ page_content='548 accuracy 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
268
+ page_content='546 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
269
+ page_content='544 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
270
+ page_content='542 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
271
+ page_content='540 train 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
272
+ page_content='538 val 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
273
+ page_content='536 0 2 4 6 8 10 12 numofEpochs0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
274
+ page_content='98 trainaccvsval acc 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
275
+ page_content='96 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
276
+ page_content='92 train 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
277
+ page_content='90 val 0 1 2 3 4 5 6 7 8 numofEpochsConfusionmatrix 320 372 280 class o(cancer) 240 True label 200 160 120 class 1(normal) 51 269 80 40 Predicted labelConfusion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
278
+ page_content='matrix 320 0 380 280 class O(cancer) 240 Truelabel 200 160 120 class.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
279
+ page_content='1(normal) 163 157 80 40 Predicted label Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
280
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
281
+ page_content='3c, Confusion matrix for Modification of model used in Thanh et al paper [18] C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
282
+ page_content=' Percsision It is calculated as the proportion of accurate positive results to those that the classifier predicted to be positive.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
283
+ page_content=' Our CNN model has medium precision, AlexNet architecture has very low precision, and the modified version of the model used in Thanh et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
284
+ page_content=" 's [18] paper has good precision due to its goodness method, as shown in fig." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
285
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
286
+ page_content='4a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
287
+ page_content=' D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
288
+ page_content=' Recall It is determined by dividing the total number of pertinent samples (all samples that should have been labeled as positive) by the total number of reliable positive results.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
289
+ page_content=' As illustrated in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
290
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
291
+ page_content='4a for our CNN model, fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
292
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
293
+ page_content='4b for the AlexNet architecture, and fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
294
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
295
+ page_content='4c for the Thanh et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
296
+ page_content=' paper [18].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
297
+ page_content=' The perfect model regarded recall is the third model.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
298
+ page_content=' The first CNN model in class 1 has a high recall but low precision.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
299
+ page_content=' This means that most of the positive examples are correctly recognized (low FN), but there are a lot of false positives.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
300
+ page_content=' Nevertheless, in class 0, low recall and high precision show that we miss a lot of positive examples (high FN), but those we predict as positive are indeed positive (low FP).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
301
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
302
+ page_content=' F1 Score The harmonic mean of recall and accuracy is the F1 score.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
303
+ page_content=' The F1 score has a range of [0, 1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
304
+ page_content=' It tells how accurate the classifier is (how many instances it classifies correctly) and how robust it is (it recognizes a significant number of instances).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
305
+ page_content=' As illustrated in fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
306
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
307
+ page_content='4a for our CNN model, fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
308
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
309
+ page_content='4b for the AlexNet architecture, and fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
310
+ page_content=' 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
311
+ page_content='4c for the Thanh et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
312
+ page_content=' paper [18].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
313
+ page_content=' These figures show that the modification of the model used in Thanh et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
314
+ page_content=" 's paper [18] is precise and robust." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
315
+ page_content=' F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
316
+ page_content=' Support Support is the number of samples accurately representing the response within that category.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
317
+ page_content=' It provides information on the precise numbers of each class in the test data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
318
+ page_content=' Figures 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
319
+ page_content='4a and 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
320
+ page_content='4b for the fundamental CNN model, 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
321
+ page_content='4b for the AlexNet architecture, and 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
322
+ page_content='8c for a modified version of the model from the Thanh et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
323
+ page_content=' work [18] serve as examples.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
324
+ page_content=' Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
325
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
326
+ page_content='4a, values of precision, recall, f1 score and support for our CNN model Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
327
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
328
+ page_content='4b, values of precision, recall, f1 score and support for AlexNet architecture Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
329
+ page_content='7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
330
+ page_content='4c, values of precision, recall, f1 score and support for Modification of model used in Thanh et al paper [18] V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
331
+ page_content=" DISCUSSION Leukemia is a malignancy that affects the body's blood- forming tissues, including the lymphatic system and bone marrow." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
332
+ page_content=' To get the most effective treatment, the patient needs early Diagnosis, so we deploy three models using the power of CNN to classify blood smears into normal and abnormal.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
333
+ page_content=' Our dataset had not been taken under the same conditions as it was collected from various resources, and it needed to be bigger to use with DL.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
334
+ page_content=' To overcome this problem, we used the power of data augmentation;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
335
+ page_content=' this solution was suitable for us;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
336
+ page_content=' our data before augmentation was 260 images, and after augmentation became 3030 images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
337
+ page_content=' Our optimizing parameters were accuracy and validation accuracy;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
338
+ page_content=' by using CNN, we trained the model: the First model consists of 3 convolutional layers with max pooling layers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
339
+ page_content=' Its accuracy was 90% and 84.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
340
+ page_content='97 % validation accuracy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
341
+ page_content=' It was terrible with our dataset due to its few layers, so we trained another model the Second model was AlexNet;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
342
+ page_content=' this architecture proved its efficiency in CNN models, so we trained it with our data, input is (RGB) color images with a resolution of 227 x 227 pixels.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
343
+ page_content=' It consists of 5 convolutional layers with three max polling layers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
344
+ page_content=' These models achieved 55.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
345
+ page_content='35% accuracy and 49.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
346
+ page_content='76 % validation accuracy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
347
+ page_content=' We found that it does not fit our dataset.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
348
+ page_content=' So we still have the same problem of low accuracy and keep looking for another model In the last model, we used a retrained model that had been used in a published paper [18];' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
349
+ page_content=' it contain7 layers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
350
+ page_content=' The first five layers perform feature extraction, and the other two layers (fully connected and SoftMax) classify the extracted Confusionmatrix 320 369 11 280 class o(cancer) 240 True label 200 160 120 class 1(normal 19 301 80 40 cer) Predicted labelprecision recall f1-score support class 0(cancerous) 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
351
+ page_content='98 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
352
+ page_content='62 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
353
+ page_content='76 78 class 1(Normal) 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
354
+ page_content='84 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
355
+ page_content='99 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
356
+ page_content='91 163precision recall f1-score class 1(cancer) 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
357
+ page_content='00 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
358
+ page_content='0 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
359
+ page_content='00 178 class e(normal) 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
360
+ page_content='49 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
361
+ page_content='00 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
362
+ page_content='66 172precision recall f1-score support class o(cancer) 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
363
+ page_content='97 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
364
+ page_content='95 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
365
+ page_content='96 373 class 1(normal) 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
366
+ page_content='94 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
367
+ page_content='97 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
368
+ page_content='95 327features.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
369
+ page_content=' The input image has a size of 128x128x3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
370
+ page_content=' This architect has an accuracy of 97.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
371
+ page_content='73 % validation accuracy is 94.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
372
+ page_content='64 %, finally, we found that this model fit our data VI.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
373
+ page_content=' CONCLUSIONS In this system, we investigated the application of deep CNNs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
374
+ page_content=' We deployed a pre-trained model for detecting and classifying the blood sample into normal and abnormal samples using microscopic blood sample images and convolutional neural network classification algorithms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
375
+ page_content=' The system was built by deep learning, which uses all features in microscopic images, not only examining changes of specific features as a classifier input.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
376
+ page_content=" We have performed the pre- trained model in a largely augmented dataset to confirm the system's accuracy and reliability." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
377
+ page_content=' By performing data augmentation, we can achieve 97.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
378
+ page_content='3% accuracy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
379
+ page_content=' The system has high accuracy, and less processing time (show results in less than 30 seconds) , minor errors, and early identification of leukemia successful in giving the patient the proper care.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
380
+ page_content=' And cheaper cost.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
381
+ page_content=' The detection system was built in three parts: 1) the acquisition part, which consists of a digital camera that has been installed at the top of the eyepiece of the microscope, 2) pre-trained CNN model responsible for the classification system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
382
+ page_content=' 3) a graphical user interface to display the image obtained from the camera and show the classification results.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
383
+ page_content=' VII.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
384
+ page_content=' FUTURE WORK Expanding the focus on classifying the subtypes of leukemia cells such as Acute Myeloid Leukemia or AML, Chronic Myeloid Leukemia or CML, Acute Lymphoid Leukemia or ALL, and Chronic Lymphoid Leukemia or CLL not only separating between cancerous and non-cancerous cells and developing a convenient environment to construct an extensive leukemia dataset as this topic of research suffer from leaks in images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
385
+ page_content=' VIII.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
386
+ page_content=' REFERENCES [1] N.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
387
+ page_content=', Ritter, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
388
+ page_content=', Cooper, “Segmentation and Border Identification of Cells in Images of Peripheral Blood Smear Slides”, 30th Australasian Computer Science Conference, Conference in Research and Practice in Information Technology, Vol.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
389
+ page_content=' 62, 2007, pp.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
390
+ page_content=' 161-169.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
391
+ page_content=' [2] D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
392
+ page_content='M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
393
+ page_content='U.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
394
+ page_content=', Sabino, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
395
+ page_content='D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
396
+ page_content='F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
397
+ page_content=', Costa, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
398
+ page_content='D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
399
+ page_content='F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
400
+ page_content=', E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
401
+ page_content='G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
402
+ page_content=', Rizzatti, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
403
+ page_content='A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
404
+ page_content=', Zago, “A Texture Approach to Leukocyte Recognition”, Real Time Imaging, Vol.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
405
+ page_content=' 10, 2004, pp.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
406
+ page_content=' 205-206.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
407
+ page_content=' [3] 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
408
+ page_content=' M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
409
+ page_content='C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
410
+ page_content=', Colunga, O.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
411
+ page_content='S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
412
+ page_content=', Siordia, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
413
+ page_content='J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
414
+ page_content=', Maybank, “Leukocyte Recognition Using EMAlgorithm”, MICAI 2009, LNAI 5845, Springer Verlag Berlin Heidelberg, 2009, pp.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
415
+ page_content=' 545-555.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
416
+ page_content=' [4] J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
417
+ page_content=' Walter,” understanding leukemia”, 2009, pp- 8-18 [5] Warren S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
418
+ page_content=' McCulloch and Walter Pitts.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
419
+ page_content=' “Neurocomputing: Foundations of Research”.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
420
+ page_content=' In: edited by James A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
421
+ page_content=' Anderson and Edward Rosenfeld.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
422
+ page_content=' Cambridge, MA, USA: MIT Press,1988.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
423
+ page_content=' Chapter A Logical Calculus of the Ideas Immanent in Nervous Activity, page15–27.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
424
+ page_content='ISBN:0-262-01097-6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
425
+ page_content='URL: http://dl.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
426
+ page_content='acm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
427
+ page_content='org/citation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
428
+ page_content='cfm?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
429
+ page_content=' id=65669.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
430
+ page_content='104377.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
431
+ page_content=' [6] ALL-IDB Acute Lymphoblastic Leukemia Image Database for Image Processing.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
432
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
433
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
434
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
435
+ page_content=' https://scotti.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
436
+ page_content='di.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
437
+ page_content='unimi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
438
+ page_content='it/all/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
439
+ page_content=' https://scotti.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
440
+ page_content='di.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
441
+ page_content='unimi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
442
+ page_content='it/all/ [7] ImageBank | Home | Regular Bank.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
443
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
444
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
445
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
446
+ page_content=' https://imagebank.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
447
+ page_content='hematology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
448
+ page_content='org/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
449
+ page_content=' https://imagebank.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
450
+ page_content='hematology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
451
+ page_content='org/ [8] Stock Photos, Vectors and Royalty Free Images from 123RF.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
452
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
453
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
454
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
455
+ page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
456
+ page_content='123rf.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
457
+ page_content='com/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
458
+ page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
459
+ page_content='123rf.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
460
+ page_content='com/ [9] Stock Images, Photos, Vectors, Video, and Music | Shutterstock.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
461
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
462
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
463
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
464
+ page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
465
+ page_content='shutterstock.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
466
+ page_content='com/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
467
+ page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
468
+ page_content='shutterstock.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
469
+ page_content='com/ [10] ATLAS OF HEMATOLOGY.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
470
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
471
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
472
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
473
+ page_content=' http://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
474
+ page_content='hematologyatlas.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
475
+ page_content='com/principalpage.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
476
+ page_content='htm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
477
+ page_content=' http://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
478
+ page_content='hematologyatlas.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
479
+ page_content='com/principalpage.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
480
+ page_content='htm [11] Chronolab.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
481
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
482
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
483
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
484
+ page_content=' https://chronolab.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
485
+ page_content='com/atlas/hemat.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
486
+ page_content=' https://chronolab.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
487
+ page_content='com/atlas/hemat [12] American Society of Hematology - Hematology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
488
+ page_content='org.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
489
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
490
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
491
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
492
+ page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
493
+ page_content='hematology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
494
+ page_content='org/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
495
+ page_content=' https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
496
+ page_content='hematology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
497
+ page_content='org/ [13] Advanced Data Augmentation Strategies | by Connor Shorten | Towards Data Science.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
498
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
499
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
500
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
501
+ page_content=' https://towardsdatascience.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
502
+ page_content='com/advanced-data-augmentation- strategies-383226cd11ba.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
503
+ page_content=' https://towardsdatascience.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
504
+ page_content='com/advanced- data-augmentation-strategies-383226cd11ba [14] Data Augmentation For Bounding Boxes: Flipping.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
505
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
506
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
507
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
508
+ page_content=' https://blog.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
509
+ page_content='paperspace.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
510
+ page_content='com/data-augmentation-for-bounding-boxes/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
511
+ page_content=' https://blog.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
512
+ page_content='paperspace.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
513
+ page_content='com/data-augmentation-for-bounding-boxes/ [15] How to Configure Image Data Augmentation in Keras - MachineLearningMastery.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
514
+ page_content='com.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
515
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
516
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
517
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
518
+ page_content=' https://machinelearningmastery.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
519
+ page_content='com/how-to-configure-image-data- augmentation-when-training-deep-learning-neural-networks/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
520
+ page_content=' https://machinelearningmastery.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
521
+ page_content='com/how-to-configure-image-data- augmentation-when-training-deep-learning-neural-networks/ [16] Data Augmentation | How to use Deep Learning when you have Limited Data — Part 2 | by Bharath Raj | NanoNets | Medium.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
522
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
523
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
524
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
525
+ page_content=' https://medium.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
526
+ page_content='com/nanonets/how-to-use-deep-learning-when-you- have-limited-data-part-2-data-augmentation-c26971dc8ced.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
527
+ page_content=' https://medium.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
528
+ page_content='com/nanonets/how-to-use-deep-learning-when-you- have-limited-data-part-2-data-augmentation-c26971dc8ced [17] mdbloice/Augmentor: Image augmentation library in Python for machine learning.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
529
+ page_content=' (n.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
530
+ page_content='d.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
531
+ page_content=').' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
532
+ page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
533
+ page_content='com/mdbloice/Augmentor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
534
+ page_content=' https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
535
+ page_content='com/mdbloice/Augmentor [18] T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
536
+ page_content=' T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
537
+ page_content=' P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
538
+ page_content=' Thanh, Caleb Vununu, Sukhrob Atoev, Suk-Hwan Lee, and Ki-Ryong Kwon , “Leukemia Blood Cell Image Classification Using Convolutional Neural Network “ International Journal of Computer Theory and Engineering, Vol.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
539
+ page_content=' 10, No.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
540
+ page_content=' 2, April 2018' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/KtE1T4oBgHgl3EQfsgWq/content/2301.03367v1.pdf'}
LdAyT4oBgHgl3EQf6fqP/content/2301.00823v1.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1ed78e4d983069731906ca99daf3c07efc1461a9a17b120a4072d4793aa992e
3
+ size 962584
LdAyT4oBgHgl3EQf6fqP/vector_store/index.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:935a24368bb95f8c1242dc3e4381d7666a556a668196cbfba95a89dfa3acdb55
3
+ size 466283
LtE1T4oBgHgl3EQftAUX/content/tmp_files/2301.03371v1.pdf.txt ADDED
@@ -0,0 +1,3133 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 1
2
+ Learning Optimal Phase-Shifts of Holographic
3
+ Metasurface Transceivers
4
+ Debamita Ghosh, IITB-Monash Research Academy, IIT Bombay, India
5
+ Manjesh K. Hanawal, MLioNS Lab, IEOR, IIT Bombay, India
6
+ Nikola Zlatanov, Innopolis University, Russia
7
+ Abstract—Holographic metasurface transceivers (HMT) is
8
+ an
9
+ emerging
10
+ technology
11
+ for
12
+ enhancing
13
+ the
14
+ coverage
15
+ and
16
+ rate of wireless communication systems. However, acquiring
17
+ accurate channel state information in HMT-assisted wireless
18
+ communication systems is critical for achieving these goals.
19
+ In this paper, we propose an algorithm for learning the
20
+ optimal
21
+ phase-shifts
22
+ at
23
+ a
24
+ HMT
25
+ for
26
+ the
27
+ far-field
28
+ channel
29
+ model.
30
+ Our
31
+ proposed
32
+ algorithm
33
+ exploits
34
+ the
35
+ structure
36
+ of
37
+ the channel gains in the far-field regions and learns the
38
+ optimal
39
+ phase-shifts
40
+ in
41
+ presence
42
+ of
43
+ noise
44
+ in
45
+ the
46
+ received
47
+ signals.
48
+ We
49
+ prove
50
+ that
51
+ the
52
+ probability
53
+ that
54
+ the
55
+ optimal
56
+ phase-shifts estimated by our proposed algorithm deviate from
57
+ the true values decays exponentially in the number of pilot
58
+ signals. Extensive numerical simulations validate the theoretical
59
+ guarantees and also demonstrate significant gains as compared
60
+ to the state-of-the-art policies.
61
+ Index Terms—Holographic Metasurface Transceivers, Channel
62
+ State Information, Uniform Exploration
63
+ I. INTRODUCTION
64
+ Future wireless network technologies, namely beyond-5G
65
+ and 6G, have been focused on millimeter wave (mmWave)
66
+ and TeraHertz (THz) communications technologies as possible
67
+ solutions to the ever growing demands for higher data rates and
68
+ lower latency. However, mmWave and THz communications
69
+ have challenges that need to be addressed before this technology
70
+ is adopted [1], [2]. One such major challenge is signal
71
+ deterioration due to reflections and absorption.
72
+ A possible solution for the signal deterioration are base
73
+ stations (BSs) with massive antennas arrays that can provide
74
+ large beamforming gains and thereby compensate for the
75
+ signal deterioration [3]. However, implementing a BS with
76
+ a massive antenna array is itself challenging due to the high
77
+ hardware costs. Holographic Metasurface Transceivers (HMTs)
78
+ are introduced as a promising solution for building a massive
79
+ antenna array [4], [5]. A HMT is comprised of a large number
80
+ of metamaterial elements densely deployed into a limited
81
+ surface area in order to form a spatially continuous transceiver
82
+ aperture. These metamaterial elements at the HMT acts as
83
+ phase-shifting antennas, where each phase-shifting element
84
+ of the HMT can change the phase of transmiting/receiving
85
+ signal and thereby beamform towards desired directions where
86
+ the users are allocated [6]. Due to these continuous apertures,
87
+ HMTs can be represented as an extension of the traditional
88
+ massive antenna arrays with discrete antennas to continuous
89
+ reflecting surfaces [6].
90
+ In this paper, we consider the HMT-assisted wireless systems
91
+ illustrated in Fig. 1, where a HMT acts as a BS that serves
92
+ multiple users. The performance of this system is dependent on
93
+ channel state information (CSI) estimates at the HMT, which
94
+ are used for accurate beamforming towards the users. The
95
+ authors in [7] and [8] have studied the effect of HMT-assisted
96
+ systems on enhancing the communication performance under
97
+ the assumption of perfect CSI. However, perfect CSI is not
98
+ available in practice. In practice, the CSI has to be estimated
99
+ via pilot signals, which results in inaccurate CSI estimates at
100
+ the HMT.
101
+ The aim of this paper is to obtain accurate CSI estimates at
102
+ the HMT, which in turn is used to set the optimal phase-shifts
103
+ at the HMT that maximize the data rate to the users when
104
+ the users are located in the far-field. To this end, we exploit
105
+ the structure of the far-field channel model between the HMT
106
+ and the users to show that the optimal phase-shifts at the
107
+ HMT can be obtained from five samples of the received pilot
108
+ signals at the HMT in a noiseless environment. We then use this
109
+ approach to develop a learning algorithm that learns the optimal
110
+ phase-shifts from the received pilot signals at the HMT in a
111
+ noisy environment. Finally, we provide theoretical guarantees
112
+ for our learning algorithm. Specifically, we prove that the
113
+ probability of the phase-shifts generated by our algorithm to
114
+ deviate by more than 𝜖 from the optimal phase-shifts is small
115
+ and decays as the number of pilot symbols increases. The
116
+ error analysis is based on tail probabilities of the non-central
117
+ Chi-squared distribution.
118
+ In summary, our main contributions are as follows:
119
+ • We propose an efficient learning algorithm for estimating
120
+ the optimal phase-shifts at an HMT in the presence of
121
+ noise for the case when the users that the HMT is serving
122
+ are located at the far-field region.
123
+ • We prove that the probability of the phase-shifts generated
124
+ by our algorithm to deviate by more than 𝜖 from the
125
+ optimal phase-shifts is small and decays exponentially as
126
+ the number of pilots used for estimation increases.
127
+ • We show numerically that the performance of the
128
+ proposed algorithm significantly outperforms existing CSI
129
+ estimation algorithms.
130
+ A. Related Works
131
+ Several channel estimation schemes, which are proposed
132
+ for the massive antenna arrays, are also applicable to the
133
+ considered HMT including exhaustive search [9], hierarchical
134
+ search [10], [11], and compressed sensing (CS) [11]. As the
135
+ exhaustive search in [9] significantly increases the training
136
+ arXiv:2301.03371v1 [eess.SP] 12 Dec 2022
137
+
138
+ 2
139
+ overhead, the authors in [10] and [11] proposed the hierarchical
140
+ search based on a predefined codebook as an improvement over
141
+ the exhaustive search. The hierarchical schemes, in general,
142
+ may incur high training overhead and system latency since they
143
+ require non-trivial coordination among the transmitter and the
144
+ user [11]. On the other hand, the proposed CS-based channel
145
+ estimation scheme in [11] provides trade-offs between accuracy
146
+ of estimation and training overhead at different computational
147
+ costs.
148
+ On the other hand, CSI estimation schemes developed
149
+ specifically for HMTs can be found in [12] and [13]. The
150
+ authors in [12] proposed the least-square estimation based
151
+ approach to study the channel estimation problem for the
152
+ uplink between a single user and the BS equipped with the
153
+ holographic surface with a large number of antennas. However,
154
+ the authors require an additional knowledge of antennas array
155
+ geometry to reduce the pilot overhead required by the channel
156
+ estimation, and hence the computational complexity scales
157
+ up with the number of antennas at the BS. In [13], the
158
+ authors proposed a scheme for the estimation of the far-field
159
+ channel between a HMT and a user that requires only five
160
+ pilots for perfect estimation in the noise-free environment.
161
+ In the noisy case, the authors of [13] proposed an iterative
162
+ algorithm that efficiently estimates the far-field channel. Unlike
163
+ the existing works, the training overhead and the computational
164
+ cost of the proposed scheme in [13] does not scale with the
165
+ number of phase-shifting elements at the HMT. The iterative
166
+ algorithm in [13] significantly outperforms the hierarchical
167
+ and CS based schemes. However, the authors in [13] did not
168
+ provide any theoretical guarantees on their proposed algorithm.
169
+ Motivated by [13], in this work, we propose an algorithm
170
+ which outperforms the one in [13], and, in addition, we also
171
+ provide theoretical guarantees for our proposed algorithm.
172
+ This paper is organized as follows. The system and channel
173
+ models for the HMT communication system are given in Sec.
174
+ II. The proposed algorithm for learning the optimal phase-shifts
175
+ is given in Sec. III and its theoretical guarantee is provided
176
+ in Sec. IV. Numerical evaluation of the proposed algorithm is
177
+ provided in Sec. V. Finally, Sec. VI concludes the paper.
178
+ II. SYSTEM AND CHANNEL MODELS
179
+ We consider a HMT-assisted wireless communication system,
180
+ shown in Fig. 1, where an HMT communicates with multiple
181
+ users in the mmWave band. We assume that there is a Line
182
+ of Sight (LoS) between the HMT and each user. As a result,
183
+ when modeling the far-field channel, we only take into account
184
+ the LoS path since its power is order of magnitude higher than
185
+ non-line-of-sight (NLoS) paths [14]. The NLoS components
186
+ are incorporated in the noise. We assume that the users send
187
+ orthogonal pilots to the HMT for channel estimation. Based
188
+ on the estimated CSI at the HMT to each user, the HMT sends
189
+ data to the users. Hence, the data rate from the HMT to the
190
+ users is directly dependent on the accuracy of the CSI estimates
191
+ at the HMT. Since in this paper our main goal is the accurate
192
+ CSI estimation at the HMT to each user, which in turn send
193
+ orthogonal pilots to the HMT, in the rest of the paper, we will
194
+ focus on the CSI estimation between the HMT and a typical
195
+ user.
196
+ RF Generator
197
+ Phase-shifting
198
+ Element
199
+ User 1
200
+ User 2
201
+ User 3
202
+ Fig. 1: The HMT-assisted wireless communication system [13].
203
+ A. HMT Model
204
+ The HMT has a rectangular surface of size 𝐿𝑥 × 𝐿𝑦, where
205
+ 𝐿𝑥 and 𝐿𝑦 are the width and the length of the surface,
206
+ respectively. The HMT’s surface is comprised of a large
207
+ number of sub-wavelength phase-shifting elements, where
208
+ each elements is assumed to be a square of size 𝐿𝑒 × 𝐿𝑒
209
+ and can change the phase of the transmit/receive signal
210
+ independently from rest of the elements. Let 𝑑𝑟 be the
211
+ distance between two neighboring phase-shifting elements.
212
+ The total number of phase-shifting elements of the HMT is
213
+ given by 𝑀 = 𝑀𝑥 × 𝑀𝑦, where 𝑀𝑥 = 𝐿𝑥/𝑑𝑟 and 𝑀𝑦 = 𝐿𝑦/𝑑𝑟.
214
+ Without loss of generality, we assume that the HMT lies
215
+ in the 𝑥 − 𝑦 plane of a Cartesian coordinate system, where
216
+ the center of the surface is at the origin of the coordinate
217
+ system. Assuming 𝑀𝑥 and 𝑀𝑦 are odd numbers, the position
218
+ of the (𝑚𝑥,𝑚𝑦)𝑡ℎ phase-shifting element in the Cartesian
219
+ coordinate system is given as (𝑥, 𝑦) = (𝑚𝑥𝑑𝑟,𝑚𝑦𝑑𝑟), where
220
+ 𝑚𝑥 ∈
221
+
222
+ − 𝑀𝑥−1
223
+ 2
224
+ ,..., 𝑀𝑥−1
225
+ 2
226
+
227
+ and 𝑚𝑦 ∈
228
+
229
+ − 𝑀𝑦−1
230
+ 2
231
+ ,..., 𝑀𝑦−1
232
+ 2
233
+
234
+ . When
235
+ 𝑀𝑥 or 𝑀𝑦 is even, the position of the (𝑚𝑥,𝑚𝑦)𝑡ℎ element can
236
+ be appropriately defined.
237
+ B. Channel Model
238
+ Consider the channel between the (𝑚𝑥,𝑚𝑦)𝑡ℎ phase-shifting
239
+ element at the HMT and the typical user. Let the beamforming
240
+ weight imposed by the (𝑚𝑥,𝑚𝑦)𝑡ℎ phase-shifting element at
241
+ the HMT be Γ𝑚𝑥𝑚𝑦 = 𝑒 𝑗𝛽𝑚𝑥 𝑚𝑦 , where 𝛽𝑚𝑥𝑚𝑦 is the phase shift
242
+ at the (𝑚𝑥,𝑚𝑦)𝑡ℎ element. Let 𝜆 denote the wavelength of
243
+ the carrier frequency, 𝑘0 = 2𝜋
244
+ 𝜆 be the wave number, 𝑑0 be the
245
+ distance between the user and the center of the HMT and
246
+ let 𝐹𝑚𝑥𝑚𝑦 denote the effect of the size and power radiation
247
+ pattern of the (𝑚𝑥,𝑚𝑦)𝑡ℎ phase-shifting element on the channel
248
+ coefficient [15]. Due to the far-field assumptions, the radiation
249
+ pattern of all the phase-shifting elements of the HMT are
250
+ identical, i.e., 𝐹𝑚𝑥𝑚𝑦 = 𝐹, ∀𝑚𝑥,𝑚𝑦 holds. Finally, let 𝜃 and
251
+ 𝜙 denote the elevation and azimuth angles of the impinging
252
+ wave from the user to the center of the HMT, see Fig. 2.
253
+ Now, if the phase-shift imposed by the (𝑚𝑥,𝑚𝑦)𝑡ℎ element,
254
+ 𝛽𝑚𝑥,𝑚𝑦, is set to
255
+ 𝛽𝑚𝑥𝑚𝑦 = −
256
+ mod (𝑘0𝑑𝑟 (𝑚𝑥𝛽1 +𝑚𝑦𝛽2),2𝜋),∀𝑚𝑥,𝑚𝑦,
257
+
258
+ 3
259
+ User
260
+ HMT
261
+ Fig. 2: Distance between the (𝑚𝑥,𝑚𝑦)-th phase-shifting element at
262
+ the HMT and the user [13].
263
+ where 𝛽1 and 𝛽2 are the phase-shift parameters [13], [16],
264
+ [17], which are the only degrees of freedom within the
265
+ phase-shift 𝛽𝑚𝑥𝑚𝑦, then the HMT-user channel in the far-field
266
+ is approximated accurately by [13], [16], [17]
267
+ 𝐻(𝛽1, 𝛽2) =
268
+ �√
269
+ 𝐹𝜆𝑒−𝑗𝑘0𝑑0
270
+ 4𝜋𝑑0
271
+
272
+ 𝐿𝑥𝐿𝑦 ×sinc
273
+
274
+ 𝐾𝑥𝜋(𝛼1 − 𝛽1)
275
+
276
+ ×sinc
277
+
278
+ 𝐾𝑦𝜋(𝛼2 − 𝛽2)
279
+
280
+ ,
281
+ (1)
282
+ where 𝐾𝑥 = 𝐿𝑥
283
+ 𝜆 ,𝐾𝑦 = 𝐿𝑦
284
+ 𝜆 ,𝛼1 = sin(𝜃) cos(𝜙),𝛼2 = sin(𝜃) sin(𝜙),
285
+ and sinc(𝑥) = sin(𝑥)
286
+ 𝑥
287
+ . Please note that 𝛼1 ∈ [−1,1] and 𝛼2 ∈
288
+ [−1,1], and their values depend on the location of the user,
289
+ i.e., on 𝜃 and 𝜙.
290
+ From (1), it is clear that the absolute value of the HMT-user
291
+ channel is maximized when the two sinc functions attain
292
+ their maximum values, which occurs when the phase-shifting
293
+ parameters, 𝛽1 and 𝛽2, are set to 𝛽1 = 𝛼1 and 𝛽2 = 𝛼2, where
294
+ (𝛼1,𝛼2) are unknown to the HMT since they depend on the
295
+ location of the user. Therefore, in the far-field case, the problem
296
+ of finding the optimal phase-shifts of the elements at the HMT
297
+ reduces to estimating the two parameters, 𝛼1 and 𝛼2 at the
298
+ HMT.
299
+ Remark 1. Fig. 3 shows an example of |𝐻(𝛽1, 𝛽2)| as a
300
+ function of (𝛽1, 𝛽2). As can be seen from Fig. 3, the graph
301
+ of |𝐻(𝛽1, 𝛽2)| hits zero periodically and has several lobes.
302
+ The optimal value (𝛼1,𝛼2) = (0.68,−0.45) is attained at the
303
+ central lobe which has the highest peak and is attained for
304
+ (𝛽∗
305
+ 1, 𝛽∗
306
+ 2) = (𝛼1,𝛼2) = (0.68,−0.45).
307
+ III. PROPOSED CHANNEL ESTIMATION STRATEGY
308
+ In this section, we propose an algorithm that estimates the
309
+ optimal phase-shifting parameters 𝛽1 and 𝛽2 that maximize
310
+ |𝐻(𝛽1, 𝛽2)| in (1) in the presence of noise.
311
+ A. Problem Formulation
312
+ In the channel estimation procedure, the user sends a
313
+ pilot symbol 𝑥𝑝 =
314
+
315
+ 𝑃 to the HMT, where 𝑃 is the pilot
316
+ transmit power. Then, the received signal at the HMT for
317
+ fixed phase-shifting parameters (𝛽1, 𝛽2), denoted by 𝑦(𝛽1, 𝛽2),
318
+ is given by
319
+ 𝑦(𝛽1, 𝛽2) =
320
+
321
+ 𝑃 × 𝐻(𝛽1, 𝛽2) + 𝜁,
322
+ (2)
323
+ Fig. 3: |𝐻(𝛽1, 𝛽2)| v/s (𝛽1, 𝛽2) for values of (𝛼1,𝛼2) = (0.68,−0.45).
324
+ where 𝜁 is the complex-valued additive white Gaussian noise
325
+ (AWGN) with zero mean and variance 𝜎2 at the HMT. The
326
+ received signal in (2) is then squared in order to obtain the
327
+ received signal squared, denoted by 𝑟(𝛽1, 𝛽2), and given by
328
+ 𝑟(𝛽1, 𝛽2) = |𝑦(𝛽1, 𝛽2)|2 =
329
+ ���
330
+
331
+ 𝑃 × 𝐻(𝛽1, 𝛽2) + 𝜁
332
+ ���
333
+ 2
334
+ .
335
+ (3)
336
+ Objective: Our goal is to identify the optimal phase-shifting
337
+ parameters, denoted by (𝛽∗
338
+ 1, 𝛽∗
339
+ 2), at the HMT that maximizes
340
+ 𝑟(𝛽1, 𝛽2) given by (3). Specifically, we aim to solve the
341
+ following optimisation problem
342
+ (𝛽∗
343
+ 1, 𝛽∗
344
+ 2) = argmax
345
+ 𝛽1∈[−1,1]
346
+ 𝛽2∈[−1,1]
347
+ 𝑟(𝛽1, 𝛽2).
348
+ (4)
349
+ The expected value of 𝑟(𝛽1, 𝛽2), denoted by 𝜇(𝛽1, 𝛽2), is given
350
+ by
351
+ 𝜇(𝛽1, 𝛽2) = E [𝑟(𝛽1, 𝛽2)]
352
+ =
353
+ ���
354
+
355
+ 𝑃 × 𝐻(𝛽1, 𝛽2)
356
+ ���
357
+ 2
358
+ + 𝜎2.
359
+ (5)
360
+ Using (5), the optimization problem in (4) can be written
361
+ equivalently as
362
+ (𝛽∗
363
+ 1, 𝛽∗
364
+ 2) = argmax
365
+ 𝛽1∈[−1,1]
366
+ 𝛽2∈[−1,1]
367
+ 𝜇(𝛽1, 𝛽2).
368
+ (6)
369
+ In order to obtain an intuition on how to solve (6), we first
370
+ assume that 𝜇(𝛽1, 𝛽2) in (5) is known perfectly at the HMT for
371
+ five specific values of the pair (𝛽1, 𝛽2). Later, we use the same
372
+ intuition to solve (6) when 𝜇(𝛽1, 𝛽2) are not known perfectly
373
+ but can be estimated.
374
+ B. The Optimal Phase-Shifting Parameters When 𝜇(𝛽1, 𝛽2)
375
+ Are Known In Advance
376
+ For notational convenience, let us define the set B as
377
+ B =
378
+
379
+ (𝛽0
380
+ 1, 𝛽0
381
+ 2), (𝛽0
382
+ 1 +𝑣, 𝛽0
383
+ 2), (𝛽0
384
+ 1 −𝑣, 𝛽0
385
+ 2),
386
+ (𝛽0
387
+ 1, 𝛽0
388
+ 2 + 𝑤), (𝛽0
389
+ 1, 𝛽0
390
+ 2 − 𝑤)
391
+
392
+ .
393
+ (7)
394
+ The set B is comprised of five pairs of the phase-shifting
395
+ parameters (𝛽1, 𝛽2), where 𝛽0
396
+ 1 and 𝛽0
397
+ 2 are some initial arbitrarily
398
+ selected phase-shifting parameters, 𝑣 and 𝑤 are numbers chosen
399
+ such that 𝐾𝑥𝑣 ∈ N and 𝐾𝑦𝑤 ∈ N hold, where N is the set of
400
+ natural numbers. Please note that for a selected (𝛽0
401
+ 1, 𝛽0
402
+ 2) and a
403
+
404
+ 1
405
+ (α_1,α2)=(0.68,-0.45)
406
+ 0.8
407
+ [H(β1, β2) /2
408
+ 0.5
409
+ 0.6
410
+ 0.4
411
+ 0.4
412
+ β1
413
+ 01
414
+ 0.8
415
+ 0.6
416
+ -0.6
417
+ 0.2
418
+ 0.4
419
+ 0.8
420
+ 0.2
421
+ 2
422
+ β2
423
+ 0.4
424
+ 14
425
+ chosen 𝑣 and 𝑤, if
426
+ ��𝛽0
427
+ 1 ±𝑣
428
+ �� ≥ 1 then we set
429
+ ��𝛽0
430
+ 1 ±𝑣
431
+ �� = 1. In the
432
+ same way, if
433
+ ��𝛽0
434
+ 2 ± 𝑤
435
+ �� ≥ 1 then we set
436
+ ��𝛽0
437
+ 2 ± 𝑤
438
+ �� = 1.
439
+ Theorem 1. If the HMT can obtain 𝜇(𝛽0
440
+ 1, 𝛽0
441
+ 2), 𝜇(𝛽0
442
+ 1 + 𝑣, 𝛽0
443
+ 2),
444
+ 𝜇(𝛽0
445
+ 1 − 𝑣, 𝛽0
446
+ 2), 𝜇(𝛽0
447
+ 1, 𝛽0
448
+ 2 + 𝑤) and 𝜇(𝛽0
449
+ 1, 𝛽0
450
+ 2 − 𝑤), i.e., obtain
451
+ 𝜇(𝛽1, 𝛽2) for the five phase-shifting parameters in (𝛽1, 𝛽2) ∈ B
452
+ given in (7), then the optimal phase-shifting parameters 𝛽∗
453
+ 1
454
+ and 𝛽∗
455
+ 2, which are the solutions of (6), are given by
456
+ 𝛽∗
457
+ 1 =
458
+
459
+ 𝛼(𝑖)
460
+ 1 +𝛼( 𝑗)
461
+ 1
462
+ 2
463
+ :
464
+ min
465
+ 𝑖∈{1,2}, 𝑗 ∈{3,4}
466
+ ���𝛼(𝑖)
467
+ 1 −𝛼( 𝑗)
468
+ 1
469
+ ���
470
+
471
+ ,
472
+ (8)
473
+ where
474
+ 𝛼(1)/(2)
475
+ 1
476
+ = 𝛽0
477
+ 1 +
478
+ 𝑣
479
+
480
+ √︄����
481
+ 𝜇(𝛽0
482
+ 1,𝛽0
483
+ 2)−𝜎2
484
+ 𝜇(𝛽0
485
+ 1+𝑣,𝛽0
486
+ 2)−𝜎2
487
+ ����
488
+ 𝛼(3)/(4)
489
+ 1
490
+ = 𝛽0
491
+ 1 −
492
+ 𝑣
493
+
494
+ √︄����
495
+ 𝜇(𝛽0
496
+ 1,𝛽0
497
+ 2)−𝜎2
498
+ 𝜇(𝛽0
499
+ 1−𝑣,𝛽0
500
+ 2)−𝜎2
501
+ ����
502
+ and
503
+ 𝛽∗
504
+ 2 =
505
+
506
+ 𝛼(𝑖)
507
+ 2 +𝛼( 𝑗)
508
+ 2
509
+ 2
510
+ :
511
+ min
512
+ 𝑖∈{1,2}, 𝑗 ∈{3,4}
513
+ ���𝛼(𝑖)
514
+ 2 −𝛼( 𝑗)
515
+ 2
516
+ ���
517
+
518
+ ,
519
+ (9)
520
+ where
521
+ 𝛼(1)/(2)
522
+ 2
523
+ = 𝛽0
524
+ 2 +
525
+ 𝑣
526
+
527
+ √︄����
528
+ 𝜇(𝛽0
529
+ 1,𝛽0
530
+ 2)−𝜎2
531
+ 𝜇(𝛽0
532
+ 1,𝛽0
533
+ 2+𝑤)−𝜎2
534
+ ����
535
+ 𝛼(3)/(4)
536
+ 2
537
+ = 𝛽0
538
+ 2 −
539
+ 𝑣
540
+
541
+ √︄����
542
+ 𝜇(𝛽0
543
+ 1,𝛽0
544
+ 2)−𝜎2
545
+ 𝜇(𝛽0
546
+ 1,𝛽0
547
+ 2−𝑤)−𝜎2
548
+ ����
549
+ Proof. By using (5) and (1) for any (𝛽1, 𝛽2) = (𝛽0
550
+ 1, 𝛽0
551
+ 2), we
552
+ have the following
553
+ 𝜇(𝛽0
554
+ 1, 𝛽0
555
+ 2) − 𝜎2 =
556
+ �����
557
+
558
+ 𝑃
559
+ �√
560
+ 𝐹𝜆𝑒− 𝑗𝑘0𝑑0
561
+ 4𝜋𝑑0
562
+
563
+ 𝐿𝑥𝐿𝑦sinc
564
+
565
+ 𝐾𝑥(𝛼1 − 𝛽0
566
+ 1)
567
+
568
+ ×sinc
569
+
570
+ 𝐾𝑦(𝛼2 − 𝛽0
571
+ 2)
572
+ ������
573
+ 2
574
+ .
575
+ (10)
576
+ For (𝛽1, 𝛽2) = (𝛽0
577
+ 1 +𝑣, 𝛽0
578
+ 2), where 𝑣 is any arbitrary parameter
579
+ such that 𝐾𝑥𝑣 ∈ N and
580
+ ��𝛽0
581
+ 1 ±𝑣
582
+ �� ≤ 1 holds, we have
583
+ 𝜇(𝛽0
584
+ 1 +𝑣, 𝛽0
585
+ 2) − 𝜎2 =
586
+ �����
587
+
588
+ 𝑃
589
+ �√
590
+ 𝐹𝜆𝑒−𝑗𝑘0𝑑0
591
+ 4𝜋𝑑0
592
+
593
+ 𝐿𝑥𝐿𝑦
594
+ ×sinc
595
+
596
+ 𝐾𝑦(𝛼2 − 𝛽0
597
+ 2)
598
+ ������
599
+ 2
600
+ .
601
+ (11)
602
+ Dividing (10) by (11), we obtain
603
+ 𝜇(𝛽0
604
+ 1, 𝛽0
605
+ 2) − 𝜎2
606
+ 𝜇(𝛽0
607
+ 1 +𝑣, 𝛽0
608
+ 2) − 𝜎2 =
609
+ ����sinc
610
+
611
+ 𝐾𝑥(𝛼1 − 𝛽0
612
+ 1)
613
+ �����
614
+ 2
615
+ ����sinc
616
+
617
+ 𝐾𝑥(𝛼1 − 𝛽0
618
+ 1 −𝑣)
619
+ �����
620
+ 2
621
+ 𝜇(𝛽0
622
+ 1, 𝛽0
623
+ 2) − 𝜎2
624
+ 𝜇(𝛽0
625
+ 1 +𝑣, 𝛽0
626
+ 2) − 𝜎2 =
627
+ ����
628
+ sin(𝐾𝑥 𝜋(𝛼1−𝛽0
629
+ 1))
630
+ 𝐾𝑥 𝜋(𝛼1−𝛽0
631
+ 1)
632
+ ����
633
+ 2
634
+ ����
635
+ sin(𝐾𝑥 𝜋(𝛼1−𝛽0
636
+ 1−𝑣))
637
+ 𝐾𝑥 𝜋(𝛼1−𝛽0
638
+ 1−𝑣)
639
+ ����
640
+ 2 .
641
+ (12)
642
+ If
643
+ 𝑣
644
+ is
645
+ selected
646
+ such
647
+ that
648
+ 𝐾𝑥𝑣 ∈ N,
649
+ then
650
+ we
651
+ have
652
+ ����sin
653
+
654
+ 𝐾𝑥𝜋(𝛼1 − 𝛽0
655
+ 1 ±𝑣)
656
+ ����� =
657
+ ����sin
658
+
659
+ 𝐾𝑥𝜋(𝛼1 − 𝛽0
660
+ 1)
661
+ �����. As a result,
662
+ (12) is simplified to
663
+ 𝜇(𝛽0
664
+ 1, 𝛽0
665
+ 2) − 𝜎2
666
+ 𝜇(𝛽0
667
+ 1 +𝑣, 𝛽0
668
+ 2) − 𝜎2 =
669
+ �����
670
+ 𝛼1 − 𝛽0
671
+ 1 −𝑣
672
+ 𝛼1 − 𝛽0
673
+ 1
674
+ �����
675
+ 2
676
+ .
677
+ (13)
678
+ Since
679
+ 𝜇(𝛽1, 𝛽2) ≥ 𝜎2, it follows that
680
+ 𝜇(𝛽1, 𝛽2) − 𝜎2 =
681
+ ��𝜇(𝛽1, 𝛽2) − 𝜎2�� always holds, for all (𝛽1, 𝛽2) ∈ B. Using this
682
+ fact, (13) can be written equivalently as
683
+
684
+
685
+ ������
686
+ 𝜇(𝛽0
687
+ 1, 𝛽0
688
+ 2) − 𝜎2
689
+ 𝜇(𝛽0
690
+ 1 +𝑣, 𝛽0
691
+ 2) − 𝜎2
692
+ ����� =
693
+ �����
694
+ 𝛼1 − 𝛽0
695
+ 1 −𝑣
696
+ 𝛼1 − 𝛽0
697
+ 1
698
+ �����.
699
+ (14)
700
+ By solving the nonlinear equation in (14) w.r.t. the unknown
701
+ 𝛼1, we obtain two solutions for 𝛼1, denoted by 𝛼(1)
702
+ 1
703
+ and 𝛼(2)
704
+ 1 ,
705
+ given by
706
+ 𝛼(1)/(2)
707
+ 1
708
+ = 𝛽0
709
+ 1 +
710
+ 𝑣
711
+
712
+ √︂���
713
+ 𝜇(𝛽0
714
+ 1,𝛽0
715
+ 2)−𝜎2
716
+ 𝜇(𝛽0
717
+ 1+𝑣,𝛽0
718
+ 2)−𝜎2
719
+ ���
720
+ .
721
+ (15)
722
+ It is not known which of the two values 𝛼(1)
723
+ 1
724
+ and 𝛼(2)
725
+ 1
726
+ is
727
+ equal to 𝛼1. To identify the correct solution for 𝛼1 of the
728
+ two solutions given by (15), we need the value of 𝜇(𝛽1, 𝛽2)
729
+ for (𝛽1, 𝛽2) = (𝛽0
730
+ 1 − 𝑣, 𝛽0
731
+ 2). Following the same procedure as
732
+ for (10)-(15), but now by using the values of 𝜇(𝛽1, 𝛽2) for
733
+ (𝛽1, 𝛽2) = (𝛽0
734
+ 1, 𝛽0
735
+ 2) and (𝛽1, 𝛽2) = (𝛽0
736
+ 1 −𝑣, 𝛽0
737
+ 2), we obtain
738
+
739
+
740
+ ������
741
+ 𝜇(𝛽0
742
+ 1, 𝛽0
743
+ 2) − 𝜎2
744
+ 𝜇(𝛽0
745
+ 1 −𝑣, 𝛽0
746
+ 2) − 𝜎2
747
+ ����� =
748
+ �����
749
+ 𝛼1 − 𝛽0
750
+ 1 +𝑣
751
+ 𝛼1 − 𝛽0
752
+ 1
753
+ �����.
754
+ (16)
755
+ By solving (16), we obtain
756
+ 𝛼(3)/(4)
757
+ 1
758
+ = 𝛽0
759
+ 1 −
760
+ 𝑣
761
+
762
+ √︂���
763
+ 𝜇(𝛽0
764
+ 1,𝛽0
765
+ 2)−𝜎2
766
+ 𝜇(𝛽0
767
+ 1−𝑣,𝛽0
768
+ 2)−𝜎2
769
+ ���
770
+ .
771
+ (17)
772
+ One of the solutions in (15) is identical to one of the solutions
773
+ in (17). Therefore, using (15) and (17), the correct solution of
774
+ 𝛼1 can be obtained as1
775
+ 𝛼1 =
776
+
777
+ 𝛼(𝑖)
778
+ 1 +𝛼( 𝑗)
779
+ 1
780
+ 2
781
+ :
782
+ min
783
+ 𝑖∈{1,2}, 𝑗 ∈{3,4}
784
+ ���𝛼(𝑖)
785
+ 1 −𝛼( 𝑗)
786
+ 1
787
+ ���
788
+
789
+ .
790
+ (18)
791
+ In order to obtain 𝛼2, we need the value of 𝜇(𝛽1, 𝛽2) for
792
+ (𝛽1, 𝛽2) = (𝛽0
793
+ 1, 𝛽0
794
+ 2), which we already have, and for (𝛽1, 𝛽2) =
795
+ (𝛽0
796
+ 1, 𝛽0
797
+ 2 +𝑤), where 𝑤 is selected such that 𝐾𝑦𝑤 ∈ N,
798
+ ��𝛽0
799
+ 2 ± 𝑤
800
+ �� ≤
801
+ 1 and
802
+ ��sin(𝐾𝑦𝜋(𝛼2 − 𝛽0
803
+ 2 ± 𝑤))
804
+ �� =
805
+ ��sin(𝐾𝑦𝜋(𝛼2 − 𝛽0
806
+ 2))
807
+ ��. Then,
808
+ similar to (10)-(14), we use the values of 𝜇(𝛽1, 𝛽2) for
809
+ (𝛽1, 𝛽2) = (𝛽0
810
+ 1, 𝛽0
811
+ 2) and (𝛽1, 𝛽2) = (𝛽0
812
+ 1, 𝛽0
813
+ 2 + 𝑤) to obtain
814
+ 1Note
815
+ that
816
+ 𝛼1
817
+ can
818
+ also
819
+ be
820
+ written
821
+ equivalently
822
+ as
823
+ 𝛼1 =
824
+
825
+ 𝛼(1)
826
+ 1
827
+ , 𝛼(2)
828
+ 1
829
+ � � �
830
+ 𝛼(3)
831
+ 1
832
+ , 𝛼(4)
833
+ 1
834
+
835
+ . However, the expression in (18) is more
836
+ convenient for the case when the values of 𝜇(𝛽1, 𝛽2) need to be estimated.
837
+
838
+ 5
839
+
840
+
841
+ ������
842
+ 𝜇(𝛽0
843
+ 1, 𝛽0
844
+ 2) − 𝜎2
845
+ 𝜇(𝛽0
846
+ 1, 𝛽0
847
+ 2 + 𝑤) − 𝜎2
848
+ ����� =
849
+ �����
850
+ 𝛼2 − 𝛽0
851
+ 2 − 𝑤
852
+ 𝛼2 − 𝛽0
853
+ 2
854
+ �����.
855
+ (19)
856
+ By solving the nonlinear equation (19), we obtain two solutions
857
+ for 𝛼2, denoted by 𝛼(1)
858
+ 2
859
+ and 𝛼(2)
860
+ 2 , given by
861
+ 𝛼(1)/(2)
862
+ 2
863
+ = 𝛽0
864
+ 2 +
865
+ 𝑤
866
+
867
+ √︂���
868
+ 𝜇(𝛽0
869
+ 1,𝛽0
870
+ 2)−𝜎2
871
+ 𝜇(𝛽0
872
+ 1,𝛽0
873
+ 2+𝑤)−𝜎2
874
+ ���
875
+ .
876
+ (20)
877
+ To identify the correct solution for 𝛼2 of the two given in
878
+ (20), we need the value of 𝜇(𝛽1, 𝛽2) for (𝛽1, 𝛽2) = (𝛽0
879
+ 1, 𝛽0
880
+ 2 −𝑤).
881
+ Again, following the procedure from (10)-(15), by using the
882
+ values of 𝜇(𝛽1, 𝛽2) for (𝛽0
883
+ 1, 𝛽0
884
+ 2) and (𝛽0
885
+ 1, 𝛽0
886
+ 2 − 𝑤), we obtain
887
+ 𝛼(3)/(4)
888
+ 2
889
+ = 𝛽0
890
+ 2 −
891
+ 𝑤
892
+
893
+ √︂���
894
+ 𝜇(𝛽0
895
+ 1,𝛽0
896
+ 2)−𝜎2
897
+ 𝜇(𝛽0
898
+ 1,𝛽0
899
+ 2−𝑤)−𝜎2
900
+ ���
901
+ .
902
+ (21)
903
+ One of the solutions in (20) is exactly same as the solutions
904
+ of (21). Therefore, using (20) and (21), the correct solution of
905
+ 𝛼2 can be obtained as
906
+ 𝛼2 =
907
+
908
+ 𝛼(𝑖)
909
+ 2 +𝛼( 𝑗)
910
+ 2
911
+ 2
912
+ :
913
+ min
914
+ 𝑖∈{1,2}, 𝑗 ∈{3,4}
915
+ ���𝛼(𝑖)
916
+ 2 −𝛼( 𝑗)
917
+ 2
918
+ ���
919
+
920
+ .
921
+ (22)
922
+ Finally, by setting 𝛽∗
923
+ 1 = 𝛼1 and 𝛽∗
924
+ 2 = 𝛼2, where 𝛼1 and 𝛼2 are
925
+ given by (18) and (22), respectively, we obtain (8) and (9).
926
+
927
+ Remark 2. In [13, Sec. IV.A], the authors proposed the channel
928
+ estimation strategy under the assumption that there is no noise
929
+ in the system. However, in the noisy case, we proposed an
930
+ estimation scheme based on the assumption that 𝜇(𝛽1, 𝛽2) for
931
+ any of the phase-shifting parameters (𝛽1, 𝛽2) ∈ B are perfectly
932
+ known at the HMT.
933
+ However, in practice the exact values of 𝜇(𝛽1, 𝛽2) for any of
934
+ the phase-shifting parameters (𝛽1, 𝛽2) ∈ B cannot be known in
935
+ advance at the HMT, and therefore they need to be estimated
936
+ using pilot symbols. In the following, we propose an algorithm
937
+ that estimates 𝜇(𝛽1, 𝛽2) for the phase-shifting parameters in
938
+ B and then uses the estimated values of 𝜇(𝛽1, 𝛽2) to find the
939
+ optimal phase-shifting parameters (𝛽∗
940
+ 1, 𝛽∗
941
+ 2) in the presence of
942
+ noise.
943
+ C. Estimation Of The Optimal Phase-Shifting Parameters In
944
+ The Noisy Case
945
+ The user sends in total 𝑁 number of pilot signals to the
946
+ HMT for the estimation of the five values of 𝜇(𝛽1, 𝛽2) for the
947
+ five pairs of (𝛽1, 𝛽2) ∈ B. As a result, the proposed algorithm
948
+ works in five epochs. In the 𝑘𝑡ℎ epoch, for 𝑘 = 1,2,..., the user
949
+ transmits
950
+ � 𝑁
951
+ 5
952
+ � number of pilots to the HMT. The HMT sets
953
+ (𝛽1, 𝛽2) to the 𝑘𝑡ℎ element in B, and collects
954
+ � 𝑁
955
+ 5
956
+ � samples
957
+ of the received signal squared, given by (3). Then 𝜇(𝛽1, 𝛽2),
958
+ for (𝛽1, 𝛽2) being the 𝑘𝑡ℎ elements in B, is estimated as
959
+ ˆ𝜇(𝛽1, 𝛽2) =
960
+ 1
961
+ ⌊𝑁/5⌋
962
+ ⌊𝑁 /5⌋
963
+ ∑︁
964
+ 𝑖=1
965
+ 𝑟𝑖(𝛽1, 𝛽2),
966
+ (23)
967
+ where 𝑟𝑖(𝛽1, 𝛽2) is the 𝑖𝑡ℎ sample of 𝑟(𝛽1, 𝛽2) in (3).
968
+ Next, we replace 𝜇(𝛽1, 𝛽2) in (15), (17), (20), and (21)
969
+ by ˆ𝜇(𝛽1, 𝛽2), ∀(𝛽1, 𝛽2) ∈ B, and thereby obtain our estimates
970
+ for 𝛽∗
971
+ 1 and 𝛽∗
972
+ 2, denoted by ˆ𝛽∗
973
+ 1 and ˆ𝛽∗
974
+ 2. The pseudo-code of
975
+ the proposed algorithm is given in Two-Stage Phase-Shifts
976
+ Estimation Algorithm below. We note that the choice of the
977
+ Two-Stage Phase-Shifts Estimation Algorithm
978
+ 1: Input: 𝑁,B,𝜎2.
979
+ 2: ***Stage 1: Uniform Exploration ***
980
+ 3: for 𝑘 = 1 to 5 do
981
+ 4:
982
+ HMT sets (𝛽1, 𝛽2) to the 𝑘𝑡ℎ pair in B.
983
+ 5:
984
+ User sends ⌊𝑁/5⌋ number of pilots to the HMT.
985
+ 6:
986
+ For the 𝑖𝑡ℎ pilot, the HMT receives 𝑟𝑖(𝛽1, 𝛽2), given by
987
+ (3), for 𝑖 = 1,2,..., ⌊𝑁/5⌋ .
988
+ 7:
989
+ The HMT computes ˆ𝜇𝑘 (𝛽1, 𝛽2) using (23).
990
+ 8: end for
991
+ 9: ***Stage
992
+ 2:
993
+ Estimate
994
+ Optimal
995
+ Phase-Shifting
996
+ Parameters***
997
+ 10: Obtain ˆ𝛽∗
998
+ 1 as
999
+ ˆ𝛽∗
1000
+ 1 =
1001
+
1002
+ ˆ𝛼(𝑖)
1003
+ 1 + ˆ𝛼( 𝑗)
1004
+ 1
1005
+ 2
1006
+ :
1007
+ min
1008
+ 𝑖∈{1,2}, 𝑗 ∈{3,4}
1009
+ ��� ˆ𝛼(𝑖)
1010
+ 1 − ˆ𝛼( 𝑗)
1011
+ 1
1012
+ ���
1013
+
1014
+ ,
1015
+ (24)
1016
+ where ˆ𝛼(1)/(2)
1017
+ 1
1018
+ is obtained by replacing the value of
1019
+ 𝜇(𝛽1, 𝛽2) by ˆ𝜇(𝛽1, 𝛽2) in (15), and ˆ𝛼(3)/(4)
1020
+ 1
1021
+ is obtained
1022
+ by replacing the value of 𝜇(𝛽1, 𝛽2) by ˆ𝜇(𝛽1, 𝛽2) in (17).
1023
+ 11: Obtain ˆ𝛽∗
1024
+ 2 as
1025
+ ˆ𝛽∗
1026
+ 2 =
1027
+
1028
+ ˆ𝛼(𝑖)
1029
+ 2 + ˆ𝛼( 𝑗)
1030
+ 2
1031
+ 2
1032
+ :
1033
+ min
1034
+ 𝑖∈{1,2}, 𝑗 ∈{3,4}
1035
+ ��� ˆ𝛼(𝑖)
1036
+ 2 − ˆ𝛼( 𝑗)
1037
+ 2
1038
+ ���
1039
+
1040
+ ,
1041
+ (25)
1042
+ where ˆ𝛼(1)/(2)
1043
+ 2
1044
+ is obtained by replacing the value of
1045
+ 𝜇(𝛽1, 𝛽2) by ˆ𝜇(𝛽1, 𝛽2) in (20), and ˆ𝛼(3)/(4)
1046
+ 2
1047
+ is obtained
1048
+ by replacing the value of 𝜇(𝛽1, 𝛽2) by ˆ𝜇(𝛽1, 𝛽2) in (21).
1049
+ 12: Output: ˆ𝛽∗
1050
+ 1 and ˆ𝛽∗
1051
+ 2.
1052
+ 13: Phase-shifts at HMT Set the phase-shift of the (𝑚𝑥,𝑚𝑦)𝑡ℎ
1053
+ element at the HMT to
1054
+ 𝛽𝑚𝑥𝑚𝑦 = −
1055
+ mod (𝑘0𝑑𝑟 (𝑚𝑥 ˆ𝛽∗
1056
+ 1 +𝑚𝑦 ˆ𝛽∗
1057
+ 2),2𝜋).
1058
+ initial (𝛽0
1059
+ 1, 𝛽0
1060
+ 2) in the set B was arbitrary. The values of (𝛽0
1061
+ 1, 𝛽0
1062
+ 2)
1063
+ can effect the estimation error. In general, if the values (𝛽0
1064
+ 1, 𝛽0
1065
+ 2)
1066
+ are closer to the (𝛼1,𝛼2), the better the estimation will be. A
1067
+ good choice for (𝛽0
1068
+ 1, 𝛽0
1069
+ 2) is given in [13, Sec. V.C], which leads
1070
+ to faster learning of (𝛼1,𝛼2).
1071
+ IV. THEORETICAL GUARANTEES FOR THE PROPOSED
1072
+ ALGORITHM
1073
+ In the section, we bound the probability that the estimates,
1074
+ obtained from the proposed Two-Stage Phase-Shifts Estimation
1075
+ Algorithm Algorithm, deviate from the true values of (𝛼1,𝛼2)
1076
+ by an amount 0 ≤ 𝜖 ≤ 1. In particular, we upper bound the
1077
+ following error probability
1078
+ P
1079
+ ��
1080
+ ˆ𝛽∗
1081
+ 1 −𝛼1
1082
+ �2
1083
+ +
1084
+
1085
+ ˆ𝛽∗
1086
+ 2 −𝛼2
1087
+ �2
1088
+ ≥ 𝜖
1089
+
1090
+ .
1091
+ (26)
1092
+
1093
+ 6
1094
+ We use the following results to upper bound the error probability
1095
+ in (26).
1096
+ Lemma 1. Let {𝑋𝑛} be a sequence of random variables (RVs)
1097
+ on a probability space. Let 𝑋 be a RV defined on the same
1098
+ probability space. Then, the following holds
1099
+ P{|𝑋𝑛 − 𝑋𝑚| ≥ 𝜖} ≤ P
1100
+
1101
+ |𝑋𝑛 − 𝑋| ≥ 𝜖
1102
+ 2
1103
+
1104
+ +P
1105
+
1106
+ |𝑋𝑚 − 𝑋| ≥ 𝜖
1107
+ 2
1108
+
1109
+ .
1110
+ Proof. The proof is given in the Appendix A.
1111
+
1112
+ Let 𝜒2
1113
+ 𝑝(𝜆) denote a non-central Chi-squared distribution with
1114
+ 𝑝 degrees of freedom and non-centrality parameter 𝜆.
1115
+ Lemma 2. Let 𝑋 =
1116
+ 2
1117
+ 𝜎2 𝑟(𝛽1, 𝛽2), where 𝑟(𝛽1, 𝛽2) is given by
1118
+ (3), and let 𝜆1 =
1119
+ 2
1120
+ 𝜎2
1121
+ ���
1122
+
1123
+ 𝑃𝐻(𝛽1, 𝛽2)
1124
+ ���
1125
+ 2
1126
+ . Then, 𝑋 is distributed as
1127
+ 𝜒2
1128
+ 2(𝜆1), i.e., 𝑋 ∼ 𝜒2
1129
+ 2(𝜆1). Furthermore, if 𝑋𝑖 for 𝑖 = 1,2,...,𝑛
1130
+ are 𝑛 independently and identically distributed (i.i.d.) RVs of
1131
+ 𝜒2
1132
+ 2(𝜆1), then
1133
+ 𝑛
1134
+ ∑︁
1135
+ 𝑖=1
1136
+ 𝑋𝑖 ∼ 𝜒2
1137
+ 2𝑛(𝑛𝜆1).
1138
+ Proof. The proof is given in the Appendix B.
1139
+
1140
+ The following theorem provides an upper bound on the error
1141
+ probability in (26).
1142
+ Theorem 2. Let us perform uniform exploration on the set B
1143
+ given in (7). For any 0 ≤ 𝜖 ≤ 1, the error probability in (26)
1144
+ is upper bounded as
1145
+ P
1146
+ ��
1147
+ ˆ𝛽∗
1148
+ 1 −𝛼1
1149
+ �2
1150
+ +
1151
+
1152
+ ˆ𝛽∗
1153
+ 2 −𝛼2
1154
+ �2
1155
+ ≥ 𝜖
1156
+
1157
+ ≤ 4
1158
+
1159
+ 𝑒− 𝑛
1160
+ 32
1161
+ � 𝜖 𝜆2
1162
+ 1+𝜆2
1163
+ �2
1164
+ + 𝑒− 𝑛
1165
+ 32
1166
+ � 𝜖 𝜆3
1167
+ 1+𝜆3
1168
+ �2
1169
+ + 𝑒− 𝑛
1170
+ 32
1171
+ � 𝜖 𝜆4
1172
+ 1+𝜆4
1173
+ �2
1174
+ + 𝑒− 𝑛
1175
+ 32
1176
+ � 𝜖 𝜆5
1177
+ 1+𝜆5
1178
+ �2�
1179
+ ,
1180
+ (27)
1181
+ where
1182
+ 𝜆1 =
1183
+ 2
1184
+ ���
1185
+
1186
+ 𝑃𝐻(𝛽0
1187
+ 1, 𝛽0
1188
+ 2)
1189
+ ���
1190
+ 2
1191
+ 𝜎2
1192
+ ,
1193
+ 𝜆2 =
1194
+ 2
1195
+ ���
1196
+
1197
+ 𝑃𝐻(𝛽0
1198
+ 1 +𝑣, 𝛽0
1199
+ 2)
1200
+ ���
1201
+ 2
1202
+ 𝜎2
1203
+ ,
1204
+ 𝜆3 =
1205
+ 2
1206
+ ���
1207
+
1208
+ 𝑃𝐻(𝛽0
1209
+ 1 −𝑣, 𝛽0
1210
+ 2)
1211
+ ���
1212
+ 2
1213
+ 𝜎2
1214
+ ,
1215
+ 𝜆4 =
1216
+ 2
1217
+ ���
1218
+
1219
+ 𝑃𝐻(𝛽0
1220
+ 1, 𝛽0
1221
+ 2 + 𝑤)
1222
+ ���
1223
+ 2
1224
+ 𝜎2
1225
+ ,
1226
+ 𝜆5 =
1227
+ 2
1228
+ ���
1229
+
1230
+ 𝑃𝐻(𝛽0
1231
+ 1, 𝛽0
1232
+ 2 − 𝑤)
1233
+ ���
1234
+ 2
1235
+ 𝜎2
1236
+ .
1237
+ Proof. Let us denote the estimate of 𝜇(𝛽0
1238
+ 1, 𝛽0
1239
+ 2) by ˆ𝜇(𝛽0
1240
+ 1, 𝛽0
1241
+ 2)
1242
+ which is given by
1243
+ ˆ𝜇(𝛽0
1244
+ 1, 𝛽0
1245
+ 2) = 1
1246
+ 𝑛
1247
+ 𝑛
1248
+ ∑︁
1249
+ 𝑖=1
1250
+ 𝑟𝑖(𝛽0
1251
+ 1, 𝛽0
1252
+ 2) = 𝜎2
1253
+ 2𝑛
1254
+ 𝑛
1255
+ ∑︁
1256
+ 𝑖=1
1257
+ 𝑋𝑖.
1258
+ Using Lemma 2, we have
1259
+ ˆ𝜇1 := 2𝑛
1260
+ 𝜎2 ˆ𝜇(𝛽0
1261
+ 1, 𝛽0
1262
+ 2) ∼ 𝜒2
1263
+ 2𝑛(𝑛𝜆1)
1264
+ (28)
1265
+ ˆ𝜇2 := 2𝑛
1266
+ 𝜎2 ˆ𝜇(𝛽0
1267
+ 1 +𝑣, 𝛽0
1268
+ 2) ∼ 𝜒2
1269
+ 2𝑛(𝑛𝜆2)
1270
+ (29)
1271
+ ˆ𝜇3 := 2𝑛
1272
+ 𝜎2 ˆ𝜇(𝛽0
1273
+ 1 −𝑣, 𝛽0
1274
+ 2) ∼ 𝜒2
1275
+ 2𝑛(𝑛𝜆3)
1276
+ (30)
1277
+ ˆ𝜇4 := 2𝑛
1278
+ 𝜎2 ˆ𝜇(𝛽0
1279
+ 1, 𝛽0
1280
+ 2 + 𝑤) ∼ 𝜒2
1281
+ 2𝑛(𝑛𝜆4)
1282
+ (31)
1283
+ ˆ𝜇5 := 2𝑛
1284
+ 𝜎2 ˆ𝜇(𝛽0
1285
+ 1, 𝛽0
1286
+ 2 − 𝑤) ∼ 𝜒2
1287
+ 2𝑛(𝑛𝜆5)
1288
+ (32)
1289
+ where 𝜆1,𝜆2,𝜆3,𝜆4 and 𝜆5 is given in Theorem 2.
1290
+ The random variables ˆ𝜇1, ˆ𝜇2, ˆ𝜇3, ˆ𝜇4, and ˆ𝜇5 are mutually
1291
+ independent, since they are sampled at different epochs. The
1292
+ estimated optimal phase-shifting parameters ( ˆ𝛽∗
1293
+ 1, ˆ𝛽∗
1294
+ 2), are given
1295
+ by (24) and (25), where the values of ˆ𝛼(1)
1296
+ 1 , ˆ𝛼(2)
1297
+ 1 , ˆ𝛼(3)
1298
+ 1 , ˆ𝛼(4)
1299
+ 1 ,
1300
+ and, ˆ𝛼(1)
1301
+ 2 , ˆ𝛼(2)
1302
+ 2 , ˆ𝛼(3)
1303
+ 2 , and ˆ𝛼(4)
1304
+ 2
1305
+ are given by
1306
+ ˆ𝛼(1)/(2)
1307
+ 1
1308
+ = 𝛽0
1309
+ 1 +
1310
+ 𝑣
1311
+
1312
+ √︄����
1313
+ ˆ𝜇(𝛽0
1314
+ 1,𝛽0
1315
+ 2)−𝜎2
1316
+ ˆ𝜇(𝛽0
1317
+ 1+𝑣,𝛽0
1318
+ 2)−𝜎2
1319
+ ����
1320
+ (33)
1321
+ ˆ𝛼(3)/(4)
1322
+ 1
1323
+ = 𝛽0
1324
+ 1 −
1325
+ 𝑣
1326
+
1327
+ √︄����
1328
+ ˆ𝜇(𝛽0
1329
+ 1,𝛽0
1330
+ 2)−𝜎2
1331
+ ˆ𝜇(𝛽0
1332
+ 1−𝑣,𝛽0
1333
+ 2)−𝜎2
1334
+ ����
1335
+ (34)
1336
+ ˆ𝛼(1)/(2)
1337
+ 2
1338
+ = 𝛽0
1339
+ 2 +
1340
+ 𝑤
1341
+
1342
+ √︄����
1343
+ ˆ𝜇(𝛽0
1344
+ 1,𝛽0
1345
+ 2)−𝜎2
1346
+ ˆ𝜇(𝛽0
1347
+ 1,𝛽0
1348
+ 2+𝑤)−𝜎2
1349
+ ����
1350
+ (35)
1351
+ ˆ𝛼(3)/(4)
1352
+ 2
1353
+ = 𝛽0
1354
+ 2 −
1355
+ 𝑤
1356
+
1357
+ √︄����
1358
+ ˆ𝜇(𝛽0
1359
+ 1,𝛽0
1360
+ 2)−𝜎2
1361
+ ˆ𝜇(𝛽0
1362
+ 1,𝛽0
1363
+ 2−𝑤)−𝜎2
1364
+ ����
1365
+ .
1366
+ (36)
1367
+ By inserting (28), (29), (30), (31), and (32) into (33), (34),
1368
+ (35) and (36), we obtain
1369
+ ˆ𝛼(1)/(2)
1370
+ 1
1371
+ = 𝛽0
1372
+ 1 +
1373
+ 𝑣
1374
+
1375
+ √︂��� ˆ𝜇1−2𝑛
1376
+ ˆ𝜇2−2𝑛
1377
+ ���
1378
+ (37)
1379
+ ˆ𝛼(3)/(4)
1380
+ 1
1381
+ = 𝛽0
1382
+ 1 −
1383
+ 𝑣
1384
+
1385
+ √︂��� ˆ𝜇1−2𝑛
1386
+ ˆ𝜇3−2𝑛
1387
+ ���
1388
+ (38)
1389
+ ˆ𝛼(1)/(2)
1390
+ 2
1391
+ = 𝛽0
1392
+ 2 +
1393
+ 𝑤
1394
+
1395
+ √︂��� ˆ𝜇1−2𝑛
1396
+ ˆ𝜇4−2𝑛
1397
+ ���
1398
+ (39)
1399
+ ˆ𝛼(3)/(4)
1400
+ 2
1401
+ = 𝛽0
1402
+ 2 −
1403
+ 𝑤
1404
+
1405
+ √︂��� ˆ𝜇1−2𝑛
1406
+ ˆ𝜇5−2𝑛
1407
+ ���
1408
+ .
1409
+ (40)
1410
+ Let us denote
1411
+ 𝐼 := P
1412
+ ��� ˆ𝛽∗
1413
+ 1 −𝛼1
1414
+ �� ≥
1415
+ √︂
1416
+ 𝜖
1417
+ 2
1418
+
1419
+ 𝐼𝐼 := P
1420
+ ��� ˆ𝛽∗
1421
+ 2 −𝛼2
1422
+ �� ≥
1423
+ √︂
1424
+ 𝜖
1425
+ 2
1426
+
1427
+ .
1428
+ Now, applying Lemma 1 in (26), we obtain
1429
+ P
1430
+ ��
1431
+ ˆ𝛽∗
1432
+ 1 −𝛼1
1433
+ �2
1434
+ +
1435
+
1436
+ ˆ𝛽∗
1437
+ 2 −𝛼2
1438
+ �2
1439
+ ≥ 𝜖
1440
+
1441
+ ≤ P
1442
+ ��
1443
+ ˆ𝛽∗
1444
+ 1 −𝛼1
1445
+ �2
1446
+ ≥ 𝜖
1447
+ 2
1448
+
1449
+ +P
1450
+ ��
1451
+ ˆ𝛽∗
1452
+ 2 −𝛼2
1453
+ �2
1454
+ ≥ 𝜖
1455
+ 2
1456
+
1457
+ ≤ 𝐼 + 𝐼𝐼.
1458
+ (41)
1459
+ We upper bound each of the term in right-hand side of (41).
1460
+ We begin with the first term P
1461
+ ��� ˆ𝛽∗
1462
+ 1 −𝛼1
1463
+ �� ≥ √︁ 𝜖
1464
+ 2
1465
+
1466
+ , denoted as I.
1467
+ G Step 1: Upper bound on I
1468
+ From (24), we have
1469
+
1470
+ 7
1471
+ P
1472
+ ��� ˆ𝛽∗
1473
+ 1 −𝛼1
1474
+ �� ≥
1475
+ √︂
1476
+ 𝜖
1477
+ 2
1478
+
1479
+ = P
1480
+ � ������
1481
+ ˆ𝛼(1)
1482
+ 1
1483
+ + ˆ𝛼(3)
1484
+ 1
1485
+ 2
1486
+ −𝛼1
1487
+ ����� ≥
1488
+ √︂
1489
+ 𝜖
1490
+ 2
1491
+
1492
+ � ������
1493
+ ˆ𝛼(1)
1494
+ 1
1495
+ + ˆ𝛼(4)
1496
+ 1
1497
+ 2
1498
+ −𝛼1
1499
+ ����� ≥
1500
+ √︂
1501
+ 𝜖
1502
+ 2
1503
+
1504
+ � ������
1505
+ ˆ𝛼(2)
1506
+ 1
1507
+ + ˆ𝛼(3)
1508
+ 1
1509
+ 2
1510
+ −𝛼1
1511
+ ����� ≥
1512
+ √︂
1513
+ 𝜖
1514
+ 2
1515
+
1516
+ � ������
1517
+ ˆ𝛼(2)
1518
+ 1
1519
+ + ˆ𝛼(4)
1520
+ 1
1521
+ 2
1522
+ −𝛼1
1523
+ ����� ≥
1524
+ √︂
1525
+ 𝜖
1526
+ 2
1527
+ � �
1528
+ ≤ P
1529
+ ������
1530
+ ˆ𝛼(1)
1531
+ 1
1532
+ + ˆ𝛼(3)
1533
+ 1
1534
+ 2
1535
+ −𝛼1
1536
+ ����� ≥
1537
+ √︂
1538
+ 𝜖
1539
+ 2
1540
+
1541
+ +P
1542
+ ������
1543
+ ˆ𝛼(1)
1544
+ 1
1545
+ + ˆ𝛼(4)
1546
+ 1
1547
+ 2
1548
+ −𝛼1
1549
+ ����� ≥
1550
+ √︂
1551
+ 𝜖
1552
+ 2
1553
+
1554
+ +P
1555
+ ������
1556
+ ˆ𝛼(2)
1557
+ 1
1558
+ + ˆ𝛼(3)
1559
+ 1
1560
+ 2
1561
+ −𝛼1
1562
+ ����� ≥
1563
+ √︂
1564
+ 𝜖
1565
+ 2
1566
+
1567
+ +P
1568
+ ������
1569
+ ˆ𝛼(2)
1570
+ 1
1571
+ + ˆ𝛼(4)
1572
+ 1
1573
+ 2
1574
+ −𝛼1
1575
+ ����� ≥
1576
+ √︂
1577
+ 𝜖
1578
+ 2
1579
+
1580
+ =
1581
+ ∑︁
1582
+ 𝑖=1,2
1583
+ 𝑗=3,4
1584
+ P
1585
+ �����
1586
+
1587
+ ˆ𝛼(𝑖)
1588
+ 1 −𝛼1
1589
+
1590
+ +
1591
+
1592
+ ˆ𝛼( 𝑗)
1593
+ 1
1594
+ −𝛼1
1595
+ ����� ≥ 2
1596
+ √︂
1597
+ 𝜖
1598
+ 2
1599
+
1600
+
1601
+ ∑︁
1602
+ 𝑖=1,2
1603
+ 𝑗=3,4
1604
+
1605
+ P
1606
+ ���� ˆ𝛼(𝑖)
1607
+ 1 −𝛼1
1608
+ ��� ≥
1609
+ √︂
1610
+ 𝜖
1611
+ 2
1612
+
1613
+ +P
1614
+ ���� ˆ𝛼( 𝑗)
1615
+ 1
1616
+ −𝛼1
1617
+ ��� ≥
1618
+ √︂
1619
+ 𝜖
1620
+ 2
1621
+ � �
1622
+ = 2
1623
+
1624
+ P
1625
+ ���� ˆ𝛼(1)
1626
+ 1
1627
+ −𝛼1
1628
+ ��� ≥
1629
+ √︂
1630
+ 𝜖
1631
+ 2
1632
+
1633
+ +P
1634
+ ���� ˆ𝛼(2)
1635
+ 1
1636
+ −𝛼1
1637
+ ��� ≥
1638
+ √︂
1639
+ 𝜖
1640
+ 2
1641
+
1642
+ +P
1643
+ ���� ˆ𝛼(3)
1644
+ 1
1645
+ −𝛼1
1646
+ ��� ≥
1647
+ √︂
1648
+ 𝜖
1649
+ 2
1650
+
1651
+ +P
1652
+ ���� ˆ𝛼(4)
1653
+ 1
1654
+ −𝛼1
1655
+ ��� ≥
1656
+ √︂
1657
+ 𝜖
1658
+ 2
1659
+ � �
1660
+ ,
1661
+ (42)
1662
+ where we applied the union bound to get the first inequality
1663
+ and applied Lemma 1 for the second inequality. We now
1664
+ bound each term in (42) separately.
1665
+ ® Upper bound of P
1666
+ ���� ˆ𝜶(1)
1667
+ 1
1668
+ −𝜶1
1669
+ ��� ≥
1670
+ √︁ 𝝐
1671
+ 2
1672
+
1673
+ : Substituting
1674
+ the
1675
+ values
1676
+ of
1677
+ ˆ𝛼(1)
1678
+ 1 ,
1679
+ as
1680
+ given
1681
+ by
1682
+ (37),
1683
+ in
1684
+ P
1685
+ ���� ˆ𝜶(1)
1686
+ 1
1687
+ −𝜶1
1688
+ ��� ≥
1689
+ √︁ 𝝐
1690
+ 2
1691
+
1692
+ , we obtain
1693
+ P
1694
+ ���� ˆ𝛼(1)
1695
+ 1
1696
+ −𝛼1)
1697
+ ��� ≥
1698
+ √︂
1699
+ 𝜖
1700
+ 2
1701
+
1702
+ = P
1703
+ ������
1704
+ ������
1705
+ ���������
1706
+ 𝑣
1707
+ 1+
1708
+ √︂���� ˆ𝜇1−2𝑛
1709
+ ˆ𝜇2−2𝑛
1710
+ ���
1711
+ − (𝛼1 − 𝛽0
1712
+ 1)
1713
+ ���������
1714
+
1715
+ √︂
1716
+ 𝜖
1717
+ 2
1718
+ ������
1719
+ ������
1720
+ .
1721
+ (43)
1722
+ Note that the following holds.
1723
+ ���������
1724
+ 𝑣
1725
+ 1+
1726
+ √︂��� ˆ𝜇1−2𝑛
1727
+ ˆ𝜇2−2𝑛
1728
+ ���
1729
+ − (𝛼1 − 𝛽0
1730
+ 1)
1731
+ ���������
1732
+
1733
+ ���������
1734
+ 𝑣
1735
+ 1+
1736
+ √︂��� ˆ𝜇1−2𝑛
1737
+ ˆ𝜇2−2𝑛
1738
+ ���
1739
+ ���������
1740
+ +
1741
+ �����𝛼1 − 𝛽0
1742
+ 1
1743
+ �����.
1744
+ (44)
1745
+ By applying (44) in (43), we obtain
1746
+ P
1747
+ ���� ˆ𝛼(1)
1748
+ 1
1749
+ −𝛼1)
1750
+ ��� ≥
1751
+ √︂
1752
+ 𝜖
1753
+ 2
1754
+
1755
+ ≤ P
1756
+ ������
1757
+ ������
1758
+ ���������
1759
+ 1
1760
+ 1+
1761
+ √︂��� ˆ𝜇1−2𝑛
1762
+ ˆ𝜇2−2𝑛
1763
+ ���
1764
+ ���������
1765
+ ≥ 1
1766
+ 𝑣
1767
+ �√︂
1768
+ 𝜖
1769
+ 2 −
1770
+ �����𝛼1 − 𝛽0
1771
+ 1
1772
+ �����
1773
+ �������
1774
+ ������
1775
+ .
1776
+ (45)
1777
+ For the RVs ˆ𝜇1 and ˆ𝜇2,
1778
+ 1
1779
+ 1+
1780
+ √︂���
1781
+ ˆ𝜇1−2𝑛
1782
+ ˆ𝜇2−2𝑛
1783
+ ���
1784
+ is always positive.
1785
+ Using this fact in (45), we obtain
1786
+ P
1787
+ ���� ˆ𝛼(1)
1788
+ 1
1789
+ −𝛼1)
1790
+ ��� ≥
1791
+ √︂
1792
+ 𝜖
1793
+ 2
1794
+
1795
+ ≤ P
1796
+ ������
1797
+ ������
1798
+ 1
1799
+ 1+
1800
+ √︂��� ˆ𝜇1−2𝑛
1801
+ ˆ𝜇2−2𝑛
1802
+ ���
1803
+ ≥ 1
1804
+ 𝑣
1805
+ �√︂
1806
+ 𝜖
1807
+ 2 −
1808
+ �����𝛼1 − 𝛽0
1809
+ 1
1810
+ �����
1811
+ �������
1812
+ ������
1813
+ .
1814
+ Let 𝑎 = 1
1815
+ 𝑣
1816
+ �√︁ 𝜖
1817
+ 2 −
1818
+ ��𝛼1 − 𝛽0
1819
+ 1
1820
+ ��
1821
+
1822
+ . We have
1823
+ P
1824
+ ���� ˆ𝛼(1)
1825
+ 1
1826
+ −𝛼1)
1827
+ ��� ≥
1828
+ √︂
1829
+ 𝜖
1830
+ 2
1831
+
1832
+ ≤ P
1833
+ ��
1834
+ ��
1835
+ 1+
1836
+ √︄����
1837
+ ˆ𝜇1 −2𝑛
1838
+ ˆ𝜇2 −2𝑛
1839
+ ���� ≤ 1
1840
+ 𝑎
1841
+ ��
1842
+ ��
1843
+ = P
1844
+ �����
1845
+ ˆ𝜇1 −2𝑛
1846
+ ˆ𝜇2 −2𝑛
1847
+ ���� ≤
1848
+
1849
+ 1− 1
1850
+ 𝑎
1851
+ �2�
1852
+ .
1853
+ (46)
1854
+ ® Upper bound of P
1855
+ ���� ˆ𝜶(2)
1856
+ 1
1857
+ −𝜶1
1858
+ ��� ≥
1859
+ √︁ 𝝐
1860
+ 2
1861
+
1862
+ : Substituting
1863
+ the
1864
+ values
1865
+ of
1866
+ ˆ𝛼(2)
1867
+ 1 ,
1868
+ as
1869
+ given
1870
+ by
1871
+ (37),
1872
+ in
1873
+ P
1874
+ ���� ˆ𝜶(2)
1875
+ 1
1876
+ −𝜶1
1877
+ ��� ≥
1878
+ √︁ 𝝐
1879
+ 2
1880
+
1881
+ , we obtain
1882
+ P
1883
+ ���� ˆ𝛼(2)
1884
+ 1
1885
+ −𝛼1)
1886
+ ��� ≥
1887
+ √︂
1888
+ 𝜖
1889
+ 2
1890
+
1891
+ = P
1892
+ ������
1893
+ ������
1894
+ ���������
1895
+ 𝑣
1896
+ 1−
1897
+ √︂��� ˆ𝜇1−2𝑛
1898
+ ˆ𝜇2−2𝑛
1899
+ ���
1900
+ − (𝛼1 − 𝛽0
1901
+ 1)
1902
+ ���������
1903
+
1904
+ √︂
1905
+ 𝜖
1906
+ 2
1907
+ ������
1908
+ ������
1909
+ .
1910
+ (47)
1911
+ Note that the following holds.
1912
+ ���������
1913
+ 𝑣
1914
+ 1−
1915
+ √︂��� ˆ𝜇1−2𝑛
1916
+ ˆ𝜇2−2𝑛
1917
+ ���
1918
+ − (𝛼1 − 𝛽0
1919
+ 1)
1920
+ ���������
1921
+
1922
+ ���������
1923
+ 𝑣
1924
+ 1−
1925
+ √︂��� ˆ𝜇1−2𝑛
1926
+ ˆ𝜇2−2𝑛
1927
+ ���
1928
+ ���������
1929
+ +
1930
+ �����𝛼1 − 𝛽0
1931
+ 1
1932
+ �����.
1933
+ (48)
1934
+ By applying (48) in the right-hand side of (47), we
1935
+
1936
+ 8
1937
+ obtain
1938
+ P
1939
+ ���� ˆ𝛼(2)
1940
+ 1
1941
+ −𝛼1)
1942
+ ��� ≥
1943
+ √︂
1944
+ 𝜖
1945
+ 2
1946
+
1947
+ ≤ P
1948
+ ���
1949
+ ���
1950
+ ������
1951
+ 1−
1952
+ √︄����
1953
+ ˆ𝜇1 −2𝑛
1954
+ ˆ𝜇2 −2𝑛
1955
+ ����
1956
+ ������
1957
+ ≤ 1
1958
+ 𝑎
1959
+ ���
1960
+ ���
1961
+ = P
1962
+ ��
1963
+ 1− 1
1964
+ 𝑎
1965
+ �2
1966
+
1967
+ ����
1968
+ ˆ𝜇1 −2𝑛
1969
+ ˆ𝜇2 −2𝑛
1970
+ ���� ≤
1971
+
1972
+ 1+ 1
1973
+ 𝑎
1974
+ �2�
1975
+ ≤ P
1976
+ �����
1977
+ ˆ𝜇1 −2𝑛
1978
+ ˆ𝜇2 −2𝑛
1979
+ ���� ≤
1980
+
1981
+ 1+ 1
1982
+ 𝑎
1983
+ �2�
1984
+ −P
1985
+ �����
1986
+ ˆ𝜇1 −2𝑛
1987
+ ˆ𝜇2 −2𝑛
1988
+ ���� ≤
1989
+
1990
+ 1− 1
1991
+ 𝑎
1992
+ �2�
1993
+ .
1994
+ (49)
1995
+ ® Upper bound of P
1996
+ ���� ˆ𝜶(3)
1997
+ 1
1998
+ −𝜶1
1999
+ ��� ≥
2000
+ √︁ 𝝐
2001
+ 2
2002
+
2003
+ : Substituting
2004
+ the
2005
+ values
2006
+ of
2007
+ ˆ𝛼(3)
2008
+ 1 ,
2009
+ as
2010
+ given
2011
+ by
2012
+ (38),
2013
+ in
2014
+ P
2015
+ ���� ˆ𝜶(3)
2016
+ 1
2017
+ −𝜶1
2018
+ ��� ≥
2019
+ √︁ 𝝐
2020
+ 2
2021
+
2022
+ and
2023
+ following
2024
+ similar
2025
+ steps
2026
+ to bound P
2027
+ ���� ˆ𝜶(1)
2028
+ 1
2029
+ −𝜶1
2030
+ ��� ≥
2031
+ √︁ 𝝐
2032
+ 2
2033
+
2034
+ , we obtain
2035
+ P
2036
+ ���� ˆ𝛼(3)
2037
+ 1
2038
+ −𝛼1)
2039
+ ��� ≥
2040
+ √︂
2041
+ 𝜖
2042
+ 2
2043
+
2044
+ ≤ P
2045
+ �����
2046
+ ˆ𝜇1 −2𝑛
2047
+ ˆ𝜇3 −2𝑛
2048
+ ���� ≤
2049
+
2050
+ 1− 1
2051
+ 𝑎
2052
+ �2�
2053
+ .
2054
+ (50)
2055
+ ® Upper bound of P
2056
+ ���� ˆ𝜶(4)
2057
+ 1
2058
+ −𝜶1
2059
+ ��� ≥
2060
+ √︁ 𝝐
2061
+ 2
2062
+
2063
+ : Substituting
2064
+ the
2065
+ values
2066
+ of
2067
+ ˆ𝛼(4)
2068
+ 1 ,
2069
+ as
2070
+ given
2071
+ by
2072
+ (38),
2073
+ in
2074
+ P
2075
+ ���� ˆ𝜶(4)
2076
+ 1
2077
+ −𝜶1
2078
+ ��� ≥
2079
+ √︁ 𝝐
2080
+ 2
2081
+
2082
+ and
2083
+ following
2084
+ similar
2085
+ steps
2086
+ to bound P
2087
+ ���� ˆ𝜶(2)
2088
+ 2
2089
+ −𝜶1
2090
+ ��� ≥
2091
+ √︁ 𝝐
2092
+ 2
2093
+
2094
+ , we obtain
2095
+ P
2096
+ ���� ˆ𝛼(4)
2097
+ 1
2098
+ −𝛼1)
2099
+ ��� ≥
2100
+ √︂
2101
+ 𝜖
2102
+ 2
2103
+
2104
+ ≤ P
2105
+ �����
2106
+ ˆ𝜇1 −2𝑛
2107
+ ˆ𝜇3 −2𝑛
2108
+ ���� ≤
2109
+
2110
+ 1+ 1
2111
+ 𝑎
2112
+ �2�
2113
+ −P
2114
+ �����
2115
+ ˆ𝜇1 −2𝑛
2116
+ ˆ𝜇3 −2𝑛
2117
+ ���� ≤
2118
+
2119
+ 1− 1
2120
+ 𝑎
2121
+ �2�
2122
+ .
2123
+ (51)
2124
+ By inserting the bounds (46), (49), (50) and (51) to (42)
2125
+ we obtain
2126
+ P
2127
+ ��� ˆ𝛽∗
2128
+ 1 −𝛼1
2129
+ �� ≥
2130
+ √︂
2131
+ 𝜖
2132
+ 2
2133
+
2134
+ ≤ 2
2135
+
2136
+ P
2137
+ �����
2138
+ ˆ𝜇2 −2𝑛
2139
+ ˆ𝜇1 −2𝑛
2140
+ ���� ≥ 𝛾1
2141
+
2142
+ +P
2143
+ �����
2144
+ ˆ𝜇3 −2𝑛
2145
+ ˆ𝜇1 −2𝑛
2146
+ ���� ≥ 𝛾1
2147
+ ��
2148
+ ,
2149
+ (52)
2150
+ where we set 𝛾1 =
2151
+
2152
+ 1
2153
+ 1+(1/𝑎)
2154
+ �2
2155
+ .
2156
+ We next upper bound each term on the right-hand side of
2157
+ (52). The bounds are derived using the properties of the
2158
+ sub-exponential distributions which we introduce below.
2159
+ G Step 2: Sub-exponential Distributions and its Tail
2160
+ Bound
2161
+ Definition IV.1 (sub-exponential distribution). A RV 𝑋
2162
+ with mean 𝜇 is said to be sub-exponential with parameters
2163
+ (𝜈,𝛼), for 𝛼 > 0, if
2164
+ E
2165
+
2166
+ exp
2167
+
2168
+ 𝑡(𝑋 − 𝜇)
2169
+ ��
2170
+ ≤ exp
2171
+ �𝑡2𝜈2
2172
+ 2
2173
+
2174
+ , for |𝑡| < 1
2175
+ 𝛼 .
2176
+ Theorem
2177
+ 3
2178
+ ([18]). Let
2179
+ 𝑋𝑘
2180
+ for
2181
+ 𝑘 = 1,2,...,𝑛 be
2182
+ independent RVs where 𝑋𝑘 is sub-exponential with
2183
+ parameters
2184
+ (𝜈𝑘,𝑏𝑘), and mean
2185
+ 𝜇𝑘 = E [𝑋𝑘]. Then
2186
+ 𝑛�
2187
+ 𝑘=1
2188
+ (𝑋𝑘 − 𝜇𝑘) is a sub-exponential RV with parameters
2189
+ (𝜈∗,𝑏∗) where
2190
+ 𝑏∗ =
2191
+ max
2192
+ 𝑘=1,2...,𝑛𝑏𝑘,
2193
+ and
2194
+ 𝜈∗ =
2195
+
2196
+ � 𝑛
2197
+ ∑︁
2198
+ 𝑘=1
2199
+ 𝜈2
2200
+ 𝑘.
2201
+ Furthermore, its tail probability can be bounded as
2202
+ P
2203
+ ������
2204
+ 1
2205
+ 𝑛
2206
+ 𝑛
2207
+ ∑︁
2208
+ 𝑘=1
2209
+ (𝑋𝑘 − 𝜇𝑘)
2210
+ ����� ≥ 𝑡
2211
+
2212
+
2213
+ ���
2214
+ ���
2215
+ 2𝑒
2216
+
2217
+ 𝑛𝑡2
2218
+ 2(𝜈2∗ /𝑛) ,
2219
+ for 0 ≤ 𝑡 ≤
2220
+ 𝜈2
2221
+
2222
+ 𝑛𝑏∗
2223
+ 2𝑒− 𝑛𝑡
2224
+ 2𝑏∗ ,
2225
+ for 𝑡 ≥
2226
+ 𝜈2
2227
+
2228
+ 𝑛𝑏∗ .
2229
+ Proof. The proof is given in Appendix C.
2230
+
2231
+ Corollary
2232
+ 1.
2233
+ Let
2234
+ 𝑋𝑘
2235
+ for
2236
+ 𝑘 = 1,2...,𝑛
2237
+ be
2238
+ i.i.d.
2239
+ sub-exponential RVs with parameters (2(2+2𝑎),4) each
2240
+ with mean 2+ 𝑎. Then,
2241
+ P
2242
+ ������
2243
+ 1
2244
+ 𝑛
2245
+ 𝑛
2246
+ ∑︁
2247
+ 𝑘=1
2248
+ (𝑋𝑘 − 𝜇𝑘)
2249
+ ����� ≥ 𝑡
2250
+
2251
+ ≤ 2𝑒
2252
+
2253
+ 𝑛𝑡2
2254
+ 8(2+2𝑎)2 ,
2255
+ for 𝑡 > 0.
2256
+ Proof. The proof is given in then Appendix D.
2257
+
2258
+ We use Corollary 1 to upper bound of the right-hand
2259
+ side terms in (52). The following lemma establishes
2260
+ the connection between the non-central chi-squared
2261
+ distribution and the sub-exponential distributions.
2262
+ Lemma 3. Let 𝑋 ∼ 𝜒2
2263
+ 𝑝(𝑎). Then, 𝑋 is sub-exponential
2264
+ with parameters �2(𝑝 +2𝑎),4�.
2265
+ Proof. The proof is given in then Appendix E.
2266
+
2267
+ G Step 3: Upper Bounding Eq. (52)
2268
+ – Recall that ˆ𝜇1 ∼ 𝜒2
2269
+ 2𝑛(𝑛𝜆1) and ˆ𝜇2 ∼ 𝜒2
2270
+ 2𝑛(𝑛𝜆2). Let
2271
+ 𝑓 ˆ𝜇1 denote the pdf of ˆ𝜇1. We upper bound the term
2272
+ P
2273
+ ���� ˆ𝜇2−2𝑛
2274
+ ˆ𝜇1−2𝑛
2275
+ ��� ≥ 𝛾1
2276
+
2277
+ as follows
2278
+ P
2279
+ �����
2280
+ ˆ𝜇2 −2𝑛
2281
+ ˆ𝜇1 −2𝑛
2282
+ ���� ≥ 𝛾1
2283
+
2284
+ =
2285
+
2286
+
2287
+ 0
2288
+ P
2289
+ ����� ˆ𝜇2 −2𝑛
2290
+ ���� ≥ 𝛾1
2291
+ ����𝑢 −2𝑛
2292
+ ����
2293
+
2294
+ 𝑓 ˆ𝜇1(𝑢)𝑑𝑢
2295
+ =
2296
+
2297
+
2298
+ 0
2299
+ P
2300
+ ����� ˆ𝜇2 −2𝑛 −𝑛𝜆2 +𝑛𝜆2
2301
+ ���� ≥ 𝛾1
2302
+ ����𝑢 −2𝑛
2303
+ ����
2304
+
2305
+ 𝑓 ˆ𝜇1(𝑢)𝑑𝑢
2306
+
2307
+
2308
+
2309
+ 0
2310
+ P
2311
+ �1
2312
+ 𝑛
2313
+ ���� ˆ𝜇2 −𝑛(2+𝜆2)
2314
+ ���� ≥ 𝛾1|𝑢 −2𝑛| −𝑛𝜆2
2315
+ 𝑛
2316
+
2317
+ 𝑓 ˆ𝜇1(𝑢)𝑑𝑢
2318
+ (53)
2319
+ Note that, if 𝛾1 |𝑢−2𝑛|−𝑛𝜆2
2320
+ 𝑛
2321
+ < 0, then P
2322
+ ���� ˆ𝜇2−2𝑛
2323
+ ˆ𝜇1−2𝑛
2324
+ ��� ≥ 𝛾1
2325
+
2326
+
2327
+ 1 as P
2328
+
2329
+ 1
2330
+ 𝑛
2331
+ ���� ˆ𝜇2 −𝑛(2+𝜆2)
2332
+ ���� ≥ 𝛾1 |𝑢−2𝑛|−𝑛𝜆2
2333
+ 𝑛
2334
+
2335
+ = 1, which is
2336
+ trivial.
2337
+
2338
+ 9
2339
+ For 𝛾1 |𝑢−2𝑛|−𝑛𝜆2
2340
+ 𝑛
2341
+ ≥ 0, using the assumption 0 ≤ 𝜖 ≤ 1 in
2342
+ (53), we have
2343
+ P
2344
+ �����
2345
+ ˆ𝜇2 −2𝑛
2346
+ ˆ𝜇1 −2𝑛
2347
+ ���� ≥ 𝛾1
2348
+
2349
+
2350
+
2351
+
2352
+ 0
2353
+ P
2354
+ �1
2355
+ 𝑛
2356
+ ���� ˆ𝜇2 −𝑛(2+𝜆2)
2357
+ ���� ≥ 𝜖
2358
+ � 𝛾1|𝑢 −2𝑛| −𝑛𝜆2
2359
+ 𝑛
2360
+ � �
2361
+ × 𝑓 ˆ𝜇1(𝑢)𝑑𝑢.
2362
+ (54)
2363
+ The last inequality follows from Lemma 1. Let
2364
+ 𝑡1 := 𝑡1(𝑢) = 𝜖
2365
+
2366
+ 𝛾1 |𝑢−2𝑛|−𝑛𝜆2
2367
+ 𝑛
2368
+
2369
+ . As E [ ˆ𝜇2] = 2𝑛 +𝑛𝜆2, by
2370
+ applying Corollary 1, we obtain
2371
+ P
2372
+ �1
2373
+ 𝑛
2374
+ ���� ˆ𝜇2 −𝑛(2+𝜆2)
2375
+ ���� ≥ 𝑡1
2376
+
2377
+ ≤ 2𝑒
2378
+
2379
+ 𝑛𝑡2
2380
+ 1
2381
+ 8(2+2𝜆2)2 ,
2382
+ 𝑡1 ≥ 0.
2383
+ (55)
2384
+ By applying (55) to (54), we obtain
2385
+ P
2386
+ ����� ˆ𝜇2 −2𝑛
2387
+ ���� ≥ 𝛾1
2388
+ ���� ˆ𝜇1 −2𝑛
2389
+ ����
2390
+
2391
+
2392
+
2393
+
2394
+ 0
2395
+ 2𝑒
2396
+
2397
+ 𝑛𝑡2
2398
+ 1
2399
+ 8(2+2𝜆2)2 𝑓 ˆ𝜇1(𝑢)𝑑𝑢,
2400
+ =
2401
+ 2𝑛
2402
+
2403
+ 0
2404
+ 2𝑒
2405
+
2406
+ 𝑛
2407
+
2408
+ 𝜖𝑛
2409
+
2410
+ 𝛾1 (2𝑛−𝑢)−𝑛𝜆2
2411
+ ��2
2412
+ 8(2+2𝜆2)2
2413
+ 𝑓 ˆ𝜇1(𝑢)𝑑𝑢
2414
+ +
2415
+
2416
+
2417
+ 2𝑛
2418
+ 2𝑒
2419
+
2420
+ 𝑛
2421
+
2422
+ 𝜖𝑛
2423
+
2424
+ 𝛾1 (𝑢−2𝑛)−𝑛𝜆2
2425
+ ��2
2426
+ 8(2+2𝜆2)2
2427
+ 𝑓 ˆ𝜇1 (𝑢)𝑑𝑢.
2428
+ (56)
2429
+ For 0 ≤ 𝑢 ≤ 2𝑛, we have
2430
+ 2𝑒
2431
+
2432
+ 𝑛
2433
+
2434
+ 𝜖𝑛
2435
+
2436
+ 𝛾1 (2𝑛−𝑢)−𝑛𝜆2
2437
+ ��2
2438
+ 8(2+2𝜆2)2
2439
+ ≤ 2𝑒
2440
+
2441
+ 𝑛�
2442
+ 𝜖 𝜆2
2443
+ �2
2444
+ 8(2+2𝜆2)2 .
2445
+ (57)
2446
+ For 2𝑛 ≤ 𝑢 ≤ ∞, we have
2447
+ 2𝑒
2448
+
2449
+ 𝑛
2450
+
2451
+ 𝜖𝑛
2452
+
2453
+ 𝛾1 (𝑢−2𝑛)−𝑛𝜆2
2454
+ ��2
2455
+ 8(2+2𝜆2)2
2456
+ ≤ 2𝑒
2457
+
2458
+ 𝑛�
2459
+ 𝜖 𝜆2
2460
+ �2
2461
+ 8(2+2𝜆2)2 .
2462
+ (58)
2463
+ Using (57) and (58) in (56), we obtain
2464
+ P
2465
+ ����� ˆ𝜇2 −2𝑛
2466
+ ���� ≥ 𝛾1
2467
+ ���� ˆ𝜇1 −2𝑛
2468
+ ����
2469
+
2470
+ ≤ 2𝑒
2471
+
2472
+ 𝑛�
2473
+ 𝜖 𝜆2
2474
+ �2
2475
+ 8(2+2𝜆2)2 P{0 < ˆ𝜇1 < 2𝑛}
2476
+ +2𝑒
2477
+
2478
+ 𝑛�
2479
+ 𝜖 𝜆2
2480
+ �2
2481
+ 8(2+2𝜆2)2 P{ ˆ𝜇1 > 2𝑛}
2482
+ P
2483
+ ����� ˆ𝜇2 −2𝑛
2484
+ ���� ≥ 𝛾1
2485
+ ���� ˆ𝜇1 −2𝑛
2486
+ ����
2487
+
2488
+ ≤ 2𝑒
2489
+
2490
+ 𝑛�
2491
+ 𝜖 𝜆2
2492
+ �2
2493
+ 8(2+2𝜆2)2 .
2494
+ (59)
2495
+ – We next upper bound P
2496
+ ���� ˆ𝜇3−2𝑛
2497
+ ˆ𝜇1−2𝑛
2498
+ ��� ≥ 𝛾1
2499
+
2500
+ . Set 𝑡2 =
2501
+ 𝜖
2502
+
2503
+ 𝛾1 |𝑢−2𝑛|−𝑛𝜆3
2504
+ 𝑛
2505
+
2506
+ . Recall that ˆ𝜇3 ∼ 𝜒2
2507
+ 2𝑛(𝑛𝜆3). Following
2508
+ the steps similar to the derivation of the bound in (59),
2509
+ we obtain
2510
+ P
2511
+ ����� ˆ𝜇3 −2𝑛
2512
+ ���� ≥ 𝛾1
2513
+ ���� ˆ𝜇1 −2𝑛
2514
+ ����
2515
+
2516
+ ≤ 2𝑒
2517
+
2518
+ 𝑛�
2519
+ 𝜖 𝜆3
2520
+ �2
2521
+ 8(2+2𝜆3)2 .
2522
+ (60)
2523
+ Combining (59) and (60) we obtain the following upper
2524
+ bound on (52)
2525
+ P
2526
+ ��� ˆ𝛽∗
2527
+ 1 −𝛼1
2528
+ �� ≥
2529
+ √︂
2530
+ 𝜖
2531
+ 2
2532
+
2533
+ ≤ 4
2534
+
2535
+ 𝑒− 𝑛
2536
+ 32
2537
+ � 𝜖 𝜆2
2538
+ 1+𝜆2
2539
+ �2
2540
+ + 𝑒− 𝑛
2541
+ 32
2542
+ � 𝜖 𝜆3
2543
+ 1+𝜆3
2544
+ �2�
2545
+ . (61)
2546
+ G Step 4: Upper bound on II
2547
+ By following the same steps for deriving the upper bound
2548
+ of P
2549
+ ��� ˆ𝛽∗
2550
+ 1 −𝛼1
2551
+ �� ≥ √︁ 𝜖
2552
+ 2
2553
+
2554
+ , we can obtain the following bound
2555
+ P
2556
+ ��� ˆ𝛽∗
2557
+ 2 −𝛼2
2558
+ �� ≥
2559
+ √︂
2560
+ 𝜖
2561
+ 2
2562
+
2563
+ ≤ 4
2564
+
2565
+ 𝑒− 𝑛
2566
+ 32
2567
+ � 𝜖 𝜆4
2568
+ 1+𝜆4
2569
+ �2
2570
+ + 𝑒− 𝑛
2571
+ 32
2572
+ � 𝜖 𝜆5
2573
+ 1+𝜆5
2574
+ �2�
2575
+ . (62)
2576
+ Combining (61) and (62), we obtain the required upper
2577
+ bound in (27).
2578
+
2579
+ V. NUMERICAL SIMULATIONS
2580
+ We estimate the initial value of (𝛽0
2581
+ 1, 𝛽0
2582
+ 2) as given in [13, Sec.
2583
+ V.C]. Based on the the initial value of (𝛽0
2584
+ 1, 𝛽0
2585
+ 2) we set B as
2586
+ given in (7), where 𝑣 and 𝑤 are selected such that 𝐾𝑥𝑣 ∈ N and
2587
+ 𝐾𝑦𝑤 ∈ N respectively. In addition to the LoS path, we assume
2588
+ that there are 4 NLoS path components due to scatters between
2589
+ the user and the HMT. The elevation and azimuth angles of
2590
+ each NLoS path from these scatters to the center of HMT follow
2591
+ the uniform distribution, i.e., 𝑈(0,2𝜋). Moreover, we consider
2592
+ the path coefficient of each NLoS path as a complex Gaussian
2593
+ distribution, i.e., 𝐶𝑁(0,𝜎2
2594
+ 𝑠 ), where 𝜎2
2595
+ 𝑠 is 20 dB weaker than
2596
+ the power of the LoS component [19]. The system parameters
2597
+ for numerical simulations are listed in Table I.
2598
+ TABLE I: A list of system parameters for numerical simulations
2599
+ Parameters
2600
+ Values
2601
+ Description
2602
+ 𝑓𝑐
2603
+ 30 GHz
2604
+ Carrier frequency
2605
+ 𝜆
2606
+ 1 cm
2607
+ Wavelength
2608
+ 𝐿𝑥
2609
+ 1 m
2610
+ Width of the HMT
2611
+ 𝐿𝑦
2612
+ 1 m
2613
+ Length of the HMT
2614
+ 𝑑𝑟
2615
+ 𝜆/4
2616
+ Unit element spacing
2617
+ 𝐿𝑒
2618
+ 𝑑𝑟
2619
+ Width
2620
+ and
2621
+ length
2622
+ of
2623
+ each
2624
+ phase-shifting element
2625
+ 𝑃
2626
+ 20 dBm
2627
+ Transmission power of the HMT
2628
+ during data transmission
2629
+ 𝜎2
2630
+ -115 dBm
2631
+ Noise power for 200 KHz 2
2632
+ A. Comparison
2633
+ Between
2634
+ the
2635
+ Proposed
2636
+ Algorithm
2637
+ and
2638
+ Benchmark Scheme
2639
+ According to the approximated channel model, where the
2640
+ phase-shift parameters at the HMT are given by 𝛽1 and 𝛽2,
2641
+ the achieved data rate at the user of the HMT-assisted wireless
2642
+ communication system is given by
2643
+ 𝑅(𝛽1, 𝛽2) = 𝑙𝑜𝑔2
2644
+
2645
+ 1+ 𝑃|𝐻(𝛽1, 𝛽2)|2
2646
+ 𝜎2
2647
+
2648
+ ,
2649
+ (63)
2650
+
2651
+ 10
2652
+ where 𝑃 is the transmission power at the HMT. The HMT
2653
+ uses the acquired CSI during the channel estimation period to
2654
+ maximize the received data rate by the user. Hence, we consider
2655
+ the achieved data rate by the user, using the acquired CSI as
2656
+ a performance metric. We applied the proposed algorithm in
2657
+ two different cases when the distance between the user and
2658
+ the center of the HMT (𝑑0 = 200 m and when 𝑑0 = 10 m. We
2659
+ compared our proposed algorithm with two benchmarks, the
2660
+ proposed algorithm in [13] and the oracle scheme where 𝛼1
2661
+ and 𝛼2 are estimated perfectly and thereby the maximum rate
2662
+ is achieved.
2663
+ In Fig. 4, we compared the achievable rates, given by (63),
2664
+ of the proposed scheme and the benchmark schemes. We
2665
+ considered both 𝑑0 = 200 m and 𝑑0 = 10 m regions of the
2666
+ HMT with respect to the transmit power of the pilot signals
2667
+ when the number of the pilot signals is fixed to 23. For all
2668
+ the algorithms we use the same number of pilots, i.e. 23,
2669
+ for both the cases. Our proposed algorithm uses four pilots
2670
+ in each epoch and there are five epochs, which makes the
2671
+ total number of pilots equals to 20. We require additional three
2672
+ number of pilots to estimate (𝛽0
2673
+ 1, 𝛽0
2674
+ 2). We run the simulation for
2675
+ 1000 times. We see that in both cases, the proposed Two-Stage
2676
+ Phase-Shifts Estimation Algorithm gives higher rates than other
2677
+ two benchmark schemes.
2678
+ 20
2679
+ 10
2680
+ 0
2681
+ 10
2682
+ 20
2683
+ 30
2684
+ 40
2685
+ Power of pilot signals (in dBm)
2686
+ 15
2687
+ 20
2688
+ 25
2689
+ 30
2690
+ 35
2691
+ 40
2692
+ 45
2693
+ 50
2694
+ Achievable Rate (in bits/symbol)
2695
+ Maximum Rate d0=200
2696
+ Proposed Algorithm d0=200
2697
+ Algorithm from [13] d0=200
2698
+ Maximum Rate d0=10
2699
+ Proposed Algorithm d0=10
2700
+ Algorithm from [13] d0=10
2701
+ Fig. 4: Achievable rate vs. the transmit power of the pilot signals
2702
+ (in dBm).
2703
+ B. Convergence of The Proposed Algorithm
2704
+ We now numerically evaluate the convergence of the upper
2705
+ bound of the proposed algorithm, given by (27). We also
2706
+ compare the actual probability, given by (26), that we obtain
2707
+ by simulations.
2708
+ In Fig. 5, we show the convergence property of the error
2709
+ probability and its upper bound of the proposed algorithm
2710
+ for increasing values of 𝜖 = {0.01,0.05,0.1} when the power
2711
+ of the pilot signal is 𝑃 = 10 dBm and 𝑑0 = 200 m. We run
2712
+ the simulation for 1000 times. We see that for each value
2713
+ of 𝜖, the proposed algorithm converges towards zero as we
2714
+ increase the number of pilots. Moreover, the upper bound of
2715
+ the error probability also converges to the error probability as
2716
+ the number of pilot signals increases.
2717
+ 2This setting corresponds to the noise power spectrum density at the HMT
2718
+ is −174 dBm/Hz and signal bandwidth is 200 KHz, assuming the noise figure
2719
+ of each user to be 6 dB [8].
2720
+ 0
2721
+ 500
2722
+ 1000
2723
+ 1500
2724
+ 2000
2725
+ 2500
2726
+ 3000
2727
+ Number of Pilots
2728
+ 10
2729
+ 1
2730
+ 100
2731
+ Error Probability
2732
+ Proposed Scheme, = 0.01
2733
+ Upper Bound, = 0.01
2734
+ Proposed Scheme, = 0.05
2735
+ Upper Bound, = 0.05
2736
+ Proposed Scheme, = 0.1
2737
+ Upper Bound, = 0.1
2738
+ Fig. 5: Error Probability Bound v/s Number of Pilots for 𝜖 =
2739
+ {0.01,0.05,0.1} for 𝑃 = 10 dBm.
2740
+ In Fig. 6, we compare the convergence property of the error
2741
+ probability of the proposed algorithm with respect to 𝜖 = 0.05
2742
+ and 𝑑0 = 200 m for different levels of power of the pilot signals,
2743
+ 𝑃 = {5,10,20} dBm. As we increase the power of the pilot
2744
+ signals, the estimation accuracy of 𝛼1 and 𝛼2 increases and
2745
+ hence the error probability decreases. This is so because, as
2746
+ we increase the power of pilot signals the received signals will
2747
+ be less noisy which increases the chances of estimating the 𝛼1
2748
+ and 𝛼2 more accurately.
2749
+ 0
2750
+ 500
2751
+ 1000
2752
+ 1500
2753
+ 2000
2754
+ 2500
2755
+ 3000
2756
+ Number of Pilots
2757
+ 10
2758
+ 2
2759
+ 10
2760
+ 1
2761
+ 100
2762
+ Error Probability
2763
+ Proposed Scheme, Pilot Power = 5 dBm
2764
+ Upper Bound, Pilot Power = 5 dBm
2765
+ Proposed Scheme, Pilot Power = 10 dBm
2766
+ Upper Bound, Pilot Power = 10 dBm
2767
+ Proposed Scheme, Pilot Power = 20 dBm
2768
+ Upper Bound, Pilot Power = 20 dBm
2769
+ Fig. 6: Error Probability Bound v/s Number of Pilots for 𝜖 = 0.05
2770
+ for 𝑃 = {5,10,20} dBm.
2771
+ VI. CONCLUSION
2772
+ We investigated the problem of estimation of the optimal
2773
+ phase-shift at the HMT-assisted wireless communication system
2774
+ in a noisy environment. We proposed a learning algorithm to
2775
+ estimate the optimal phase-shifting parameters and showed that
2776
+ the probability that the phase-shifting parameters generated by
2777
+ the proposed algorithm to deviate by more than 𝜖 from the
2778
+ optimal values decay exponentially fast as the number of pilots
2779
+ grows. Our proposed algorithm exploited structural properties
2780
+ of the channel gains in the far-field regions.
2781
+
2782
+ 11
2783
+ APPENDIX
2784
+ A. Proof of Proposition 1
2785
+ Proof. Let us define the following events.
2786
+ 𝐴𝑛,𝑚 = |𝑋𝑛 − 𝑋𝑚| > 𝜖,
2787
+ 𝐴𝑛 = |𝑋𝑛 − 𝑋| > 𝜖
2788
+ 2,
2789
+ and
2790
+ 𝐴𝑚 = |𝑋𝑚 − 𝑋| > 𝜖
2791
+ 2
2792
+ By the triangle inequality, we have
2793
+ |𝑋𝑛 − 𝑋𝑚| ≤ |𝑋𝑛 − 𝑋| + |𝑋𝑚 − 𝑋|.
2794
+ (64)
2795
+ Using (64), the event 𝐴𝑛,𝑚 can be written as
2796
+ |𝑋𝑛 − 𝑋𝑚| ≥ 𝜖 =⇒ |𝑋𝑛 − 𝑋| + |𝑋𝑚 − 𝑋| ≥ 𝜖
2797
+ Therefore, we have
2798
+ 𝐴𝑛,𝑚 ⊂ {|𝑋𝑛 − 𝑋| + |𝑋 − 𝑋𝑚| > 𝜖}
2799
+
2800
+
2801
+ |𝑋𝑛 − 𝑋| > 𝜖
2802
+ 2
2803
+
2804
+ |𝑋 − 𝑋𝑚| > 𝜖
2805
+ 2
2806
+
2807
+ (65)
2808
+ Note that for any two events 𝐴 and 𝐵 where 𝐴 ⊂ 𝐵, then
2809
+ P{𝐴} ≤ P{𝐵}. We use this fact in (65), and we get
2810
+ P{|𝑋𝑛 − 𝑋𝑚| > 𝜖} ≤ P
2811
+
2812
+ |𝑋𝑛 − 𝑋| > 𝜖
2813
+ 2
2814
+
2815
+ +P
2816
+
2817
+ |𝑋𝑚 − 𝑋| > 𝜖
2818
+ 2
2819
+
2820
+
2821
+ B. Proof of Lemma 2
2822
+ Proof. We consider 𝑟(𝛽1, 𝛽2) as given in (3) which comprises
2823
+ of two complex-valued factors
2824
+
2825
+ 𝑃 × 𝐻(𝛽1, 𝛽2) (see (1)) and 𝜁.
2826
+ Write 𝜁 = 𝑛1 + 𝑗𝑛2, where 𝑛1 and 𝑛2 follows 𝑁
2827
+
2828
+ 0, 𝜎2
2829
+ 2
2830
+
2831
+ and
2832
+ are independent, and write
2833
+
2834
+ 𝑃 × 𝐻(𝛽1, 𝛽2) = 𝑎 + 𝑗𝑏, where 𝑎
2835
+ and 𝑏 are real values. Therefore,
2836
+ 𝑟(𝛽1, 𝛽2) = |𝑦(𝛽1, 𝛽2)|2 = (𝑎 +𝑛1)2 + (𝑏 +𝑛2)2.
2837
+ (66)
2838
+ Note that 𝑎+𝑛1
2839
+ 𝜎/
2840
+
2841
+ 2 ∼ 𝑁
2842
+
2843
+ 𝑎
2844
+ 𝜎/
2845
+
2846
+ 2,1
2847
+
2848
+ and 𝑏+𝑛2
2849
+ 𝜎/
2850
+
2851
+ 2 ∼ 𝑁
2852
+
2853
+ 𝑏
2854
+ 𝜎/
2855
+
2856
+ 2,1
2857
+
2858
+ and they
2859
+ are independent. Therefore,
2860
+ 2
2861
+ 𝜎2
2862
+
2863
+ (𝑎 +𝑛1)2 + (𝑏 +𝑛2)2
2864
+
2865
+ ∼ 𝜒2
2866
+ 2
2867
+ � 2
2868
+ 𝜎2
2869
+
2870
+ 𝑎2 + 𝑏2��
2871
+ .
2872
+ (67)
2873
+ Applying (67) in (66), we get 𝑋 =
2874
+ 2
2875
+ 𝜎2 𝑟(𝛽1, 𝛽2) ∼ 𝜒2
2876
+ 2 (𝜆1) ,
2877
+ where 𝜆1 =
2878
+ 2
2879
+ 𝜎2
2880
+ ���
2881
+
2882
+ 𝑃 × 𝐻(𝛽1, 𝛽2)
2883
+ ���
2884
+ 2
2885
+ . The second part of the lemma
2886
+ follows from the additive property of non-central Chi-squared
2887
+ distribution of the sum of 𝑛 i.i.d. RVs of 𝜒2
2888
+ 2 (𝜆1) .
2889
+
2890
+ C. Proof of Theorem 3
2891
+ As 𝑋𝑘,∀𝑘 are independent, applying the definition IV.1, the
2892
+ moment generating function of
2893
+ 𝑛�
2894
+ 𝑘=1
2895
+ (𝑋𝑘 − 𝜇𝑘) is given by
2896
+ E
2897
+
2898
+ 𝑒
2899
+ 𝑡
2900
+ 𝑛�
2901
+ 𝑘=1
2902
+ (𝑋𝑘−𝜇𝑘)
2903
+
2904
+ ≤ 𝑒
2905
+ 𝜆2
2906
+ 2
2907
+ 𝑛�
2908
+ 𝑘=1
2909
+ 𝜈2
2910
+ 𝑘,
2911
+ ∀|𝑡| < ��
2912
+
2913
+ 1
2914
+ max
2915
+ 𝑘=1,2,...,𝑛𝑏𝑘
2916
+ ��
2917
+
2918
+ .
2919
+ Since the moment generating functions uniquely determines
2920
+ the distribution, comparing with the definition IV.1, it follows
2921
+ that
2922
+ 𝑛�
2923
+ 𝑘=1
2924
+ (𝑋𝑘 − 𝜇𝑘) is a sub-exponential (𝜈∗,𝑏∗) random variable,
2925
+ where
2926
+ 𝑏∗ =
2927
+ max
2928
+ 𝑘=1,2...,𝑛𝑏𝑘
2929
+ and
2930
+ 𝜈∗ =
2931
+
2932
+ � 𝑛
2933
+ ∑︁
2934
+ 𝑘=1
2935
+ 𝜈2
2936
+ 𝑘.
2937
+ To prove the second part of the Theorem we use the following
2938
+ tail bound on a sub-exponential distribution proved in [18].
2939
+ Proposition 1 ([18] Proposition 2.9). Let 𝑋 is sub-exponential
2940
+ random variable with parameters (𝜈,𝑏) and E [𝑋] = 𝜇. Then
2941
+ P{|𝑋 − 𝜇| ≥ 𝑡} ≤
2942
+
2943
+ 2𝑒− 𝑡2
2944
+ 2𝜈2 ,
2945
+ if 0 ≤ 𝑡 ≤ 𝜈2
2946
+ 𝑏
2947
+ 2𝑒− 𝑡
2948
+ 2𝑏 ,
2949
+ if 𝑡 ≥ 𝜈2
2950
+ 𝑏 .
2951
+ The claim immediately follows by applying the above result on
2952
+ 𝑍𝑛 :=
2953
+ 𝑛�
2954
+ 𝑘=1
2955
+ (𝑋𝑘 − 𝜇𝑘), which is sub-exponential (𝜈∗,𝑏∗), where
2956
+ 𝑏∗ =
2957
+ max
2958
+ 𝑘=1,2...,𝑛𝑏𝑘 and 𝜈∗ =
2959
+ √︂ 𝑛�
2960
+ 𝑘=1
2961
+ 𝜈2
2962
+ 𝑘.
2963
+ D. Proof of Corollary 1
2964
+ From Theorem 3, �𝑛
2965
+ 𝑘=1(𝑋𝑘 −𝜇𝑘) is sub-exponential (𝜈∗,𝑏∗),
2966
+ where 𝑏∗ = 4 and 𝜈∗ = 2√𝑛(2 + 2𝑎). Using the parameters
2967
+ (𝜈∗,𝑏∗) = (2√𝑛(2+2𝑎),4) in Proposition 1, we get the required
2968
+ upper bound as
2969
+ P
2970
+ ������
2971
+ 1
2972
+ 𝑛
2973
+ 𝑛
2974
+ ∑︁
2975
+ 𝑘=1
2976
+ (𝑋𝑘 − 𝜇𝑘)
2977
+ ����� ≥ 𝑡
2978
+
2979
+
2980
+ ��
2981
+ ��
2982
+ 2𝑒
2983
+
2984
+ 𝑛𝑡2
2985
+ 8(2+2𝑎)2 ,
2986
+ 0 ≤ 𝑡 ≤ (2+2𝑎)2
2987
+ 2𝑒− 𝑛𝑡
2988
+ 8 ,
2989
+ 𝑡 ≥ (2+2𝑎)2
2990
+ P
2991
+ ������
2992
+ 1
2993
+ 𝑛
2994
+ 𝑛
2995
+ ∑︁
2996
+ 𝑘=1
2997
+ (𝑋𝑘 − 𝜇𝑘)
2998
+ ����� ≥ 𝑡
2999
+
3000
+ ≤ 2𝑒
3001
+
3002
+ 𝑛𝑡2
3003
+ 8(2+2𝑎)2 ,
3004
+ 𝑡 > 0.
3005
+ E. Proof of Lemma 3
3006
+ If 𝑋 ∼ 𝜒2
3007
+ 𝑝(𝑎), then according to [20], the moment-generating
3008
+ function (MGF) of 𝑋 is given by
3009
+ E [exp{𝑡(𝑋 − (𝑝 + 𝑎)}] = 𝑒−𝑡 ( 𝑝+𝑎)E
3010
+
3011
+ 𝑒𝑡𝑋�
3012
+ = 𝑒−𝑡 ( 𝑝+𝑎)𝑒
3013
+ 𝑎𝑡
3014
+ 1−2𝑡
3015
+ (1−2𝑡) 𝑝/2
3016
+ = 𝑒
3017
+ 2𝑎𝑡2
3018
+ 1−2𝑡
3019
+ 𝑒−𝑝𝑡
3020
+ (1−2𝑡) 𝑝/2 ,
3021
+ for 𝑡 < 1
3022
+ 2.
3023
+ (68)
3024
+ By following some calculus, refer [21], [18, Example 2.8], we
3025
+ obtain
3026
+ 𝑒−𝑝𝑡
3027
+ (1−2𝑡) 𝑝/2 ≤ 𝑒2𝑝𝑡2,
3028
+ for |𝑡| ≤ 1
3029
+ 4.
3030
+ (69)
3031
+ For |𝑡| ≤ 1
3032
+ 4, we have
3033
+ 𝑒
3034
+ 2𝑎𝑡2
3035
+ 1−2𝑡 ≤ 𝑒4𝑎𝑡2.
3036
+ (70)
3037
+ Applying (69) and (70) to (68), we obtain
3038
+ E [exp{𝑡(𝑋 − (𝑝 + 𝑎)}] ≤ 𝑒2( 𝑝+2𝑎)𝑡2,
3039
+ ∀|𝑡| ≤ 1
3040
+ 4.
3041
+ (71)
3042
+ Therefore, by (71), 𝑋 is Sub-exponential distribution with
3043
+ parameters �2(𝑝 +2𝑎),4�.
3044
+ REFERENCES
3045
+ [1] Z. Wan, Z. Gao, F. Gao, M. Di Renzo, and M.-S. Alouini, “Terahertz
3046
+ massive mimo with holographic reconfigurable intelligent surfaces,” IEEE
3047
+ Transactions on Communications, 2021.
3048
+
3049
+ 12
3050
+ [2] V. Jamali, A. M. Tulino, G. Fischer, R. R. Müller, and R. Schober,
3051
+ “Intelligent surface-aided transmitter architectures for millimeter-wave
3052
+ ultra massive mimo systems,” IEEE Open Journal of the Communications
3053
+ Society, vol. 2, pp. 144–167, 2020.
3054
+ [3] E. G. Larsson, O. Edfors, F. Tufvesson, and T. L. Marzetta, “Massive
3055
+ mimo for next generation wireless systems,” IEEE communications
3056
+ magazine, vol. 52, no. 2, pp. 186–195, 2014.
3057
+ [4] Q. Wu and R. Zhang, “Intelligent reflecting surface enhanced wireless
3058
+ network via joint active and passive beamforming,” IEEE Transactions
3059
+ on Wireless Communications, vol. 18, no. 11, pp. 5394–5409, 2019.
3060
+ [5] C. Huang, S. Hu, G. C. Alexandropoulos, A. Zappone, C. Yuen, R. Zhang,
3061
+ M. Di Renzo, and M. Debbah, “Holographic mimo surfaces for 6g
3062
+ wireless networks: Opportunities, challenges, and trends,” IEEE Wireless
3063
+ Communications, vol. 27, no. 5, pp. 118–125, 2020.
3064
+ [6] S. Hu, F. Rusek, and O. Edfors, “Beyond massive mimo: The potential
3065
+ of data transmission with large intelligent surfaces,” IEEE Transactions
3066
+ on Signal Processing, vol. 66, no. 10, pp. 2746–2758, 2018.
3067
+ [7] I. Yoo and D. R. Smith, “Holographic metasurface antennas for uplink
3068
+ massive mimo systems,” arXiv preprint arXiv:2108.12513, 2021.
3069
+ [8] H. Zhang, N. Shlezinger, F. Guidi, D. Dardari, M. F. Imani, and Y. C.
3070
+ Eldar, “Beam focusing for near-field multi-user mimo communications,”
3071
+ IEEE Transactions on Wireless Communications, 2022.
3072
+ [9] F. Dai and J. Wu, “Efficient broadcasting in ad hoc wireless networks
3073
+ using directional antennas,” IEEE Transactions on Parallel and
3074
+ Distributed Systems, vol. 17, no. 4, pp. 335–347, 2006.
3075
+ [10] Z. Xiao, T. He, P. Xia, and X.-G. Xia, “Hierarchical codebook design
3076
+ for beamforming training in millimeter-wave communication,” IEEE
3077
+ Transactions on Wireless Communications, vol. 15, no. 5, pp. 3380–3392,
3078
+ 2016.
3079
+ [11] K. Chen and C. Qi, “Beam training based on dynamic hierarchical
3080
+ codebook for millimeter wave massive mimo,” IEEE Communications
3081
+ Letters, vol. 23, no. 1, pp. 132–135, 2018.
3082
+ [12] Ö. T. Demir, E. Björnson, and L. Sanguinetti, “Channel modeling and
3083
+ channel estimation for holographic massive mimo with planar arrays,”
3084
+ IEEE Wireless Communications Letters, vol. 11, no. 5, pp. 997–1001,
3085
+ 2022.
3086
+ [13] M. Ghermezcheshmeh, V. Jamali, H. Gacanin, and N. Zlatanov,
3087
+ “Channel estimation for large intelligent surface-based transceiver using
3088
+ a parametric channel model,” arXiv preprint arXiv:2112.02874, 2021.
3089
+ [14] M. R. Akdeniz, Y. Liu, M. K. Samimi, S. Sun, S. Rangan, T. S. Rappaport,
3090
+ and E. Erkip, “Millimeter wave channel modeling and cellular capacity
3091
+ evaluation,” IEEE journal on selected areas in communications, vol. 32,
3092
+ no. 6, pp. 1164–1179, 2014.
3093
+ [15] S. W. Ellingson, “Path loss in reconfigurable intelligent surface-enabled
3094
+ channels,” in 2021 IEEE 32nd Annual International Symposium on
3095
+ Personal, Indoor and Mobile Radio Communications (PIMRC).
3096
+ IEEE,
3097
+ 2021, pp. 829–835.
3098
+ [16] K. T. Selvan and R. Janaswamy, “Fraunhofer and fresnel distances:
3099
+ Unified
3100
+ derivation
3101
+ for
3102
+ aperture
3103
+ antennas.”
3104
+ IEEE
3105
+ Antennas
3106
+ and
3107
+ Propagation Magazine, vol. 59, no. 4, pp. 12–15, 2017.
3108
+ [17] M. Najafi, V. Jamali, R. Schober, and H. V. Poor, “Physics-based modeling
3109
+ and scalable optimization of large intelligent reflecting surfaces,” IEEE
3110
+ Transactions on Communications, vol. 69, no. 4, pp. 2673–2691, 2020.
3111
+ [18] M. J. Wainwright, High-dimensional statistics: A non-asymptotic
3112
+ viewpoint.
3113
+ Cambridge University Press, 2019, vol. 48.
3114
+ [19] W. Wang and W. Zhang, “Joint beam training and positioning for
3115
+ intelligent reflecting surfaces assisted millimeter wave communications,”
3116
+ IEEE Transactions on Wireless Communications, vol. 20, no. 10, pp.
3117
+ 6282–6297, 2021.
3118
+ [20] A.
3119
+ E.
3120
+ El-Sayed,
3121
+ A.
3122
+ I.
3123
+ Sahar,
3124
+ and
3125
+ Y.
3126
+ A.
3127
+ Yassmen,
3128
+ “Moment
3129
+ generating function of the unbalanced non-central chi-square distribution,”
3130
+ International Journal of Engineering, vol. 4, no. 3, p. 8269, 2013.
3131
+ [21] M. Ghosh, “Exponential tail bounds for chisquared random variables,”
3132
+ Journal of Statistical Theory and Practice, vol. 15, no. 2, pp. 1–6, 2021.
3133
+
LtE1T4oBgHgl3EQftAUX/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
MdE4T4oBgHgl3EQfjA06/content/2301.05138v1.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e400011ee067cd1003f61cf68685774149e9b19ace08a122d62bde10cb8b0f10
3
+ size 221759
MdE4T4oBgHgl3EQfjA06/vector_store/index.faiss ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24099b49eb603882b655e324ecea415f4bab11ccfdb22544dd84a2b813b07035
3
+ size 1179693
MdE4T4oBgHgl3EQfjA06/vector_store/index.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f2d2f1a3ed4ada774fe78738fca0f20c6423c42af47f6a72a772f0855b0ced5
3
+ size 57419
NNFJT4oBgHgl3EQfzi34/content/tmp_files/2301.11644v1.pdf.txt ADDED
@@ -0,0 +1,1013 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Realization of valley-spin polarized current via parametric pump in monolayer MoS2
2
+ Kai-Tong Wang,1, 2 Hui Wang,2 Fuming Xu,1, ∗ Yunjin Yu,1 and Yadong Wei1
3
+ 1College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, China
4
+ 2School of Physics and Engineering, Henan University of Science and Technology, Luoyang 471023, China
5
+ Monolayer MoS2 is a typical valleytronic material with valley-spin locked valence bands.
6
+ We
7
+ numerically investigate the valley-spin polarized current in monolayer MoS2 via adiabatic electron
8
+ pumping. By introducing an exchange field to break the energy degeneracy of monolayer MoS2, the
9
+ top of its valence bands is valley-spin polarized and tunable by the exchange field. A device with
10
+ spin-up polarized left lead, spin-down polarized right lead, and untuned central region is constructed
11
+ through applying different exchange fields in the corresponding regions. Then, equal amount of
12
+ pumped currents with opposite valley-spin polarization are simultaneously generated in the left and
13
+ right leads when periodically varying two pumping potentials. Numerical results show that the phase
14
+ difference between the pumping potentials can change the direction and hence polarization of the
15
+ pumped currents. It is found that the pumped current exhibits resonant behavior in the valley-spin
16
+ locked energy window, which depends strongly on the system size and is enhanced to resonant current
17
+ peaks at certain system lengths. More importantly, the pumped current periodically oscillates as
18
+ a function of the system length, which is closely related to the oscillation of transmission. The
19
+ effects of other system parameters, such as the pumping amplitude and the static potential, are also
20
+ thoroughly discussed.
21
+ I.
22
+ INTRODUCTION
23
+ Valleytronics has attracted enormous attention on ac-
24
+ count of its potential for information processing1–16. In
25
+ many crystalline materials, there are two or more min-
26
+ ima(maxima) at the conduction(valence) band in the mo-
27
+ mentum space, known as valleys.
28
+ The degenerate but
29
+ inequivalent valley states constitute new pseudospin de-
30
+ gree of freedom for low energy carriers. Similar to spin-
31
+ tronics, the essential of valleytronics is to generate and
32
+ manipulate valley polarization to encode and store infor-
33
+ mation. Various materials have been explored to real-
34
+ ize valley polarization, including silicon17,18, bismuth19,
35
+ diamond20,21, carbon nanotube22,23, etc. In particular,
36
+ two-dimensional(2D) honeycomb lattice materials such
37
+ as graphene or transition metal dichalcogenides (TMDs)
38
+ provide a perfect platform to investigate valleytronics.
39
+ Compared to graphene, TMDs labeled as MX2 (M =
40
+ Mo, W, X = S, Se, Te), also have two well-separated val-
41
+ leys in the Brillouin zone2,24. However, due to inversion
42
+ symmetry breaking, TMDs are natural gapped semicon-
43
+ ductors, which makes TMDs the promising candidates of
44
+ valleytronic materials25–31.
45
+ As a typical TMDs material, monolayer MoS2 has a
46
+ strong spin-orbit coupling(SOC) interaction26,32, which
47
+ leads to the locking between valley and spin at the top of
48
+ its valance band. The valley-spin locking means that the
49
+ valley and spin can be polarized together, and the lifetime
50
+ of polarization can be enhanced due to the large spacing
51
+ between K and K′ valleys. In the presence of an exchange
52
+ field, TMDs exhibit interesting phenomena, such as the
53
+ quantum anomalous Hall effect33,34, spin and valley Hall
54
+ effects35, and unconventional superconductivity36.
55
+ Be-
56
+ sides, an exchange field can induce polarized valleys,
57
+ which can be inverted by tuning the spin polarization.
58
+ Through the ferromagnetic proximity effect37,38 or mag-
59
+ netic doping39,40, the exchange field can be introduced
60
+ into TMDs materials, which provides an effective way to
61
+ manipulate its valley/spin degree of freedom. In exper-
62
+ iments, the exchange field for valley splitting has been
63
+ realized in Fe-doped41 or Co-doped monolayer MoS2.42
64
+ EuS as a ferromagnetic substrate can efficiently induce
65
+ the magnetic exchange field in monolayer TMDs.43,44
66
+ Based on the valley optical selection rules, the opti-
67
+ cal pumping of valley polarization has been experimen-
68
+ tally realized by circular polarized light in 2D TMDs45–47.
69
+ Very recently, the spin-valley coupled dynamics at the
70
+ MoS2-MoSe2 interface is experimentally studied using
71
+ optical pumping48; photoinduced valley-selective polar-
72
+ ization in monolayer WS2 has been realized with cir-
73
+ cularly polarized light pumping.49 Besides, The line
74
+ defects50, nonmagnetic disorders51, and spatially vary-
75
+ ing potentials16 were predicted to achieve the valley po-
76
+ larization in monolayer MoS2. In terms of applications,
77
+ it is desirable to obtain pure valley polarized current
78
+ by electrical methods.
79
+ Accordingly, we propose that
80
+ quantum parametric pump can drive valley and spin po-
81
+ larized currents in monolayer MoS2 through adiabati-
82
+ cally varying two gate voltages. The parametric pump
83
+ can produce dc current by periodically varying system
84
+ parameters, which has been generalized to various 2D
85
+ materials52–55. Specially, spin pump has been reported
86
+ in several nanostructures56–58, where pure spin current
87
+ and zero charge current are obtained.
88
+ In this paper, we numerically study the generation and
89
+ manipulation of valley-spin polarized currents via adia-
90
+ batic pump in monolayer MoS2.
91
+ The system setup is
92
+ shown in Fig.1. By magnetic doping, an exchange inter-
93
+ action is introduced in the left and right leads, which in-
94
+ duces locked valley-spin polarization at the top of valence
95
+ band as shown in Fig.1(a) and 1(c). When the pumping
96
+ potentials periodically change, fully valley-spin polarized
97
+ dc currents are driven into the leads. At one moment,
98
+ the current with K valley and spin up is pumped into the
99
+ arXiv:2301.11644v1 [cond-mat.mes-hall] 27 Jan 2023
100
+
101
+ 2
102
+ FIG. 1: Schematics of the band structures of monolayer MoS2
103
+ for (a) with exchange field M, (b) without exchange field, (c)
104
+ with exchange field −M. The red and blue valance bands de-
105
+ note valley K with spin up and K′ with spin down. (d) The
106
+ pump setup based on monolayer MoS2 consisting of left/right
107
+ leads and the scattering region, whose band structures are
108
+ correspondingly shown in (a) to (c), respectively. The MoS2
109
+ lattice is represented by the simple honeycomb lattice. The
110
+ pumping potentials V1 and V2 are added in the scattering re-
111
+ gion, adjacent to the leads. As V1 and V2 periodically change,
112
+ electric currents with opposite valley-spin polarizations are si-
113
+ multaneously pumped into the left and right leads, as shown
114
+ by the block arrows.
115
+ left lead while the current with opposite valley-spin po-
116
+ larization flows into the right lead. The polarized current
117
+ exhibits resonant behavior in the valley-spin locked en-
118
+ ergy window, which mainly depends on the system size.
119
+ With the increasing of the system length, the pumped
120
+ currents show periodic oscillation behavior and robust
121
+ resonant current peaks can be observed. We also inves-
122
+ tigate the influence of other system parameters, includ-
123
+ ing the phase difference, the Rashba SOC strength, the
124
+ static potential, and the Fermi energy. It is found that
125
+ the phase difference and static potentials can invert the
126
+ direction and hence polarization of the pumped current.
127
+ The paper is organized as follows. In Sec. II, we in-
128
+ troduce the Hamiltonian of monolayer MoS2 and the for-
129
+ malism of adiabatic parametric pumping. In Sec. III,
130
+ numerical results and relevant discussions are presented.
131
+ Finally, a brief summary is given in Sec. IV.
132
+ II.
133
+ MODEL AND FORMALISM
134
+ In monolayer MoS2, the low-energy spectrum at K
135
+ and K′ valleys consists of three d orbitals of Mo, i.e.,
136
+ dz2, dx2−y2, dxy.
137
+ The relations between these orbitals
138
+ and basis wave functions satisfy: |ϕc⟩ = |dz2⟩, |ϕλ
139
+ υ⟩ =
140
+ (|dx2−y2⟩+iλ|dxy⟩)/
141
+
142
+ 2, where the subscript c/υ denotes
143
+ the conduction/valence band and λ = ±1 corresponds
144
+ to different valleys K and K′. Based on above low-lying
145
+ states, the effective Hamiltonian of monolayer MoS2 has
146
+ the following form26,59
147
+ H0(k) = at(λkxσx + kyσy) + ∆σz − tSOλσz − 1
148
+ 2
149
+ τz, (1)
150
+ where a and t are the lattice constant and hopping
151
+ strength, respectively. σx,y,z and τz represent the Pauli
152
+ matrices of basis functions(|ϕc⟩ and |ϕυ⟩) and spin(↑ and
153
+ ↓). ∆ is the mass term and the last term is the intrinsic
154
+ SOC with strength tSO.
155
+ We employ the tight-binding model of MoS2, which
156
+ treats monolayer MoS2 as a simplified honeycomb lattice.
157
+ The lattice includes A and B sublattices, corresponding
158
+ to the dz2 orbit and dx2−y2 + iλdxy orbits of Mo, respec-
159
+ tively. In the tight-binding approximation, the Hamilto-
160
+ nian can be expressed as60,61
161
+ H0 =
162
+
163
+ i
164
+ ϵic†
165
+ iαciα + t
166
+
167
+ <i,j>
168
+ c†
169
+ iαciα + HSO,
170
+ (2)
171
+ with
172
+ HSO = 2itSO
173
+ 3
174
+
175
+ 3
176
+
177
+ ≪i,j≫,α,α′
178
+ υijc†
179
+ iατz,αα′cjα′,
180
+ (3)
181
+ where c†
182
+ iα(ciα) is the creation(annihilation) operator at
183
+ site i with spin α = ±1, ϵi is the on-site energy. HSO
184
+ denotes the intrinsic SOC term and the summation over
185
+ the second nearest-neighbor sites only involves B sublat-
186
+ tice. Besides, υij = +1(−1) if an electron moves from
187
+ site j to site i with taking a left(right) turn62.
188
+ Based on this model, we consider a monolayer MoS2
189
+ setup, which contains three parts: the central scattering
190
+ region, the left and right leads, as shown in Fig.1(d).
191
+ By magnetic doping, different valley-spin polarizations
192
+ can be induced in the left and right leads due to the
193
+ exchange field39,63. The schematic band structures with
194
+ or without the exchange field are depicted in Fig.1(a)-(c).
195
+ In the presence of Rashba spin-orbit coupling (RSOC),
196
+ the Hamiltonians for the scattering region with pumping
197
+ potentials and the leads can be written as
198
+ HC = H0+ 3itR
199
+ 4
200
+
201
+ <i,j>,α,α′
202
+ (ταα′×dij)zc†
203
+ iαcjα′+V (x, y, t),
204
+ (4)
205
+ HL/R = H0 ± M
206
+
207
+ i,α,α′
208
+ τz,αα′c†
209
+ iαciα′,
210
+ (5)
211
+ where M and tR denote the strengths of the exchange
212
+ field and RSOC, respectively. ταα′ = (τx, τy, τz) is the
213
+ Pauli matrix for spin, and dij is the lattice vector con-
214
+ necting sites i and j.
215
+ The potential term V (x, y, t) =
216
+ Vs(x, y) + Vt(x, y, t), where Vs = V0
217
+
218
+ i Πi(x, y) corre-
219
+ sponds to the static potential defining the shape of the
220
+ pumping region. Vt = Vp
221
+
222
+ i Πi(x, y)cos(ωt + ϕi) is the
223
+ periodic pumping potential. V0 and Vp are the ampli-
224
+ tudes of Vs and Vt. i = 1, 2 are the indices of the po-
225
+ tential and Πi represents the potential profile, which is
226
+
227
+ (a)
228
+ (b)
229
+ (c)
230
+ :
231
+ spin up
232
+ spin down
233
+ K
234
+ K'
235
+ K
236
+ K'
237
+ K
238
+ K
239
+ (d)
240
+ 个,K
241
+ *,K'
242
+ Lead-L
243
+ Vi(t)
244
+ Scattering region
245
+ V2(t)
246
+ Lead-R3
247
+ highlighted in green in Fig.1(d). ϕi is the initial phase of
248
+ the pumping potential.
249
+ To evaluate the adiabatic valley-spin pump, we need to
250
+ calculate the average current flowing into lead β. Con-
251
+ sider a slowly varying time-dependent pumping potential
252
+ Vt,i, the average current in one period is expressed as64
253
+ Iβ = qω
254
+
255
+ � T
256
+ 0
257
+ dt[ dNβ
258
+ dVt,1
259
+ dVt,1
260
+ dt
261
+ + dNβ
262
+ dVt,2
263
+ dVt,2
264
+ dt ],
265
+ (6)
266
+ where the period of Vt,i is T = 2π/ω with frequency ω
267
+ and β = L/R labels the lead.
268
+ The emissivity
269
+ dNβ
270
+ dVi
271
+ is
272
+ defined in terms of the scattering matrix Sββ′ as65,66
273
+ dNβ
274
+ dVi
275
+ =
276
+ � dE
277
+ 2π (−∂Ef)
278
+
279
+ β′
280
+ Im∂Sββ′
281
+ ∂Vi
282
+ S∗
283
+ ββ′,
284
+ (7)
285
+ with f the Fermi distribution function. Under the adi-
286
+ abatic condition, the pumped current is independent of
287
+ the pumping frequency ω, hence we set ω = 1 in the
288
+ calculation.
289
+ In the language of nonequilibrium Green’s functions,
290
+ the pumped current is expressed as67–69
291
+ Iβ = − q
292
+
293
+ � 2π
294
+ 0
295
+ dt
296
+
297
+ dE(∂Ef)Tr[ΓβGr dVt
298
+ dt Ga].
299
+ (8)
300
+ Here Gr/Ga is the retarded/advanced Green’s function
301
+ of the central scattering region, which is defined as Gr =
302
+ Ga,† = [E − HC − �
303
+ β Σr
304
+ β]−1. HC is the corresponding
305
+ Hamiltonian. Σr
306
+ β is the retarded self-energy of lead β,
307
+ which can be calculated by surface Green’s function70,71.
308
+ Γβ = i(Σr
309
+ β − Σa
310
+ β) denotes the linewidth function.
311
+ As shown by block arrows in Fig.1(d), at one moment
312
+ of the pumping period, polarized current with K valley
313
+ and spin up (pink arrow) is driven into the left lead, and
314
+ equal amount current with K′ valley and spin down (blue
315
+ arrow) flows in the right lead. Detailed numerical results
316
+ are shown in the following section.
317
+ We use the short
318
+ term, the pumped current, to stand for the valley-spin
319
+ polarized currents in the leads.
320
+ III.
321
+ RESULTS AND DISCUSSION
322
+ In the calculations, the on-site energy is ϵi = ±0.83 eV
323
+ for A and B sublattices. Other parameters26,60 are set
324
+ as t = 1.27 eV, tSO = 0.038 eV, and the lattice constant
325
+ a=0.32 nm. Without loss of generality, we set the ex-
326
+ change field strength M = 0.06 eV, and eV is taken as
327
+ the energy unit throughout the calculation. The periodic
328
+ boundary condition (PBC) is considered for monolayer
329
+ MoS2, and thus the edge effect is removed. To realize
330
+ PBC, the upper and lower edges of a zigzag MoS2 rib-
331
+ bon shown in Fig.1(d) are connected with appropriate
332
+ hopping interactions, which is also called the cylinder
333
+ boundary.
334
+ FIG. 2: The dispersion relation of monolayer MoS2 ribbon
335
+ with an exchange field: (a) M = 0.06, (b) M = −0.06. Both
336
+ valley and spin are polarized together, where ∆E is defined
337
+ as the valley-spin locked energy window.
338
+ A.
339
+ Valley-spin polarized current
340
+ The dispersion of monolayer MoS2 with an exchange
341
+ field is plotted in Fig.2. From Fig.2(a), it is clear that
342
+ the valley K with spin up is polarized at the valence band
343
+ top. However, as the exchange field changes, the polar-
344
+ ization of both valley and spin is inverted as shown in
345
+ Fig.2(b). In the valley-spin locked window ∆E, perfect
346
+ polarization can be realized. To investigate the valley-
347
+ spin polarized current, we consider only the ∆E energy
348
+ range. In Fig.3, We study the dependence of the pumped
349
+ current on the phase difference ϕ12 between V1 and V2.
350
+ Due to the inverse valley-spin polarization of the left and
351
+ right leads, the holes are forbidden to propagate through
352
+ the scattering region, so there is no pumped current gen-
353
+ erated in this setup at tR = 0.
354
+ Introducing RSOC in
355
+ the scattering region, it is found that the pumped cur-
356
+ rent arises and flows into left or right lead. We attribute
357
+ such a dc current to the spin flip process induced by the
358
+ RSOC. Importantly, nonzero valley-spin polarized cur-
359
+ rents are pumped into different leads, which depends on
360
+ the exchange field in leads. Our results show that IL,R is
361
+ an odd function about ϕ12, i.e., I(ϕ12) = −I(−ϕ12). The
362
+ maximum of the pumped current appears at ϕ12 = ±π/2.
363
+ From Fig.3, when the phase difference ϕ12 shifts from
364
+ −π to 0, the valley-polarized holes with spin up will be
365
+ pumped out of the scattering region and flow into the
366
+ left lead. On the contrary, the opposite valley-polarized
367
+ holes with spin down will spread into the right lead when
368
+ ϕ12 shifts from 0 to π. It means that the direction of the
369
+ pumped current can be tuned by the phase difference ϕ12,
370
+ then different valley-spin polarized currents are pumped
371
+ into different leads. We calculate the pumped current as
372
+ a function of the phase difference ϕ12 for different RSOC
373
+ strengths tR. Consequently, with the increasing of tR,
374
+ the pumped current increases. IL versus tR at ϕ12 = π
375
+ 2 is
376
+ plotted in the inset. The result is understandable: since
377
+ the spin-flip efficiency increases when tR is increased, the
378
+ spin-up carriers from one lead can be more easily flipped
379
+ as spin-down carriers and flow into the other lead, which
380
+
381
+ 1.5
382
+ (a)
383
+ (b)
384
+ 1
385
+ 0.5
386
+ spin up
387
+ E(eV)
388
+ spin down
389
+ 0
390
+ -0.5
391
+ K
392
+ K'
393
+ K
394
+ K'
395
+ △E
396
+ -1
397
+ -1.5
398
+ 0
399
+ 0.5
400
+ 1
401
+ 1.5
402
+ 2 0
403
+ 0.5
404
+ 1
405
+ 1.5
406
+ 2
407
+ k(π/a)
408
+ k (π/a)
409
+ X4
410
+ -3
411
+ -2
412
+ -1
413
+ 0
414
+ 1
415
+ 2
416
+ 3
417
+ -0.003
418
+ -0.002
419
+ -0.001
420
+ 0.000
421
+ 0.001
422
+ 0.002
423
+ 0.003
424
+ 0.00
425
+ 0.02
426
+ 0.04
427
+ 0.06
428
+ 0.08
429
+ 0.10
430
+ 0.000
431
+ 0.002
432
+ 0.004
433
+ 0.006
434
+ 0.008
435
+ ϕ12
436
+ Pumped current
437
+ tR=0
438
+ tR=0.02
439
+ tR=0.04
440
+ tR=0.05
441
+ Pumped current
442
+ tR
443
+ FIG. 3: The pumped current IL as a function of the phase
444
+ difference ϕ12 between V1 and V2. Inset: the pumped current
445
+ versus RSOC strength tR at ϕ12 = π/2. Other parameters:
446
+ Ef = −0.74, L = 10a, V0 = 0.1, Vp = 0.05.
447
+ -0.76
448
+ -0.75
449
+ -0.74
450
+ -0.73
451
+ -0.72
452
+ -0.71
453
+ -0.70
454
+ 0.000
455
+ 0.002
456
+ 0.004
457
+ 0.006
458
+ 0.00
459
+ 0.02
460
+ 0.04
461
+ 0.06
462
+ 0.08
463
+ -0.006
464
+ -0.003
465
+ 0.000
466
+ 0.003
467
+ 0.006
468
+ IL
469
+ (b)
470
+ Transmission
471
+ Pumped current
472
+ Ef
473
+ (a)
474
+ 0.00
475
+ 0.03
476
+ 0.06
477
+ 0.09
478
+ T
479
+
480
+ Pumped current
481
+ Vp
482
+ IL,V0=0.05
483
+ IR,V0=0.05
484
+ IL,V0=0.07
485
+ IR,V0=0.07
486
+ IL,V0=0.1
487
+ IR,V0=0.1
488
+ FIG. 4:
489
+ (a) The pumped current and transmission versus
490
+ Fermi energy at V0 = 0.1 and Vp = 0.05. (b) The polarized
491
+ current IL and IR versus the pumping potential Vp for dif-
492
+ ferent static potentials V0 at Ef = −0.74. Other parameters:
493
+ ϕ12 = π/2, L = 10a, tR = 0.05.
494
+ is required by the conservation of charge.
495
+ In Fig.4(a), the pumped current and the transmission
496
+ versus Fermi energy Ef are plotted at ϕ12 =
497
+ π
498
+ 2 .
499
+ Ob-
500
+ viously, a broad transmission peak arises in the locked
501
+ window ∆E, and the pumped current exhibits similar
502
+ behavior as the transmission. This result indicates that
503
+ the transport of polarized holes is dominated by quan-
504
+ tum resonance, which originates from quantum interfer-
505
+ ence effect. In fact, the resonance assisted transport is a
506
+ common property of electron pump66. Besides, an im-
507
+ passable interval for the pumped current is generated
508
+ as the Fermi energy is away from the resonant peak as
509
+ 10
510
+ 20
511
+ 30
512
+ 40
513
+ 50
514
+ 0.00
515
+ 0.02
516
+ 0.04
517
+ 0.06
518
+ 10
519
+ 20
520
+ 30
521
+ 40
522
+ 50
523
+ 0.000
524
+ 0.002
525
+ 0.004
526
+ 0.006
527
+
528
+ (a)
529
+ Transmission
530
+ 46
531
+ 34
532
+ 22
533
+ 10
534
+ (b)
535
+ Pumped current
536
+ L/a
537
+ tR=0.02
538
+ tR=0.05
539
+ tR=0.08
540
+ FIG. 5: (a) The transmission versus the length L of scattering
541
+ region for tR = 0.05. (b) The pumped current IL as a function
542
+ of system length L for different RSOC strengths at ϕ12 =
543
+ π/2. The numbers label the lengths of the scattering region
544
+ where current peaks emerge. Other parameters are the same
545
+ as Fig.4.
546
+ shown in Fig.4(a).
547
+ The variation of polarized current
548
+ is well correlated to the transmission. The pumped cur-
549
+ rent IL,R versus the pumping potential for different static
550
+ potentials is plotted in Fig.4(b). Due to the particle con-
551
+ servation, the current flowing into the scattering region
552
+ must satisfy IL = −IR, which is confirmed through the
553
+ symmetric curves of IL and IR. Furthermore, the results
554
+ show that the valley-spin polarized current linearly in-
555
+ creases with the increasing of pumping potential. For a
556
+ relatively small Vp, the static potential V0 can enhance
557
+ the magnitude of pumped currents.
558
+ B.
559
+ Size effect on the pumped current
560
+ In the following, we study the influence of the sys-
561
+ tem size on the pumped current. Transmission and the
562
+ pumped current versus the length of the scattering re-
563
+ gion are plotted in Fig.5.
564
+ It is interesting that some
565
+ robust peaks of the pumped current appear at certain
566
+ lengths, but the currents for other lengths are almost
567
+ zero.
568
+ Moreover, these peaks show periodic oscillation
569
+ behavior with the period length 12a.
570
+ In Fig.5(a), we
571
+ plot the transmission as a function of the length L for
572
+ tR = 0.05.
573
+ The transmission exhibits a similar be-
574
+ havior as the pumped current, where the transmission
575
+ peaks also appear at certain system lengths. It is clear
576
+ that these transmission peaks correspond exactly to the
577
+ pumped current peaks. It is reasonable that the peri-
578
+ odic behavior of the pumped current originates from the
579
+ spin precession72–74 induced by Rashba SOC. When the
580
+ current carriers travel through the central scattering re-
581
+
582
+ 5
583
+ 0.000
584
+ 0.002
585
+ 0.004
586
+ 0.000
587
+ 0.001
588
+ 0.002
589
+ 10
590
+ 20
591
+ 30
592
+ 40
593
+ 50
594
+ 0.000
595
+ 0.001
596
+ 0.002
597
+ IL
598
+ (c)
599
+ (b)
600
+
601
+ Pumped current
602
+ (a)
603
+ 0.00
604
+ 0.05
605
+ 0.10
606
+ T
607
+
608
+
609
+ Pumped current
610
+ 0.00
611
+ 0.05
612
+ 0.10
613
+
614
+ Transmission
615
+ Transmission
616
+ Transmission
617
+ Pumped current
618
+ L/a
619
+ 0.00
620
+ 0.05
621
+ 0.10
622
+
623
+ FIG. 6: The pumped current and transmission versus the
624
+ length L of the scattering region for different Fermi energies.
625
+ (a): Ef = −0.72, (b): Ef = −0.735, (c): Ef = −0.75. Other
626
+ parameters are the same as Fig.4.
627
+ gion with RSOC interaction, carrier spin keeps precess-
628
+ ing and spin flip occurs, which results in the periodic
629
+ oscillating behavior of the pumped current. Our calcu-
630
+ lation further demonstrates that the width of scattering
631
+ region has almost no influence on the periodic behavior
632
+ of polarized current as long as Fermi energy lies in the
633
+ valley-spin locked energy window. To evaluate the influ-
634
+ ence of RSOC, We show in Fig.5(b) the pumped currents
635
+ for different RSOC strengths. It is clear that the RSOC
636
+ strength has less influence on the resonant period, which
637
+ is L = 12a. However, with the increasing of tR, the reso-
638
+ nant current peaks become significant. The peak current
639
+ value grows larger as tR increases, which is consistent
640
+ with the results shown in Fig.3.
641
+ Notice that this periodic oscillation behavior of the
642
+ pumped current is different from the even-odd conduc-
643
+ tance oscillation of carbon-atom chains.75,76 When two
644
+ metallic electrodes are attached to a carbon atomic chain,
645
+ the electronic structure of the carbon chain is modified,
646
+ which leads to the difference in the density of states be-
647
+ tween odd- and even-number carbon chains.75,76 This
648
+ leads to the even-odd conductance oscillation driven by
649
+ dc bias.
650
+ However, for the periodic oscillation of the
651
+ valley-spin polarized currents, it is due to the spin-
652
+ flipping process induced by Rashba SOC and driven by
653
+ periodic pumping potentials.
654
+ In Fig.6, we plot the pumped current and transmission
655
+ versus the system length for different Fermi energies. Ap-
656
+ parently, the periodic oscillation behavior of the pumped
657
+ current persists as the Fermi energy changes.
658
+ The re-
659
+ sult shows that, with the increasing of Ef, the number of
660
+ peaks increases while the corresponding periodic length
661
+ FIG. 7: The pumped current with respect to both the system
662
+ length L and the Fermi energy Ef.
663
+ decreases. Besides, the magnitude of current peaks will
664
+ decrease as the Fermi energy increases. By calculating
665
+ the transmission, it can be seen that the behavior of the
666
+ pumped current is still consistent with the transmission,
667
+ which means the periodic oscillation is a universal phe-
668
+ nomenon in the setup.
669
+ To exhibit an overall view of the pumped current,
670
+ in Fig.7, we provide a two-dimensional diagram of the
671
+ valley-spin polarized current as a function of both the
672
+ system length L and the Fermi energy Ef.
673
+ By vary-
674
+ ing L and Ef, we find that there are five curves with
675
+ discrete extrema of currents. The periodic dependence
676
+ of the pumped current on the system length L is clearly
677
+ shown. When Ef approaches the top of valley, the largest
678
+ pumped current appears and becomes more sharp as
679
+ shown by the orange regions.
680
+ Moreover, it is further
681
+ confirmed that, multiple resonant peaks of the pumped
682
+ current arise under an appropriate system length L. Our
683
+ numerical results reveal that it is possible to design a
684
+ high-efficiency device setup for generating valley-spin po-
685
+ larized currents.
686
+ C.
687
+ Influence of the static potential
688
+ In this section, we focus on the dependence of IL on
689
+ the static potential V0.
690
+ For this purpose, the system
691
+ length and the RSOC strength are fixed at L = 10a and
692
+ tR = 0.05. In Fig.8(a), we plot the pumped current as
693
+ well as transmission coefficient versus the static potential
694
+ V0. With the increasing of the static potential, a current
695
+ peak first emerges at V0 = 0.126 labeled by the blue point
696
+ γ3. As V0 scans the critical point γ1 at V0 = 0.151, the
697
+ direction of IL is reversed. Continuing to increase V0, we
698
+ can see a negative current peak. The result shows that
699
+ the static potential can also change the direction and
700
+ hence the polarization of the pumped current. Besides,
701
+ in the vicinity of the critical point γ1, a transmission
702
+
703
+ 0.008
704
+ 50
705
+ 0.006
706
+ 0.004
707
+ 40
708
+ 0.002
709
+ a
710
+ 30
711
+ 0
712
+ 20
713
+ 10.
714
+ -0.76
715
+ -0.74
716
+ -0.72
717
+ -0.706
718
+ 0.00
719
+ 0.06
720
+ 0.12
721
+ 0.18
722
+ 0.24
723
+ -0.002
724
+ 0.000
725
+ 0.002
726
+ 0.004
727
+ -0.750
728
+ -0.745
729
+ -0.740
730
+ -0.735
731
+ -0.730
732
+ 0.1
733
+ 0.2
734
+ 0.3
735
+ 0.4
736
+ γ3
737
+ γ2
738
+ Pumped current
739
+ V0
740
+ IL
741
+ T/50
742
+ (a)
743
+ γ1
744
+ (b)
745
+ V0&T
746
+ Ef
747
+ γ1 transition V0
748
+ γ2 maximum T
749
+ 0.000
750
+ 0.002
751
+ 0.004
752
+ 0.006
753
+ 0.008
754
+ Maximum IL
755
+ γ3 maximum IL
756
+ FIG. 8: (a) The pumped current as well as the transmission
757
+ coefficient T as a function of the static potential V0. A factor
758
+ of 1/50 is multiplied to T for better illustration. γ1, γ2, γ3
759
+ label the critical points. (b) The maximum of IL, transition
760
+ point of V0 and corresponding maximum of T versus the Fermi
761
+ energy. Other parameters: Vp = 0.03, ϕ12 = π/2, L = 10a.
762
+ peak is clear, which suggests the influence of the static
763
+ potential also results from quantum resonance.
764
+ In Fig.8(b), The critical point of V0 and the maxima
765
+ of both T and IL versus Fermi energy Ef are plotted.
766
+ It is found that the critical value of V0 increases linearly
767
+ with the increasing of Ef, which indicates the resonant
768
+ energy level depends on the static potential. Besides, the
769
+ curves of the peak values for transmission and pumped
770
+ current grow with the Fermi energy. The variation of the
771
+ pumped current with the Fermi energy is consistent with
772
+ the results in Fig.6.
773
+ In this work, the valley-spin polarized current is gener-
774
+ ated by electric pumping in adiabatic regime, which re-
775
+ quires two independently varying system parameters. We
776
+ emphasize that similar mechanism can also be achieved
777
+ with optical pumping, which is in the non-adiabatic
778
+ regime due to the high frequency of light wave. In this
779
+ case, the light frequency can serve as a pumping param-
780
+ eter. Therefore, non-adiabatic parametric pumping us-
781
+ ing electric or optical ways will certainly bring in more
782
+ physics.
783
+ IV.
784
+ CONCLUSIONS
785
+ In conclusion, we study the valley-spin polarized cur-
786
+ rent in monolayer MoS2 ribbon via parametric electron
787
+ pump. In the proposed setup, different valley-spin polar-
788
+ ized currents can be controlled to flow into different leads
789
+ in the valley-spin locked energy window. The phase dif-
790
+ ference between the pumping potentials can change the
791
+ direction and hence polarization of the pumped current,
792
+ where quantum resonance dominates the transport pro-
793
+ cess. Furthermore, the size effect on the valley-spin po-
794
+ larized current is numerically investigated. As the length
795
+ of the scattering region changes, resonant peaks of the
796
+ pumped current arise and show periodic oscillation due
797
+ to the spin precession.
798
+ With the increasing of Fermi
799
+ energy, the number of peaks decreases while the peak
800
+ height increases within a fixed length range. The depen-
801
+ dence of the resonant pumped current on the scattering
802
+ region length and the Fermi energy is numerically re-
803
+ vealed in a two-dimensional diagram.
804
+ It is also found
805
+ that the direction of valley-spin polarized currents can
806
+ be inverted by the static pumping potential. As a po-
807
+ tential valleytronic device, the pump setup proposed in
808
+ this work can serve as a valley-spin polarization source,
809
+ which can simultaneously generate opposite polarization
810
+ in one device.
811
+ The polarized signals can be efficiently
812
+ manipulated by many system parameters, and significant
813
+ resonant enhancement has been demonstrated.
814
+ ACKNOWLEDGMENTS
815
+ This
816
+ work
817
+ was
818
+ supported
819
+ by
820
+ the
821
+ National
822
+ Natural
823
+ Science
824
+ Foundation
825
+ of
826
+ China
827
+ (Grant
828
+ Nos.
829
+ 12034014
830
+ and
831
+ 61674052)
832
+ and
833
+ the
834
+ Natu-
835
+ ral
836
+ Science
837
+ Foundation
838
+ of
839
+ Shenzhen
840
+ (Grant
841
+ Nos.
842
+ 20200812092737002,
843
+ JCYJ20190808115415679,
844
+ and
845
+ JCYJ20190808152801642). Hui Wang also acknowledges
846
+ supports from the Outstanding Youth Foundation of
847
+ Henan Scientific Committee (212300410041) and the
848
+ Key Scientific and Technological Projects in Henan
849
+ Province (212102210223).
850
851
+ 1 A. Rycerz, J. Tworzyd�lo, and C. W. J. Beenakker, Nat.
852
+ Phys., 2007, 3, 172-175.
853
+ 2 D. Xiao, W. Yao, and Q. Niu, Phys. Rev. Lett., 2007, 99,
854
+ 236809.
855
+ 3 S.-G. Cheng, J. Zhou, H. Jiang, and Q.-F. Sun, New J.
856
+ Phys., 2016, 18, 103024.
857
+ 4 J. R. Schaibley, H. Yu, G. Clark, P. Rivera, J. S. Ross,
858
+ K. L. Seyler, W. Yao, and X. Xu, Nat. Rev. Mater., 2016,
859
+ 1, 16055.
860
+ 5 D. Gunlycke and C. T. White, Phys. Rev. Lett., 2011, 106,
861
+ 136806.
862
+ 6 Y. S. Ang, S. A. Yang, C. Zhang, Z. Ma, and L. K. Ang,
863
+ Phys. Rev. B, 2017, 96, 245410.
864
+ 7 X. Chen, Y. Xu, J. Wang, and H. Guo, Phys. Rev. B, 2019,
865
+ 99, 064302.
866
+
867
+ 7
868
+ 8 Z.-M. Yu, S. Guan, X.-L. Sheng, W. Gao, and S. A. Yang,
869
+ Phys. Rev. Lett., 2020, 124, 037701.
870
+ 9 L. Zhang, K. Gong, J. Chen, L. Liu, Y. Zhu, D. Xiao, and
871
+ H. Guo, Phys. Rev. B, 2014, 90, 195428.
872
+ 10 Z. Yu, F. Xu, and J. Wang, Carbon, 2016, 99, 451.
873
+ 11 L. Zhang, Z. Yu, F. Xu, and J. Wang, Carbon, 2018, 126,
874
+ 183.
875
+ 12 Z. P. Niu, New J. Phys., 2018 20, 103021.
876
+ 13 J. Pawlowski, D. Zebrowski, and S. Bednarek, Phys. Rev.
877
+ B, 2018, 97, 155412.
878
+ 14 S.-G. Cheng, H. Liu, H. Jiang, Q.-F. Sun, and X. C. Xie,
879
+ Phys. Rev. Lett., 2018, 121, 156801.
880
+ 15 C. Lei, Y. Ma, T. Zhang, X. Xu, B. Huang, and Y. Dai,
881
+ New J. Phys., 2020 22, 033002.
882
+ 16 D. Gut, M. Prokop, D. Sticlet, and M. P. Nowak, Phys.
883
+ Rev. B, 2020, 101, 085425.
884
+ 17 K. Takashina, Y. Ono, A. Fujiwara, Y. Takahashi, and
885
+ Y. Hirayama, Phys. Rev. Lett., 2006 96, 236801.
886
+ 18 D. Culcer, A. L. Saraiva, B. Koiller, X. Hu, and S. D.
887
+ Sarma, Phys. Rev. Lett., 2012, 108, 126804.
888
+ 19 Z. Zhu, A. Collaudin, B. Fauque, W. Kang, and K. Behnia,
889
+ Nat. Phys., 2012, 8, 89.
890
+ 20 C. E. Nebel, Nat. Mater., 2013, 12, 690.
891
+ 21 J. Isberg, M. Gabrysch, J. Hammersberg, S. Majdi, K. K.
892
+ Kovi, and D. J. Twitchen, Nat. Mater., 2013, 12, 760.
893
+ 22 F. Xu, Z. Yu, Y. Ren, B. Wang, Y. Wei, and Z. Qiao, New
894
+ J. Phys., 2016 18, 113011.
895
+ 23 A. Palyi and G. Burkard, Phys. Rev. Lett., 2011, 106,
896
+ 086801.
897
+ 24 X. Xu, W. Yao, D. Xiao, and T. F. Heinz, Nat. Phys., 2014
898
+ 10, 343.
899
+ 25 T. Cai, S. A. Yang, X. Li, F. Zhang, J. Shi, W. Yao, and
900
+ Q. Niu, Phys. Rev. B, 2013, 88, 115140.
901
+ 26 D. Xiao, G.-B. Liu, W. Feng, X. Xu, and W. Yao, Phys.
902
+ Rev. Lett., 2012, 108, 196802.
903
+ 27 P. E. Faria Junior, K. Zollner, T. Wo´zniak, M. Kurpas,
904
+ M. Gmitra, and J. Fabian, New J. Phys., 2022 24, 083004.
905
+ 28 Y. G. Semenov and K. W. Kim, Phys. Rev. B, 2016, 93,
906
+ 041414(R).
907
+ 29 D. Oliveira, L. V. Lelovsky, M. A. G. Soler, and F. Qu,
908
+ Phys. Rev. B, 2018, 97, 125309.
909
+ 30 H. Tornatzky, A. M. Kaulitz, and J. Maultzsch, Phys. Rev.
910
+ Lett., 2018, 121, 167401.
911
+ 31 K. F. Mak, K. L. Mcgill, J. Park, and P. L. Mceuen, Sci-
912
+ ence, 2014, 344, 1489.
913
+ 32 Z. Y. Zhu, Y. C. Cheng, and U. Schwingenschl¨ogl, Phys.
914
+ Rev. B, 2011, 84, 153402.
915
+ 33 R. Yu, W. Zhang, H.-J. Zhang, S.-C. Zhang, X. Dai, and
916
+ Z. Fang, Science, 2010, 329, 61.
917
+ 34 J. Zhou, Q. Sun, and P. Jena, Phys. Rev. Lett., 2017, 119,
918
+ 046403.
919
+ 35 H. Da, Q. Song, P. Dong, H. Ye, and X. Yan, J. Appl.
920
+ Phys., 2020, 127, 023903.
921
+ 36 M. A. Rahimi, A. G. Moghaddam, C. Dykstra, M. Gover-
922
+ nale, and U. Z¨ulicke, Phys. Rev. B, 2017, 95, 104515.
923
+ 37 L.-Y. Gan, Q. Zhang, Y. Cheng, and U. Schwingenschl¨ogl,
924
+ Phys. Rev. B, 2013, 88, 235310.
925
+ 38 S. Teng, X. Mao, Z. Liu, Y. Liu, X. Xu, L. Li, X. Xie,
926
+ S. Fan, G. Zhou, J. Li, and J. Li, New J. Phys., 2020 22,
927
+ 103061.
928
+ 39 Y. C. Cheng, Q. Y. Zhang, and U. Schwingenschl¨ogl, Phys.
929
+ Rev. B, 2014, 89, 155429.
930
+ 40 J. Zhou and P. Jena, J. Phys. Chem. Lett, 2017, 8, 5764-
931
+ 5770.
932
+ 41 Q. Li, X. Zhao, L. Deng, Z. Shi, S. Liu, Q. Wei, L. Zhang,
933
+ Y. Cheng, L. Zhang, H. Lu, W. Gao, W. Huang, C. W. Qiu,
934
+ G. Xiang, S. J. Pennycook, Q. Xiong, K. P. Loh, and B.
935
+ Peng, ACS Nano 2020, 14, 4636.
936
+ 42 J. Zhou, J. Lin, H. Sims, C. Jiang, C. Cong, J. A. Brehm,
937
+ Z. Zhang, L. Niu, Y. Chen, Y. Zhou, Y. Wang, F. Liu,
938
+ C. Zhu, T Yu, K. Suenaga, R. Mishra, S. T. Pantelides,
939
+ Z.-G. Zhu, W. Gao, Z. Liu, and W. Zhou, Adv. Mater.
940
+ 2020, 32, 1906536.
941
+ 43 C. Zhao, T. Norden, P. Zhang, P. Zhao, Y. Cheng,
942
+ F. Sun, J. P. Parry, P. Taheri, J. Wang, Y. Yang,
943
+ T. Scrace, K. Kang, S. Yang, G.-x. Miao, R. Sabirianov,
944
+ G. Kioseoglou, W. Huang, A. Petrou, and H. Zeng, Nat.
945
+ Nanotechnol. 2017, 12, 757.
946
+ 44 T. Norden, C. Zhao, P. Zhang, R. Sabirianov, A. Petrou,
947
+ and H. Zeng, Nat. Commun. 2019, 10, 4163.
948
+ 45 K. F. Mak, K. He, J. Shan, and T. F. Heinz, Nat. Nan-
949
+ otechnol., 2012, 7, 494-498.
950
+ 46 H. Zeng, J. Dai, W. Yao, D. Xiao, and X. Cui, Nat. Nan-
951
+ otechnol., 2012, 7, 490-493.
952
+ 47 K. F. Mak, D. Xiao, and J. Shan, Nat. Photonics 2018,
953
+ 12, 451.
954
+ 48 A. Kumar, D. Yagodkin, N. Stetzuhn, S. Kovalchuk,
955
+ A. Melnikov, P. Elliott, S. Sharma, C. Gahl, and K. I.
956
+ Bolotin, Nano Lett. 2021, 21, 7123.
957
+ 49 S. Guddala, Y. Kawaguchi, F. Komissarenko, S. Kir-
958
+ iushechkina, A. Vakulenko, K. Chen, A. Al`u, V. M. Menon,
959
+ and A. B. Khanikaev, Nat. Commun. 2021, 12, 3746.
960
+ 50 A. Pulkin and O. V. Yazyev, Phys. Rev. B, 2016, 93,
961
+ 041419(R).
962
+ 51 X.-T. An, J. Xiao, M. W.-Y. Tu, H. Yu, V. I. Fal’ko, and
963
+ W. Yao, Phys. Rev. Lett., 2017, 118, 096602.
964
+ 52 V. I. Talyanskii, D. S. Novikov, B. D. Simons, and L. S.
965
+ Levitov, Phys. Rev. Lett. 2001, 87, 276802.
966
+ 53 Q. Zhang, K. S. Chan, and Z. Lin, Appl. Phys. Lett., 2011,
967
+ 98, 032106.
968
+ 54 M.-J. Wang, J. Wang, and J.-F. Liu, New J. Phys., 2020
969
+ 22, 013042.
970
+ 55 K.-T. Wang, F. Xu, B. Wang, Y. Yu, and Y. Wei, Front.
971
+ Phys., 2022, 17, 43501.
972
+ 56 Y. Xing, B. Wang, Y. Wei, B. Wang, and J. Wang, Phys.
973
+ Rev. B, 2004, 70, 245324.
974
+ 57 W. Y. Deng, W. Luo, H. Geng, M. N. Chen, L. Sheng, and
975
+ D. Y. Xing, New J. Phys., 2015, 17, 103018.
976
+ 58 M. N. Chen, L. Sheng, R. Shen, D. N. Sheng, and D. Y.
977
+ Xing, Phys. Rev. B, 2015 91, 125117.
978
+ 59 H. Ochoa and R. Rold´an, Phys. Rev. B, 2013, 87, 245421.
979
+ 60 J. Klinovaja and D. Loss, Phys. Rev. B, 2013, 88, 075404.
980
+ 61 I. Kleftogiannis, C.-S. Tang, and S.-J. Cheng, J. Phys.:
981
+ Condens. Matter, 2015, 27, 205302.
982
+ 62 C. L. Kane and E. J. Mele, Phys. Rev. Lett., 2005, 95,
983
+ 226801.
984
+ 63 C. Bai, Y. Yang, and L. Bai, J. Phys.: Condens. Matter,
985
+ 2018, 30, 495801.
986
+ 64 P. W. Brouwer, Phys. Rev. B, 1998, 58, R10135.
987
+ 65 M. B¨uttiker, H. Thomas, and A. Prˆetre, Z. Phys. B Con-
988
+ dens. Matter, 1994, 94, 133.
989
+ 66 Y. Wei, J. Wang, and H. Guo, Phys. Rev. B, 2000, 62,
990
+ 9947.
991
+ 67 B. Wang, J. Wang, and H. Guo, Phys. Rev. B, 2003 68,
992
+ 155326.
993
+ 68 F. Xu, Y. Xing, and J. Wang, Phys. Rev. B, 2011, 84,
994
+ 245323.
995
+ 69 Y. Yu, Y. Zhou, L. Wan, B. Wang, F. Xu, Y. Wei, and
996
+
997
+ 8
998
+ J. Wang, Nanotechnology, 2016, 27, 185202.
999
+ 70 D. H. Lee and J. D. Joannopoulos, Phys. Rev. B, 1981, 23,
1000
+ 4988.
1001
+ 71 D. H. Lee and J. D. Joannopoulos, Phys. Rev. B, 1981, 23,
1002
+ 4997.
1003
+ 72 J.-S. Yang, X.-G. He, S.-H. Chen, and C.-R. Chang, Phys.
1004
+ Rev. B, 2008, 78, 085312.
1005
+ 73 D. Culcer, C. Lechner, and R. Winkler, Phys. Rev. Lett.,
1006
+ 2006, 97, 106601.
1007
+ 74 T.-W. Chen and H.-C. Hsu, New J. Phys., 2021, 23,
1008
+ 073017.
1009
+ 75 N. D. Lang and P. Avouris, Phys. Rev. Lett., 1998, 81,
1010
+ 3515.
1011
+ 76 N. D. Lang and P. Avouris, Phys. Rev. Lett., 2000, 84,
1012
+ 358.
1013
+
NNFJT4oBgHgl3EQfzi34/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
OdAzT4oBgHgl3EQfzf5C/content/tmp_files/2301.01769v1.pdf.txt ADDED
@@ -0,0 +1,1348 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Comprehensive analysis of gene expression profiles to radiation exposure reveals
2
+ molecular signatures of low-dose radiation response
3
+ Xihaier Luo∗, Sean McCorkle∗, Gilchan Park∗, Vanessa L´opez-Marrero∗, Shinjae Yoo∗,
4
+ Edward R. Dougherty†, Xiaoning Qian∗†, Francis J. Alexander∗, Byung-Jun Yoon∗†
5
+ ∗ Computational Science Initiative, Brookhaven National Laboratory, Upton, NY
6
+ † Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX
7
+ Abstract—There are various sources of ionizing radiation ex-
8
+ posure, where medical exposure for radiation therapy or diag-
9
+ nosis is the most common human-made source. Understanding
10
+ how gene expression is modulated after ionizing radiation
11
+ exposure and investigating the presence of any dose-dependent
12
+ gene expression patterns have broad implications for health
13
+ risks from radiotherapy, medical radiation diagnostic proce-
14
+ dures, as well as other environmental exposure. In this paper,
15
+ we perform a comprehensive pathway-based analysis of gene
16
+ expression profiles in response to low-dose radiation exposure,
17
+ in order to examine the potential mechanism of gene regulation
18
+ underlying such responses. To accomplish this goal, we employ
19
+ a statistical framework to determine whether a specific group
20
+ of genes belonging to a known pathway display coordinated
21
+ expression patterns that are modulated in a manner consistent
22
+ with the radiation level. Findings in our study suggest that
23
+ there exist complex yet consistent signatures that reflect the
24
+ molecular response to radiation exposure, which differ between
25
+ low-dose and high-dose radiation.
26
+ Index Terms—Gene expression analysis, radiation biology, low-
27
+ dose radiation response, pathway analysis.
28
+ 1. Introduction
29
+ Environmental threats constitute a major factor in deter-
30
+ mining a person’s susceptibility to disease. With the progress
31
+ of industrialization and modernization, radiation exposure
32
+ has become one of the most serious environmental threats
33
+ in today’s world. Mounting evidence suggests that ionizing
34
+ radiation is linked to the development of thyroid cancers,
35
+ multiple myeloma, and myeloid leukemia in children and
36
+ adults [1]. It is well documented that the biological effects of
37
+ ionizing radiation on mammalian cells are closely related to
38
+ radiation doses and dose rates. In general, low-dose radiation
39
+ exposure is far more common than high-dose radiation ex-
40
+ posure because low-dose radiation can come from a variety
41
+ of sources, including natural sources, cosmic rays, nuclear
42
+ power, and various types of radioactive waste. However,
43
+ in contrast to the more well-defined effects of high-dose
44
+ radiation exposure, the biological effects and consequences
45
+ of low-dose radiation and mixed exposures remain poorly
46
+ understood [2], [3].
47
+ Historically, the health risks associated with low-dose
48
+ ionizing radiation exposure have been estimated by extrap-
49
+ olating from available high-dose radiation exposure data.
50
+ However, the majority of the data come from experiments
51
+ that used extremely high, even supra-lethal, doses. Extrapo-
52
+ lating the results of such studies to physiologically relevant
53
+ doses can thus be difficult [4]. Furthermore, an increasing
54
+ number of studies show that the biological reactions to
55
+ high and low doses of radiation are qualitatively distinct,
56
+ necessitating a direct examination of low-dose responses to
57
+ better understand potential risks [5].
58
+ Genome-wide expression assays using microarrays or
59
+ RNA sequencing can provide snapshots of transcriptional
60
+ activities in a biological sample, hence studying the gene
61
+ expression profiles under low doses of ionizing radiation
62
+ can provide novel insights into the biological reactions to
63
+ such radiation exposure. In fact, mining gene expression
64
+ profiles has proven useful in understanding pathophysiolog-
65
+ ical mechanisms, diagnosis and prognosis of complex dis-
66
+ eases, and deciding on treatment plans. Several studies have
67
+ demonstrated the effectiveness of using gene expression
68
+ profiles for traditionally challenging problems, for instance,
69
+ discriminating between different subtypes of a complex
70
+ disease, such as cancer [6], [7]. Despite these successful
71
+ applications, quantification and interpretation at the genetic
72
+ level of the impact from radiation exposure on the risk of
73
+ developing such diseases are still challenging. Especially,
74
+ the small sample size of typical clinical data, on the other
75
+ hand, frequently impedes meaningful analysis, making pat-
76
+ tern discovery, disease marker identification, risk prediction,
77
+ reproducibility, and validation extremely difficult [8], [9].
78
+ Adjusting for multiple hypothesis testing is another critical
79
+ issue for all microarray analysis methods. The similarities
80
+ of such signatures across different sample types have not
81
+ been demonstrated to be strong enough to conclude that
82
+ they represent a universal biological mechanism shared by
83
+ different sample types [10]–[12].
84
+ In recent years, scientists have gained a better under-
85
+ standing of the transcriptional response in cells to radiation
86
+ exposure [13]. When cells are exposed to ionizing radiation,
87
+ multiple signal transduction pathways are activated, mak-
88
+ ing pathway activity a potentially powerful and informa-
89
+ tive approach for determining disease states. Furthermore,
90
+ pathways, the most well-documented protein interactions,
91
+ arXiv:2301.01769v1 [q-bio.GN] 3 Jan 2023
92
+
93
+ are known to closely reflect functional relationships related
94
+ to molecular biological activities such as metabolic, sig-
95
+ naling, protein interaction, and gene regulation processes.
96
+ A growing body of research indicates that tasks such as
97
+ class distinction based on differences in pathway activity
98
+ can be more stable than distinction based solely on genes.
99
+ For example, [14] incorporated pathway information into
100
+ expression-based disease diagnosis and proposed a classifi-
101
+ cation method based on pathway activities inferred for each
102
+ patient. Later in [15], pathway activity patterns are used to
103
+ describe a classification scheme for human breast cancer
104
+ and to reveal complexity in intrinsic breast cancer subtypes.
105
+ The probabilistic inference of differential pathway activity
106
+ across different classes (e.g., disease states or phenotypes)
107
+ using probabilistic graphical models [16] was shown to
108
+ identify molecular signatures that can be used as robust
109
+ and reproducible disease markers. The marker identification
110
+ method in [16] was further extended in [17], where a novel
111
+ algorithm for discovering robust and effective subnetwork
112
+ markers in a human protein-protein interaction network that
113
+ can accurately predict cancer prognosis and simultaneously
114
+ discover multiple synergistic subnetwork markers. It should
115
+ be noted that at the heart of these pathway-based analyses
116
+ is determining the activity of a given pathway based on the
117
+ expression levels of the constituent genes.
118
+ The primary goal of this paper is to perform a compre-
119
+ hensive pathway-based analysis of gene expression profiles
120
+ to investigate the differential time and dose effects, primarily
121
+ in low-dose experiments, in order to uncover molecular
122
+ signatures of low-dose radiation response. Towards this goal,
123
+ we adopt the probabilistic pathway activity inference scheme
124
+ in [16], where the pathway activity level is estimated from
125
+ gene expression data via the use of a simple probabilistic
126
+ graphical model. More specifically, the scheme estimates the
127
+ log-likelihood ratio between different classes (e.g., differ-
128
+ ent levels of radiation exposure) based on the expression
129
+ level of each member gene. The log-likelihood ratios of
130
+ the member genes in a given pathway are then aggregated
131
+ for probabilistic inference of differential pathway activity.
132
+ Through this analysis, we identify the most significantly
133
+ differentially activated pathways in response to low-dose
134
+ radiation. These pathways are investigated to determine
135
+ the presence of consistent dose-dependent gene expression
136
+ patterns. Our cross-validation experiments demonstrate that
137
+ the proposed method can generate reliable and consistent
138
+ pathway analysis results even with limited data.
139
+ 2. Data
140
+ 2.1. Low-dose radiation gene expression data
141
+ The goal of the current study is to identify poten-
142
+ tial molecular signatures underlying the biological response
143
+ to low-dose ionizing radiation exposure through pathway-
144
+ based analysis of gene expression profiles. For this purpose,
145
+ we conducted a thorough literature search and preliminary
146
+ analysis to identify human gene expression data suitable
147
+ for studying the low-dose radiation response. The gene
148
+ Dose Level
149
+ Number of Samples
150
+ 0 Gy
151
+ 18
152
+ 0.005 Gy
153
+ 16
154
+ 0.01 Gy
155
+ 18
156
+ 0.025 Gy
157
+ 18
158
+ 0.05 Gy
159
+ 17
160
+ 0.1 Gy
161
+ 18
162
+ 0.5 Gy
163
+ 16
164
+ TABLE 1. DESCRIPTION OF THE GENE EXPRESSION DATASET
165
+ GSE43151 THAT WAS USED TO INVESTIGATE THE MOLECULAR
166
+ SIGNATURES OF LOW-DOSE RADIATION RESPONSE IN THIS STUDY.
167
+ expression dataset GSE431511 was identified to be the most
168
+ suitable for our study, in terms of sample size and the range
169
+ of radiation levels that were considered. Overall, GSE43151
170
+ contains gene expression measurements from 121 blood
171
+ samples, where five healthy male donors provided 400 mL
172
+ venous peripheral blood samples each [18]. A complete
173
+ blood count was performed on each whole blood sample
174
+ using an ADVIA Hematology System (Bayer HealthCare).
175
+ The standard lymphocyte proportion of 16-45 percent was
176
+ met by all samples. Heparin at a final concentration of 34
177
+ U ml−1 was added to whole blood samples. The blood
178
+ was then diluted 1:10 with Iscove’s Modified Dulbecco’s
179
+ Medium (IMDM, Life Technologies). Finally, blood samples
180
+ were incubated overnight at 37 Cina 5% CO2 concentration.
181
+ For the ex vivo irradiation, whole blood exposures were
182
+ performed at the ICO-4000 facility (Fontenay-aux-Roses,
183
+ France) with a Co source at a low dose rate (50 mGy
184
+ min−1). Exposures were carried out independently on each
185
+ donor’s blood sample. The kerma rate was calculated us-
186
+ ing a Physikalisch-Technische Werkst¨atten (PTW) ionization
187
+ chamber that was irradiated under the same conditions as the
188
+ samples. Doses of 5, 10, 25, 50, 100, and 500 mGy were
189
+ tested (See Table. 1), as well as sham irradiated conditions.
190
+ Following ex vivo irradiation, blood samples were incubated
191
+ at 37 degrees Celsius for 150, 300, 450, and 600 minutes
192
+ in a 5% CO2 atmosphere.
193
+ A density medium was used to collect CD4+ T lym-
194
+ phocytes for cell sorting. Following that, total RNA was
195
+ extracted from CD4+ T lymphocytes using RNeasy Mini
196
+ columns from the RNeasy Mini Kit (Qiagen) as directed by
197
+ the manufacturer. For all RNA samples, the RIN (RNA in-
198
+ tegrity number) was calculated for assigning integrity values
199
+ to RNA measurements. For gene expression assays, all RIN
200
+ values were greater than the recommended value of 7.
201
+ Before performing the pathway analysis based on the
202
+ GSE43151 gene expression dataset, all 121 samples in the
203
+ dataset were normalized, filtered, and analyzed using GAGE
204
+ in R software [19]. Following the filtering step, a total of
205
+ 10,875 probes were chosen, where the basic filtering criteria
206
+ consisted of removing a probe when it was undetected in at
207
+ least 75% of the replicates considered.
208
+ 1. https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE43151
209
+
210
+ (D.1) Overall algorithm - pseudo code
211
+ (D.2) Rank pathways
212
+ Rank pathways using computed t-scores
213
+ for pathway in KEGG database
214
+ for gene in selected pathway
215
+ compute the active score
216
+ compute the t-score
217
+
218
+ hsa00010
219
+
220
+ hsa00020
221
+
222
+ hsa00030
223
+
224
+ hsa00040
225
+
226
+
227
+
228
+ hsa00400
229
+
230
+ hsa00560
231
+
232
+ hsa04907
233
+
234
+ hsa00790
235
+
236
+
237
+ Task 1: low-dose Task 2: high-dose
238
+ (C.1) Label samples
239
+ (C.2) Build conditional distributions
240
+ (C.3) Estimate activity score
241
+ Task 1: zero-dose vs low-dose Task 2: zero-dose vs high-dose
242
+ . . .
243
+
244
+ Zero-dose
245
+
246
+ Low-dose
247
+
248
+ High-dose
249
+ Compute log-likelihood ratio by Equation 1
250
+ • Task 1
251
+ • Task 2
252
+ <latexit sha1_base64="5FlLgQ/XWm6UEAeV2onZM/k+9I=">ACHicbVDLSsNAFJ3UV62vqEs3g0VoNzXRom6EghsFxXsA9oQJpNJO3SiTMTpYR+iBt/xY0LRdy4EPwbp20W2nrgXg7n3MvMPV7MqFSW9W3kFhaXlfyq4W19Y3NLXN7pyl5IjBpYM64aHtIEkY
253
+ j0lBUMdKOBUGhx0jLG1yM/dY9EZLy6FYNY+KEqBfRgGKktOSax9duyvjDCJ5DxnuwFJRg9y5BftbL8BDOSmXLFoVawI4T+yMFEGumt+dn2Ok5BECjMkZce2YuWkSCiKGRkVuokMcID1CMdTSMUEumk+NG8EArPgy40BUpOF/b6QolHIYenoyRKovZ72x+J/XSVRw5qQ0ihNFIjx9KEgYVByOk4I+FQrNtQEYUH1XyHuI4Gw0nkWdAj27MnzpHlUsU8q1ZtqsXaVxZEHe2AflIANTkENXI6aAMHsEzeAVvxpPxYrwbH9PRnJHt7I/ML5+AJo8npc=</latexit>Llow = log(f(
254
+ )/f(
255
+ ))
256
+ <latexit sha1_base64="7s+WhPIwKFCaBDQ2xfdNP0Dke4=">ACHXicbVDLSsNAFJ3UV62vqEs3g0VoNzWRom6EghsFxXsA9oQJpNJOnQyiTMToYT+iBt/xY0LRVy4Ef/GaZuFth64l8M59zJzj5cwKpVlfRuFpeWV1bXiemljc2t7x9zda8s4FZi0cMxi0fWQJIx
257
+ y0lJUMdJNBEGRx0jHG15O/M4DEZLG/E6NEuJEKOQ0oBgpLblm/cbNBjQcjOEFZHEIK0EF9u9T5Oe9Co/hvFR1zbJVs6aAi8TOSRnkaLrmZ9+PcRoRrjBDUvZsK1FOhoSimJFxqZ9KkiA8RCHpacpRKSTa8bwyOt+DCIhS6u4FT9vZGhSMpR5OnJCKmBnPcm4n9eL1XBuZNRnqSKcDx7KEgZVDGcRAV9KghWbKQJwoLqv0I8QAJhpQMt6RDs+ZMXSfukZp/W6rf1cuM6j6MIDsAhqAbnIEGuAJN0AIYPIJn8ArejCfjxXg3PmajBSPf2Qd/YHz9AEXwnu8=</latexit>Lhigh = log(f(
258
+ )/f(
259
+ ))
260
+ (B.1) GSE Database
261
+ (B.2) Identify the low-dose radiation data set
262
+ (B.3) Sample classification
263
+ Zero radiation samples
264
+ Low-dose radiation samples
265
+ ERR127303
266
+ ERR127302
267
+ ERR127305
268
+ ERR127304
269
+ ERR127309
270
+ ERR127307
271
+ ERR127306
272
+ ERR127308
273
+ 26472
274
+ 2029
275
+ 51582
276
+ 6418
277
+ 51377
278
+ 11146
279
+ 5147
280
+ 11261
281
+ 304
282
+ 10628
283
+ 336
284
+ 308
285
+ 10935
286
+ 5511
287
+ 145376
288
+ 5037
289
+ 57805
290
+ 8525
291
+ 6588
292
+ 341
293
+ 22853
294
+ 27344
295
+ 116154
296
+ 221476
297
+ 142679
298
+ 710
299
+ 7035
300
+ 1026
301
+ 9749
302
+ 27329
303
+ 153218
304
+ 51050
305
+ 5611
306
+ 8434
307
+ 5274
308
+ 25913
309
+ 331
310
+ 2646
311
+ 54577
312
+ 94274
313
+ 5627
314
+ 301
315
+ 161742
316
+ 9479
317
+ 2873
318
+ 1032
319
+ 9491
320
+ 284352
321
+ 5570
322
+ 9858
323
+ 7349
324
+ 23145
325
+ 27290
326
+ 7076
327
+ 1028
328
+ 57761
329
+ 7079
330
+ 124790
331
+ 89932
332
+ −2
333
+ 0
334
+ 2
335
+ Value
336
+ Color Key
337
+ High-dose radiation samples
338
+ (D.3) Expert interpretation
339
+ (A.1) KEGG Pathway Database
340
+ (A.2) Extract the pathway information
341
+ (A.3) Gene list
342
+ gene 1
343
+ gene 2
344
+ gene k
345
+ . . .
346
+ A. Build a gene list from a selected pathway from the KEGG database
347
+ B. Identify low-dose radiation data set from GSE database
348
+ C. Probabilistic inference of pathway activity
349
+ D. Rank pathways based on their discriminative powers
350
+ Figure 1. Overview of the pathway-based analysis of gene expression profiles in response to low-dose radiation exposure.
351
+ 2.2. Pathway database
352
+ We used the KEGG (Kyoto Encyclopedia of Genes
353
+ and Genomes) database to obtain a reliable set of known
354
+ biological pathways [20]. KEGG is a collection of manually
355
+ drawn pathway maps for understanding high-level functions
356
+ and utilities of the biological system. The genomic infor-
357
+ mation is maintained in the GENES database, which is a
358
+ collection of gene catalogs for all fully sequenced genomes
359
+ and some partially sequenced genomes with current annota-
360
+ tions of gene functions. The PATHWAY database’s higher-
361
+ order functional information is augmented with a collection
362
+ of ortholog group tables for information about conserved
363
+ subpathways, which are frequently encoded by positionally
364
+ related genes on the chromosome and are especially valuable
365
+ in predicting gene functions. In our case, we identified 343
366
+ pathways relevant to the gene expression dataset GSE43151
367
+ from the available 548 KEGG pathway maps by discarding
368
+ the pathways that did not contain any gene whose measure-
369
+ ment was included in GSE43151.
370
+ 3. Methods
371
+ In this section, we describe the technical details of the
372
+ pathway-based gene expression data analysis procedure that
373
+ was used to detect potential molecular signatures underlying
374
+ low-dose radiation response. Figure 1 provides an overview
375
+ of the overall procedure.
376
+ 3.1. Pathway activity inference
377
+ To perform the pathway analysis, we first identified the
378
+ genes whose measurements were included in the gene ex-
379
+ pression dataset GSE43151 for the pathways of our interest.
380
+
381
+ P53 SIGNALING PATHWAY
382
+ Target genes
383
+ Cyclin D
384
+ CDK416
385
+ Response
386
+ G1 arest
387
+ p21
388
+ Cyclin E
389
+ (sustaine d)
390
+ -irradlia
391
+ 143-3-
392
+ CDK2
393
+ Cell cyc le arrest
394
+ UV
395
+ /Rerrima
396
+ Genotoxic
397
+ Cyelin E
398
+ ATM
399
+ CHK2
400
+ G2 arrest
401
+ Cell cyc le
402
+ Cellular se rnescence
403
+ drugs
404
+ DNA damage
405
+ Gadd45
406
+ Cdc2
407
+ (sustained)
408
+ Nutrition
409
+ ATR
410
+ CHK1
411
+ B99
412
+ deprivation
413
+ Hypoxia
414
+ Heaticold .
415
+ Fas
416
+ shock
417
+ Nitric oxide
418
+ DRS
419
+ CASP8
420
+ PIDD
421
+ Bil
422
+ A poptosis
423
+ Stress signak
424
+ Noxa
425
+ PUIMAP53AIP
426
+ tBid
427
+ Cytc
428
+ Jncogene
429
+ Bcl-xL
430
+ Ras, BCR-ABL)
431
+ 7Sival
432
+ -CASP9
433
+ CASP3
434
+ SCYL1EPI
435
+ Bc12
436
+ Araf-1
437
+ +p
438
+ ROS
439
+ PIGs
440
+ P14ARF
441
+ MDM2
442
+ r53
443
+ DNA
444
+ IVitoc hordrior
445
+ ScotinPERPPAG608Siah
446
+ Apoptosis
447
+ AAIFM2
448
+ MDMX
449
+ IGF-BF3
450
+ HIGF
451
+ Cell cy le
452
+ PAIBAI-1KAIGDAiFTSP1Maspin
453
+ Irhibition ofangiogene sis
454
+ ar retastasis
455
+ P48p53R2Gadd45Sestins
456
+ DNA re pair and
457
+ PTENTSC2IGF-BP3
458
+ TS AF6
459
+ Exosorre rrediated
460
+ secretion
461
+ MDM2Cop-1PIRH2CyelinGSiah-1WiplANp73
462
+ p53 regative feedback
463
+ 041156/4/20
464
+ (c) Kanehisa LaboratoriesFor every pathway, member genes that were missing in the
465
+ given dataset were removed from the gene set. Consider a
466
+ pathway G that consist of n genes {gk}n
467
+ k=1 whose mea-
468
+ surements were available in the dataset. In the context of
469
+ binary classification, we assume that the expression level of
470
+ gene gk (k = 1, 2, . . . , n) has a phenotype-dependent dis-
471
+ tribution. Let us denote the conditional probability density
472
+ function (PDF) of gene gk expression level under phenotype
473
+ 1 as f 1
474
+ k(x) and the conditional PDF under phenotype 2
475
+ as f 2
476
+ k(x) with x representing the expression level of gene
477
+ gk. In our case, we classify radiation exposures into three
478
+ categories: zero-dose, low-dose, and high-dose. We compare
479
+ low-dose and high-dose samples separately to zero-dose
480
+ samples, which means that if zero-dose samples are treated
481
+ as phenotype 1, either low-dose or high-dose samples will
482
+ be treated as phenotype 2.
483
+ After examining different probability distribution mod-
484
+ els, we assumed that both f 1
485
+ k(x) and f 2
486
+ k(x) are Guassian in
487
+ this study. Having these conditional PDFs, we can calculate
488
+ the log-likelihood ratio (LLR) between the two phenotypes
489
+ at a given expression level x of gene gk as follows
490
+ Lk(x) = log[f 1
491
+ k(x)/f 2
492
+ k(x)]
493
+ (1)
494
+ For any given gene gk in the pathway G, the associated log-
495
+ likelihood ratio Lk(x) in (1) indicates which phenotype is
496
+ more likely based on the expression level x of gene gk. By
497
+ combining the evidence–in the form of LLR–from all the
498
+ member genes in the pathway, we can assess the overall
499
+ activity level of the pathway at hand to infer which of the
500
+ two phenotypes the collective expression pattern of its mem-
501
+ ber genes points to and how significantly so, as discussed
502
+ in [16]. More specifically, provided with a set {xj,k}m
503
+ j=1 of
504
+ m samples (i.e., gene expression measurements) for each
505
+ gene gk, we first calculated activity levels {Sj}m
506
+ j=1 defined
507
+ as
508
+ Sj =
509
+ n
510
+
511
+ k=1
512
+ Lk(xj,k)
513
+ (2)
514
+ The activity level Sj in (2) incorporates information from
515
+ every gene in the pathway of interest and can be used to
516
+ predict the phenotype (class label) based on the overall
517
+ activation level of the given pathway in sample j.
518
+ Note that to calculate the log-likelihood ratio Lk(x) in
519
+ (1), we must first estimate the conditional PDF f c
520
+ k(x) for
521
+ each phenotype c ∈ {1, 2}. We assume that the expression
522
+ of gene gk under the phenotype c follows a Gaussian dis-
523
+ tribution with a mean of µc
524
+ k and a standard deviation of σc
525
+ k.
526
+ These parameters were calculated using all of the available
527
+ samples that correspond to the phenotype c. After that, the
528
+ estimated conditional PDFs can be utilized to compute the
529
+ log-likelihood ratios. In practice, we often have insufficient
530
+ training data to estimate the PDFs of f 1
531
+ k(x) and f 2
532
+ k(x)
533
+ with confidence. As a result, the computation of the log-
534
+ likelihood ratio may be sensitive to relatively small changes
535
+ in the gene expression levels. To alleviate this issue, we
536
+ normalized the data as recommended in [16]. Namely, Lk(x)
537
+ was normalized to obtain �Lk(x) as follows
538
+ �Lk(x) =
539
+ Lk(x) − E[Lk(x)]
540
+
541
+ E[(Lk(x) − E[Lk(x)])2]
542
+ .
543
+ (3)
544
+ While the use of (1) and (2) without normalization for infer-
545
+ ring the pathway activity level would be equivalent to using
546
+ a Naive Bayes model (NBM) for classifying the phenotype
547
+ (class label) given the expression profile of the member
548
+ genes that belong to a given pathway, this normalization step
549
+ in (3) makes the pathway activity scoring scheme diverge
550
+ from the traditional NBM.
551
+ 3.2. Pathways as potential markers for discriminat-
552
+ ing low-dose response from high-dose response
553
+ To examine the ability of a pathway to discriminate
554
+ between two phenotypes, we computed the t-test statistics
555
+ scores using the activity levels Sj for all member genes (as
556
+ defined in (2)) and averaged the absolute value of the t-test
557
+ scores to compute an aggregated differential activity score.
558
+ The aggregated score–which we refer to as the pathway
559
+ activity score–was then used as an indicator of the pathway’s
560
+ discriminative power [21]. It should be noted that low-dose
561
+ and high-dose samples were analyzed separately to detect
562
+ most strongly differentially activated pathways under each
563
+ radiation exposure level. We had three types of samples:
564
+ zero radiation, low-dose radiation (0.005 Gy to 0.1 Gy),
565
+ and high-dose radiation (0.5 Gy). Despite the fact that
566
+ different low-dose levels of ionizing radiation have been
567
+ tested, we treated all dose levels between 0.005 Gy and 0.1
568
+ Gy as the same type (i.e., low-dose radiation). Based on this
569
+ categorization, we ranked all relevant KEGG pathways to
570
+ based on the strongest differential pathway activity between
571
+ zero-dose against low-dose radiations, and separately, based
572
+ on zero-dose against high-dose radiations. This is illustrated
573
+ in Fig. 1(C).
574
+ 4. Results
575
+ 4.1. Pathway analysis results
576
+ To begin, we evaluated all relevant pathways in the
577
+ KEGG database and ranked the pathways based on their
578
+ discriminative power following the procedures elaborated
579
+ in Sec. 3 and illustrated in Fig. 1. In particular, we ranked
580
+ the pathways based on their discriminative power, assessed
581
+ based on the aggregated differential activity score obtained
582
+ by averaging the absolute value of the t-test scores of the
583
+ member genes in a given pathway [21] and estimating the
584
+ p-value.
585
+ Fig. 2(a) shows the top five pathways that have been
586
+ identified as being the most deferentially activated in the
587
+ presence of low-dose radiation.
588
+ The top pathway was associated with Natural killer cell
589
+ mediated cytotoxicity, focusing on natural killer cells, which
590
+ are innate immune system lymphocytes involved in early
591
+
592
+ <latexit sha1
593
+ _base64="2RwxLXlY8TROIoM9
594
+ 8j2WcOjpro=">AB6nicbVDL
595
+ SgNBEOyNrxhfUY9eBoMQL2FXgn
596
+ oMevEY0TwgWcLspDcZMju7zMw
597
+ KIeQTvHhQxKtf5M2/cZLsQRML
598
+ Goqbrq7gkRwbVz328mtrW9sbu
599
+ W3Czu7e/sHxcOjpo5TxbDBYhG
600
+ rdkA1Ci6xYbgR2E4U0igQ2ApGt
601
+ zO/9YRK81g+mnGCfkQHkoecUW
602
+ OlhzI97xVLbsWdg6wSLyMlyFDv
603
+ Fb+6/ZilEUrDBNW647mJ8SdUG
604
+ c4ETgvdVGNC2YgOsGOpBFqfz
605
+ I/dUrOrNInYaxsSUPm6u+JCY20
606
+ HkeB7YyoGeplbyb+53VSE17E
607
+ y6T1KBki0VhKoiJyexv0ucKmRF
608
+ jSyhT3N5K2JAqyoxNp2BD8JZf
609
+ XiXNi4p3WaneV0u1myOPJzAKZ
610
+ TBgyuowR3UoQEMBvAMr/DmCOf
611
+ FeXc+Fq05J5s5hj9wPn8Ai5mN
612
+ Uw=</latexit>(a)
613
+ <latexit sha1
614
+ _base64="KWh0RLJ0bw8em/x3P
615
+ U2+HIlN2FQ=">AB6nicbVDL
616
+ SgNBEOyNrxhfUY9eBoMQL2FXgn
617
+ oMevEY0TwgWcLspDcZMju7zMw
618
+ KIeQTvHhQxKtf5M2/cZLsQRML
619
+ Goqbrq7gkRwbVz328mtrW9sbu
620
+ W3Czu7e/sHxcOjpo5TxbDBYhG
621
+ rdkA1Ci6xYbgR2E4U0igQ2ApGt
622
+ zO/9YRK81g+mnGCfkQHkoecUW
623
+ Olh3Jw3iuW3Io7B1klXkZKkKHe
624
+ K351+zFLI5SGCap1x3MT40+oM
625
+ pwJnBa6qcaEshEdYMdSPU/m
626
+ R+6pScWaVPwljZkobM1d8TExp
627
+ PY4C2xlRM9TL3kz8z+ukJrz2J
628
+ 1wmqUHJFovCVBATk9nfpM8VMiP
629
+ GlCmuL2VsCFVlBmbTsG4C2/
630
+ vEqaFxXvslK9r5ZqN1kceTiBUy
631
+ iDB1dQgzuoQwMYDOAZXuHNEc6
632
+ L8+58LFpzTjZzDH/gfP4AjR6N
633
+ VA=</latexit>(b)
634
+ Figure 2. Ranking of most differentially activated pathways and their
635
+ discriminative power in terms of the pathway activity score. (a) Top
636
+ differentially activated pathways under low-dose radiation exposure. The
637
+ aggregated t-test scores reflect the discriminative power of the pathways
638
+ for discriminating between zero-dose and low-dose samples. (b) Top differ-
639
+ entially activated pathways for high-dose radiation exposure (zero-dose vs
640
+ high-dose). Comparison between (a) and (b) show a significant difference
641
+ between the list of top pathways that are differentially activated under low-
642
+ dose radiation and those under high-dose radiation.
643
+ defenses against both allogeneic and autologous cells. Many
644
+ studies have been conducted to investigate the direct effects
645
+ of low-dose ionizing radiation (LDIR) on natural killer cells
646
+ and the potential mechanism [22], [23]. The results of the
647
+ experiments showed that a simplified strategy based on
648
+ LDIR leads to effective expansion and increased activity of
649
+ natural killer cells, providing a novel approach for adoptive
650
+ cellular immunotherapy.
651
+ The second pathway is related to Adherens junction (AJ),
652
+ which is the most common type of intercellular adhesion. AJ
653
+ initiates and maintains cell adhesion while also controlling
654
+ the actin cytoskeleton. In [24], three types of junctional
655
+ proteins were chosen for immunohistochemical labeling,
656
+ and experimental results showed that not only high, but
657
+ also low and moderate doses of cranial irradiation increase
658
+ cerebral vessel permeability in mice. In-vitro studies showed
659
+ that irradiation alters junctional morphology, reduces cell
660
+ number, and causes senescence in brain endothelial cells.
661
+ Another study [25] discovered that gamma-radiation, even
662
+ at low doses, rapidly disrupts tight junctions, adherens
663
+ junctions, and the actin cytoskeleton, resulting in barrier
664
+ dysfunction in the mouse colon in vivo. Radiation-induced
665
+ epithelial junction disruption and barrier dysfunction are
666
+ mediated by oxidative stress, which can be mitigated by
667
+ NAC supplementation prior to IR.
668
+ Another pathway linked to Sphingolipid metabolism was
669
+ also highly ranked. Sphingolipids, a type of membrane
670
+ lipid, are bioactive molecules that play a variety of roles
671
+ in fundamental cellular processes such as cell division,
672
+ differentiation, and cell death. Many studies on the effect
673
+ of sphingolipids on cancer treatment have been conducted.
674
+ Microbeam radiation can induce radiosensitivity in elements
675
+ within the cytoplasm, according to [26]. The effect could be
676
+ inhibited by agents that disrupt the formation of lipid rafts
677
+ (filipin), demonstrating once again that membranes could be
678
+ a target of ionizing radiation. The authors of [27] concluded
679
+ that, while other pathways are activated to induce radiation
680
+ or chemoresistance, sphingolipids play a significant role.
681
+ The JAK-STAT signaling pathway and Glycosphin-
682
+ golipid biosynthesis have also been revealed to be very
683
+ important in the study of radiation effects. For example,
684
+ erythropoietin (EPO), which was originally identified as an
685
+ erythrocyte growth factor, is now used to treat anemia and
686
+ fatigue in cancer patients receiving radiation therapy and
687
+ chemotherapy. The study in [28] demonstrated previously
688
+ unknown EPO-mediated HNSCC cell invasion via the Janus
689
+ kinase (JAK)-signal transducer and activator of transcription
690
+ (STAT) signaling pathway. On the other hand, the findings in
691
+ [29] suggest that glycosphingolipid biosynthesis on the cell
692
+ surface contributed to the activation of ionizing radiation-
693
+ induced apoptosis via ceramide production. The functional
694
+ importance of this pathway to eradicating cancer cells with
695
+ ionizing radiation has been proven, with sphingolipid break-
696
+ down activated as a mechanism of ceramide formation after
697
+ cell irradiation.
698
+ In a similar manner, Fig. 2(b) shows the top five path-
699
+ ways that have been identified as being most differentially
700
+ activated in the presence of high-dose radiation. The genes
701
+ found in the identified pathways are closely related to the
702
+ radiotherapy regimen. Graft-versus-host disease (GVHD),
703
+ for example, is a fatal complication of allogeneic hematopoi-
704
+ etic stem cell transplantation in which immunocompetent
705
+ donor T cells attack genetically diverse host cells. Many
706
+ clinical studies have found a link between GVHD severity
707
+ and radiation dose, with more severe GVHD after condi-
708
+ tioning regimens that included radiation therapy compared
709
+ to those that only included chemotherapy [30], [31]. Another
710
+ example is allograft rejection. By definition, the recipient’s
711
+ alloimmune response to nonself antigens expressed by donor
712
+ tissues causes allograft rejection. According to research,
713
+ the complex pathophysiology involves host tissue damage
714
+ caused by the conditioning regimen (chemotherapy and/or
715
+ irradiation) [32]. After nonmyeloablative conditioning with
716
+ low-dose irradiation, the use of recombinant fusion protein
717
+ promotes mixed lymphoid chimerism.
718
+ Interestingly, we can see that there is relatively small
719
+ overlap between the set of pathways there were most re-
720
+ sponsive to low-dose radiation exposure and those that were
721
+ responsive to high-dose radiation exposure. For example, as
722
+ shown in Fig. 2, only one pathway (i.e., Natural killer cell
723
+ mediated cytotoxicity) was among the top 5 differentially
724
+ activate pathways under both low-dose and high-dose ra-
725
+ diation. However, we can see more pathways in common
726
+ as we go down the list further. For example, when we
727
+ compare the top ten pathways that are the most responsive
728
+
729
+ Natural killer cell mediated cytotoxicity
730
+ Pathway Name
731
+ Adherens junction
732
+ Sphingolipid metabolism
733
+ JAK-STAT signaling pathway
734
+ Glycosphingolipid biosynthesis
735
+ 2.5
736
+ 12.5
737
+ 0.0
738
+ 5.0
739
+ 7.5
740
+ 10.0
741
+ Aggregated t-test scoreNatural killer cell mediated cytotoxicity
742
+ Graft-versus-host disease
743
+ Viral myocarditis
744
+ Allograft rejection
745
+ Autoimmune thyroid disease
746
+ 2.5
747
+ 12.5
748
+ 0.0
749
+ 5.0
750
+ 7.5
751
+ 10.0
752
+ Aggregated t-test scoreto low-dose and high-dose radiation exposure, we find four
753
+ common pathways: Natural killer cell mediated cytotoxic-
754
+ ity, Adherens junction, Glycosphingolipid biosynthesis, and
755
+ Antigen processing and presentation.
756
+ 4.2. Differential dose effect on radiation responsive
757
+ pathways
758
+ Next, we investigated the differential dose effects on the
759
+ top pathways that were most responsive to either low-dose
760
+ or high-dose radiation exposure. As noted earlier in Sec. 3.1,
761
+ the probabilistic pathway activity inference scheme [16],
762
+ which we adopted in this current study, is equivalent to using
763
+ a simple probabilistic graphical model (PGM)–namely, a
764
+ NBM–when we use (2) for calculating the pathway activity
765
+ score based on the LLRs of the member genes belonging
766
+ to the pathway. We wanted to find out whether this PGM
767
+ constructed to detect the presence of low-dose (or high-dose)
768
+ radiation exposure yields consistent activity inference results
769
+ as the radiation dose level changes.
770
+ Figure 3 shows the inference result based on the PGM
771
+ trained to discriminate between zero-dose and low-dose
772
+ samples. The y-axis shows the aggregated LLRs and the
773
+ x-axis corresponds to the radiation dose level. For each
774
+ dose level, the dots show the distribution of the pathway
775
+ activity scores for all samples radiated at the given dose
776
+ level. The results are shown for the top five pathways that
777
+ were found to be most responsive to low-dose radiation.
778
+ As we can see in Fig. 3, all low-dose responsive pathways
779
+ yielded similar trends, where the inferred differential activity
780
+ levels generally decreased as the radiation exposure level
781
+ increased. These results imply that these pathways, and the
782
+ gene expression profiles of the members therein, may reflect
783
+ potential molecular signatures underlying the biological re-
784
+ sponse to low-dose radiation exposure.
785
+ We carried out a similar analysis based on the top five
786
+ high-dose radiation response pathways that were identified
787
+ in our study. The analysis results are summarized in Fig. 4.
788
+ As before, the y-axis shows the pathway activity score
789
+ obtained by aggregating the LLRs of the member genes in
790
+ the pathway at hand. It should however be noted that, in this
791
+ case, the LLR is obtained by comparing the likelihood ratios
792
+ between zero-dose response and high-dose response. The
793
+ resulting PGM is therefore trained to discriminate between
794
+ zero-dose samples and high-dose samples. Interestingly, ex-
795
+ cept for the first pathway (i.e., Natural killer cell mediated
796
+ cytotoxicity), which was the top-ranked pathway in both
797
+ low-dose as well as high-dose differential activity analysis
798
+ (see Fig. 2), the pathway activity levels did not change
799
+ significantly as the dose level increased. Considering that
800
+ the pathway activity scores reflect the presence of potential
801
+ molecular signatures of high-dose radiation response, this
802
+ may imply that these top pathways that were responsive
803
+ to high-dose radiation exposure might not be substantially
804
+ perturbed when the radiation dose level is relatively low.
805
+ 4.3. Reproducibility of the identified pathways
806
+ We conducted cross-validation experiments to assess the
807
+ reproducibility of pathway analysis results and the signifi-
808
+ cance of the identified pathways. To begin the experiment,
809
+ we randomly selected 70% of zero-dose, low-dose, and
810
+ high-dose samples, and we repeated this process ten times,
811
+ taking into account the total size of our dataset. The top-
812
+ ranked pathways identified by the algorithm are depicted
813
+ in Fig. 5. Because the different sample selection introduces
814
+ randomness, we first counted the show-up cases of pathways
815
+ from the top ten most activated pathways. Then, we ranked
816
+ our cross-validation results based on the total number of
817
+ counts (shown in blue color). We also computed the mean
818
+ and standard deviation of the aggregated t-test scores for
819
+ each pathway (shown in red color). The cross-validation
820
+ experiments for low-dose radiation responsive pathways are
821
+ shown in see Fig. 5(a). As we can see, Fig. 5(a) demonstrates
822
+ the consistency of the identified pathways when compared
823
+ to the results originally obtained using the whole dataset
824
+ (see Fig. 2 for comparison). Pathways Natural killer cell
825
+ mediated cytotoxicity and JAK-STAT signaling pathway, for
826
+ example, have been identified as being highly related to low-
827
+ dose radiation response. We suspect that the difference is
828
+ due to the radiation dose level. As previously discussed,
829
+ we discovered a direct relationship between dose level and
830
+ activation. Such differences are expected in a mixed and
831
+ random combination of different dose levels.
832
+ Noticeably, such consistency was not observed in the
833
+ high-dose experiments shown in Fig. 5(b). In many top-
834
+ ranked pathways, as shown in Fig. 4, there is a weak dis-
835
+ tinction between high-dose samples. The last column, which
836
+ represents the distribution of the calculated aggregated t-test
837
+ scores of high-dose samples, in particular, shows a narrow-
838
+ band distribution (See Fig. 4(b), (d), and (e)). Because the
839
+ calculated statistical scores are so close, when randomness
840
+ is introduced into data sampling, the cross-validation results
841
+ in Fig. 5(b) appear more random. To validate this, we
842
+ expanded our ranked pathway list to the top 30 pathways
843
+ and found a larger number of overlapping pathways between
844
+ the experiments using full dataset and the cross-validation
845
+ experiments using only 70% of the dataset. In this case, the
846
+ average ranking of the pathways Natural killer cell mediated
847
+ cytotoxicity and Allograft rejection, for example, were 17th
848
+ and 22nd, respectively. It should be noted that the radiation
849
+ dose level that we categorized as “high-dose” in this study is
850
+ still relatively low. We expect that gene expression analysis
851
+ of samples that underwent higher-dose radiation exposure
852
+ may result in more consistent pathway identification results
853
+ with clear molecular signatures.
854
+ Finally, we also investigated the assumption regarding
855
+ the conditional distribution of the gene expression values.
856
+ We used the one-sample Kolmogorov-Smirnov (KS) test to
857
+ determine the goodness of fit. The test compares the under-
858
+ lying distribution F(x) of a sample to a given distribution
859
+ G(x), which in our case is a Gaussian distribution. The
860
+ null hypothesis holds that the two distributions are identical,
861
+ with F(x) = G(x) for all x; the alternative holds that
862
+
863
+ <latexit sha1_base64="2R
864
+ RwxLXlY8TROIoM98j2WcOjpro=">AB6nicbVDLSgNBEOyN
865
+ rxhfUY9eBoMQL2FXgnoMevEY0TwgWcLspDcZMju7zMwKIeQT
866
+ vHhQxKtf5M2/cZLsQRMLGoqbrq7gkRwbVz328mtrW9sbuW3
867
+ Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApG
868
+ tzO/9YRK81g+mnGCfkQHkoecUWOlhzI97xVLbsWdg6wSLyMl
869
+ yFDvFb+6/ZilEUrDBNW647mJ8SdUGc4ETgvdVGNC2YgOsGOp
870
+ pBFqfzI/dUrOrNInYaxsSUPm6u+JCY20HkeB7YyoGeplbyb+
871
+ 53VSE17Ey6T1KBki0VhKoiJyexv0ucKmRFjSyhT3N5K2JAq
872
+ yoxNp2BD8JZfXiXNi4p3WaneV0u1myOPJzAKZTBgyuowR3U
873
+ oQEMBvAMr/DmCOfFeXc+Fq05J5s5hj9wPn8Ai5mNUw=</la
874
+ texit>(a)
875
+ <latexit sha1_base64="KWh0RLJ0bw8em/x3PU2+HIlN2F
876
+ Q=">AB6nicbVDLSgNBEOyNrxhfUY9eBoMQL2FXgnoMevEY0TwgWcLspDcZMju7zMwKIeQTvHhQxKtf5M2/cZLsQRMLGoq
877
+ brq7gkRwbVz328mtrW9sbuW3Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApGtzO/9YRK81g+mnGCfkQHkoec
878
+ UWOlh3Jw3iuW3Io7B1klXkZKkKHeK351+zFLI5SGCap1x3MT40+oMpwJnBa6qcaEshEdYMdSPU/mR+6pScWaVPwljZkobM
879
+ 1d8TExpPY4C2xlRM9TL3kz8z+ukJrz2J1wmqUHJFovCVBATk9nfpM8VMiPGlCmuL2VsCFVlBmbTsG4C2/vEqaFxXvslK9
880
+ r5ZqN1kceTiBUyiDB1dQgzuoQwMYDOAZXuHNEc6L8+58LFpzTjZzDH/gfP4AjR6NVA=</latexit>(b)
881
+ <latexit sha1_base64="oMYcx8nAFvX5nYPedyKOzsjl4=">AB6nicbVDLSgNBEOyNrxhfUY9eBoMQL2FXgnoMevEY
882
+ 0TwgWcLspDcZMju7zMwKIeQTvHhQxKtf5M2/cZLsQRMLGoqbrq7gkRwbVz328mtrW9sbuW3Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApGtzO/9YRK81g+mnGCfkQHkoecUWOlhzI7xVLbsWdg6wSLyMlyFDvFb+6/ZilEUrDBNW647mJ
883
+ 8SdUGc4ETgvdVGNC2YgOsGOpBFqfzI/dUrOrNInYaxsSUPm6u+JCY20HkeB7YyoGeplbyb+53VSE17Ey6T1KBki0VhKoiJyexv0ucKmRFjSyhT3N5K2JAqyoxNp2BD8JZfXiXNi4p3WaneV0u1myOPJzAKZTBgyuowR3UoQEMBvAMr/DmCOfFeXc+Fq05
884
+ J5s5hj9wPn8AjqONVQ=</latexit>(c)
885
+ <latexit sha1_base64="H0
886
+ +vPkOYIB9WILVirvWzNP/lO+0=">AB6nicbVBNSwMxEJ3U
887
+ r1q/qh69BItQL2VXinosevFY0X5Au5RsNtuGZrNLkhXK0p/g
888
+ xYMiXv1F3vw3pu0etPXBwO9GWbm+Yng2jONyqsrW9sbhW3
889
+ Szu7e/sH5cOjto5TRVmLxiJWXZ9oJrhkLcONYN1EMRL5gnX8
890
+ 8e3M7zwxpXksH80kYV5EhpKHnBJjpYdqcD4oV5yaMwdeJW5O
891
+ KpCjOSh/9YOYphGThgqidc91EuNlRBlOBZuW+qlmCaFjMmQ9
892
+ SyWJmPay+alTfGaVAIexsiUNnqu/JzISaT2JfNsZETPSy95M
893
+ /M/rpSa89jIuk9QwSReLwlRgE+PZ3zjgilEjJpYQqri9FdMR
894
+ UYQam07JhuAuv7xK2hc197JWv69XGjd5HEU4gVOogtX0IA7
895
+ aEILKAzhGV7hDQn0gt7Rx6K1gPKZY/gD9PkDkCiNVg=</la
896
+ texit>(d)
897
+ <latexit sha1_base64="NUvJWY4nPSEWtRiIZudJfNYoFj
898
+ 4=">AB6nicbVDLSgNBEOyNrxhfUY9eBoMQL2FXgnoMevEY0TwgWcLspDcZMju7zMwKIeQTvHhQxKtf5M2/cZLsQRMLGoq
899
+ brq7gkRwbVz328mtrW9sbuW3Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApGtzO/9YRK81g+mnGCfkQHkoec
900
+ UWOlhzKe94olt+LOQVaJl5ESZKj3il/dfszSCKVhgmrd8dzE+BOqDGcCp4VuqjGhbEQH2LFU0gi1P5mfOiVnVumTMFa2pCFz
901
+ 9fEhEZaj6PAdkbUDPWyNxP/8zqpCa/9CZdJalCyxaIwFcTEZPY36XOFzIixJZQpbm8lbEgVZcamU7AheMsvr5LmRcW7rFTv
902
+ q6XaTRZHk7gFMrgwRXU4A7q0AGA3iGV3hzhPivDsfi9ack80cwx84nz+RrY1X</latexit>(e)
903
+ Figure 3. The pathway activity level measured in terms of the aggregated log-likelihood ratios (LLRs) in response to different levels of radiation exposure.
904
+ Dose-dependent activity level is shown for the top five pathways that were most differentially activated under low-dose radiation exposure. (a) Natural
905
+ killer cell mediated cytotoxicity (b) Adherens junction (c) Sphingolipid metabolism (d) JAK-STAT signaling pathway (e) Glycosphingolipid biosynthesis.
906
+ All plots in (a)–(e) for the top low-dose response pathways display similar trends, where the differential activity levels reflecting the presence of potential
907
+ molecular signatures of low-dose radiation response decrease as the radiation dose level increases.
908
+ <latexit sha1_base64="2R
909
+ RwxLXlY8TROIoM98j2WcOjpro=">AB6nicbVDLSgNBEOyN
910
+ rxhfUY9eBoMQL2FXgnoMevEY0TwgWcLspDcZMju7zMwKIeQT
911
+ vHhQxKtf5M2/cZLsQRMLGoqbrq7gkRwbVz328mtrW9sbuW3
912
+ Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApG
913
+ tzO/9YRK81g+mnGCfkQHkoecUWOlhzI97xVLbsWdg6wSLyMl
914
+ yFDvFb+6/ZilEUrDBNW647mJ8SdUGc4ETgvdVGNC2YgOsGOp
915
+ pBFqfzI/dUrOrNInYaxsSUPm6u+JCY20HkeB7YyoGeplbyb+
916
+ 53VSE17Ey6T1KBki0VhKoiJyexv0ucKmRFjSyhT3N5K2JAq
917
+ yoxNp2BD8JZfXiXNi4p3WaneV0u1myOPJzAKZTBgyuowR3U
918
+ oQEMBvAMr/DmCOfFeXc+Fq05J5s5hj9wPn8Ai5mNUw=</la
919
+ texit>(a)
920
+ <latexit sha1_base64="KWh0RLJ0bw8em/x3PU2+HIlN2F
921
+ Q=">AB6nicbVDLSgNBEOyNrxhfUY9eBoMQL2FXgnoMevEY0TwgWcLspDcZMju7zMwKIeQTvHhQxKtf5M2/cZLsQRMLGoq
922
+ brq7gkRwbVz328mtrW9sbuW3Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApGtzO/9YRK81g+mnGCfkQHkoec
923
+ UWOlh3Jw3iuW3Io7B1klXkZKkKHeK351+zFLI5SGCap1x3MT40+oMpwJnBa6qcaEshEdYMdSPU/mR+6pScWaVPwljZkobM
924
+ 1d8TExpPY4C2xlRM9TL3kz8z+ukJrz2J1wmqUHJFovCVBATk9nfpM8VMiPGlCmuL2VsCFVlBmbTsG4C2/vEqaFxXvslK9
925
+ r5ZqN1kceTiBUyiDB1dQgzuoQwMYDOAZXuHNEc6L8+58LFpzTjZzDH/gfP4AjR6NVA=</latexit>(b)
926
+ <latexit sha1_base64="oMYcx8nAFvX5nYPedyKOzsjl4=">AB6nicbVDLSgNBEOyNrxhfUY9eBoMQL2FXgnoMevEY
927
+ 0TwgWcLspDcZMju7zMwKIeQTvHhQxKtf5M2/cZLsQRMLGoqbrq7gkRwbVz328mtrW9sbuW3Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApGtzO/9YRK81g+mnGCfkQHkoecUWOlhzI7xVLbsWdg6wSLyMlyFDvFb+6/ZilEUrDBNW647mJ
928
+ 8SdUGc4ETgvdVGNC2YgOsGOpBFqfzI/dUrOrNInYaxsSUPm6u+JCY20HkeB7YyoGeplbyb+53VSE17Ey6T1KBki0VhKoiJyexv0ucKmRFjSyhT3N5K2JAqyoxNp2BD8JZfXiXNi4p3WaneV0u1myOPJzAKZTBgyuowR3UoQEMBvAMr/DmCOfFeXc+Fq05
929
+ J5s5hj9wPn8AjqONVQ=</latexit>(c)
930
+ <latexit sha1_base64="H0
931
+ +vPkOYIB9WILVirvWzNP/lO+0=">AB6nicbVBNSwMxEJ3U
932
+ r1q/qh69BItQL2VXinosevFY0X5Au5RsNtuGZrNLkhXK0p/g
933
+ xYMiXv1F3vw3pu0etPXBwO9GWbm+Yng2jONyqsrW9sbhW3
934
+ Szu7e/sH5cOjto5TRVmLxiJWXZ9oJrhkLcONYN1EMRL5gnX8
935
+ 8e3M7zwxpXksH80kYV5EhpKHnBJjpYdqcD4oV5yaMwdeJW5O
936
+ KpCjOSh/9YOYphGThgqidc91EuNlRBlOBZuW+qlmCaFjMmQ9
937
+ SyWJmPay+alTfGaVAIexsiUNnqu/JzISaT2JfNsZETPSy95M
938
+ /M/rpSa89jIuk9QwSReLwlRgE+PZ3zjgilEjJpYQqri9FdMR
939
+ UYQam07JhuAuv7xK2hc197JWv69XGjd5HEU4gVOogtX0IA7
940
+ aEILKAzhGV7hDQn0gt7Rx6K1gPKZY/gD9PkDkCiNVg=</la
941
+ texit>(d)
942
+ <latexit sha1_base64="NUvJWY4nPSEWtRiIZudJfNYoFj
943
+ 4=">AB6nicbVDLSgNBEOyNrxhfUY9eBoMQL2FXgnoMevEY0TwgWcLspDcZMju7zMwKIeQTvHhQxKtf5M2/cZLsQRMLGoq
944
+ brq7gkRwbVz328mtrW9sbuW3Czu7e/sHxcOjpo5TxbDBYhGrdkA1Ci6xYbgR2E4U0igQ2ApGtzO/9YRK81g+mnGCfkQHkoec
945
+ UWOlhzKe94olt+LOQVaJl5ESZKj3il/dfszSCKVhgmrd8dzE+BOqDGcCp4VuqjGhbEQH2LFU0gi1P5mfOiVnVumTMFa2pCFz
946
+ 9fEhEZaj6PAdkbUDPWyNxP/8zqpCa/9CZdJalCyxaIwFcTEZPY36XOFzIixJZQpbm8lbEgVZcamU7AheMsvr5LmRcW7rFTv
947
+ q6XaTRZHk7gFMrgwRXU4A7q0AGA3iGV3hzhPivDsfi9ack80cwx84nz+RrY1X</latexit>(e)
948
+ Figure 4. The pathway activity level measured in terms of the aggregated log-likelihood ratios (LLRs) in response to different levels of radiation exposure.
949
+ As before, dose-dependent activity level is shown for the top five pathways that were most differentially activated under high-dose radiation exposure. (a)
950
+ Natural killer cell mediated cytotoxicity (b) Graft-versus-host disease (c) Viral myocarditis (d) Allograft rejection (e) Autoimmune thyroid disease. Except
951
+ for the top pathway in (a), the differential activity levels reflecting the presence of potential molecular signatures of high-dose radiation response do not
952
+ significantly change as the radiation dose level increases. This implies that the pathways that are responsive to high-dose radiation exposure may not be
953
+ substantially perturbed under relatively lower-dose radiation exposure.
954
+ they are not. We classify the samples as having a Gaussian
955
+ distribution if the P-value is greater than 0.05; otherwise,
956
+ they have a non-Gaussian distribution. Figure 6 depicts
957
+ the computed results, which show that 70.45 percent of
958
+ the low-dose samples and 89.63 percent of the high-dose
959
+ samples adhere to the Gaussian assumption. This indicates
960
+ that during the pathway analysis, it is appropriate to assume
961
+ that the conditional distribution of the gene expression data
962
+ is Gaussian.
963
+
964
+ 1.8
965
+ 1.6
966
+ 1.4
967
+ 1.2
968
+ 1.0
969
+ 0.8
970
+ 0
971
+ 0.005
972
+ 0.01
973
+ 0.025
974
+ 0.05
975
+ 0.1
976
+ 0.5
977
+ Dose level1.6
978
+ 1.4
979
+ 1.2
980
+ 1.0
981
+ 0.8
982
+ 0
983
+ 0.005
984
+ 0.01
985
+ 0.025
986
+ 0.05
987
+ 0.1
988
+ 0.5
989
+ Dose level1.3
990
+ 1.2
991
+ 1.1
992
+ 1.0
993
+ 0.9
994
+ 0.8
995
+ 0.7
996
+ 0
997
+ 0.005
998
+ 0.01
999
+ 0.025
1000
+ 0.05
1001
+ 0.1
1002
+ 0.5
1003
+ Dose level1.4
1004
+ 1.2
1005
+ 1.0
1006
+ 0.8
1007
+ 0
1008
+ 0.005
1009
+ 0.01
1010
+ 0.025
1011
+ 0.05
1012
+ 0.1
1013
+ 0.5
1014
+ Dose level1.4
1015
+ 1.2
1016
+ 1.0
1017
+ 0.8
1018
+ 0
1019
+ 0.005
1020
+ 0.01
1021
+ 0.025
1022
+ 0.05
1023
+ 0.1
1024
+ 0.5
1025
+ Dose level1.8
1026
+ 1.6
1027
+ 1.4
1028
+ 1.2
1029
+ 1.0
1030
+ 0.8
1031
+ 0
1032
+ 0.005
1033
+ 0.01
1034
+ 0.025
1035
+ 0.05
1036
+ 0.1
1037
+ 0.5
1038
+ Dose level4
1039
+ 2
1040
+ 0
1041
+ 2
1042
+ 0
1043
+ 0.005
1044
+ 0.01
1045
+ 0.025
1046
+ 0.05
1047
+ 0.1
1048
+ 0.5
1049
+ Dose level2.0
1050
+ 1.5
1051
+ 1.0
1052
+ 0
1053
+ 0.005
1054
+ 0.01
1055
+ 0.025
1056
+ 0.05
1057
+ 0.1
1058
+ 0.5
1059
+ Dose level2
1060
+ 0
1061
+ 2
1062
+ 0
1063
+ 0.005
1064
+ 0.01
1065
+ 0.025
1066
+ 0.05
1067
+ 0.1
1068
+ 0.5
1069
+ Dose level2
1070
+ 1
1071
+ 0
1072
+ -2
1073
+ 0
1074
+ 0.005
1075
+ 0.01
1076
+ 0.025
1077
+ 0.05
1078
+ 0.1
1079
+ 0.5
1080
+ Dose level<latexit sha1_
1081
+ base64="2RwxLXlY8TROIoM98j2
1082
+ WcOjpro=">AB6nicbVDLSgNBEO
1083
+ yNrxhfUY9eBoMQL2FXgnoMevEY0T
1084
+ wgWcLspDcZMju7zMwKIeQTvHhQxK
1085
+ tf5M2/cZLsQRMLGoqbrq7gkRwbV
1086
+ z328mtrW9sbuW3Czu7e/sHxcOjpo
1087
+ 5TxbDBYhGrdkA1Ci6xYbgR2E4U0i
1088
+ gQ2ApGtzO/9YRK81g+mnGCfkQHko
1089
+ ecUWOlhzI97xVLbsWdg6wSLyMlyF
1090
+ DvFb+6/ZilEUrDBNW647mJ8SdUGc
1091
+ 4ETgvdVGNC2YgOsGOpBFqfzI/dU
1092
+ rOrNInYaxsSUPm6u+JCY20HkeB7Y
1093
+ yoGeplbyb+53VSE17Ey6T1KBki0
1094
+ VhKoiJyexv0ucKmRFjSyhT3N5K2J
1095
+ AqyoxNp2BD8JZfXiXNi4p3WaneV0
1096
+ u1myOPJzAKZTBgyuowR3UoQEMBv
1097
+ AMr/DmCOfFeXc+Fq05J5s5hj9wPn
1098
+ 8Ai5mNUw=</latexit>(a)
1099
+ <latexit sha1_
1100
+ base64="KWh0RLJ0bw8em/x3PU2+
1101
+ HIlN2FQ=">AB6nicbVDLSgNBEO
1102
+ yNrxhfUY9eBoMQL2FXgnoMevEY0T
1103
+ wgWcLspDcZMju7zMwKIeQTvHhQxK
1104
+ tf5M2/cZLsQRMLGoqbrq7gkRwbV
1105
+ z328mtrW9sbuW3Czu7e/sHxcOjpo
1106
+ 5TxbDBYhGrdkA1Ci6xYbgR2E4U0i
1107
+ gQ2ApGtzO/9YRK81g+mnGCfkQHko
1108
+ ecUWOlh3Jw3iuW3Io7B1klXkZKkK
1109
+ HeK351+zFLI5SGCap1x3MT40+oMp
1110
+ wJnBa6qcaEshEdYMdSPU/mR+6p
1111
+ ScWaVPwljZkobM1d8TExpPY4C2x
1112
+ lRM9TL3kz8z+ukJrz2J1wmqUHJFo
1113
+ vCVBATk9nfpM8VMiPGlCmuL2VsC
1114
+ FVlBmbTsG4C2/vEqaFxXvslK9r5
1115
+ ZqN1kceTiBUyiDB1dQgzuoQwMYDO
1116
+ AZXuHNEc6L8+58LFpzTjZzDH/gfP
1117
+ 4AjR6NVA=</latexit>(b)
1118
+ Figure 5. Cross validation results of the top ranked pathways. (a) Cross-
1119
+ validation results for pathways most responsive to low-dose radiation.
1120
+ (b) Cross-validation results for pathways most responsive to high-dose
1121
+ radiation.
1122
+ Figure 6. Kolmogorov-Smirnov (KS) test results. We checked the normality
1123
+ of the gene expression values in low-dose and high-dose samples using the
1124
+ KS test. Results indicate that the Gaussian assumption holds in most cases.
1125
+ 5. Conclusion
1126
+ The current study aimed to unveil molecular signatures
1127
+ of biological responses exposed to low or very low doses
1128
+ of ionizing radiation through pathway-based analysis of
1129
+ genome-wide expression profiles. Gene expression patterns
1130
+ under the radiation exposure at six different dose levels
1131
+ ranging from 5 mGy to 500 mGy were investigated, where
1132
+ the measurements in the original study [18] were made using
1133
+ blood samples obtained from five different donors during
1134
+ five independent irradiation sessions. Our investigation was
1135
+ conducted at the pathway level, as pathway-based gene
1136
+ expression analysis is known to yield more robust and repro-
1137
+ ducible results and as it may shed light on potential molecu-
1138
+ lar mechanisms underlying low-dose radiation response. To
1139
+ determine the differential activity level of a given pathway
1140
+ under different levels of radiation exposure, a probabilistic
1141
+ pathway activity inference scheme was adopted that aggre-
1142
+ gates the log-likelihood ratios (LLRs) of the member genes
1143
+ in a given pathway to infer its differential activity. This
1144
+ allows robust detection of pathways, whose member genes
1145
+ display possibly subtle yet consistent coordinated expression
1146
+ patterns in response to low-dose radiation exposure. We
1147
+ searched through the KEGG database to prioritize pathways
1148
+ based on their differential activity levels modulated by low-
1149
+ dose radiation exposure. Our analysis identified the top
1150
+ pathways that may be associated with low-dose radiation re-
1151
+ sponse. Findings in this study reflect the complicated nature
1152
+ of the biological response to low-dose ionizing radiation,
1153
+ as well as the fact that low-dose exposures affect many
1154
+ different gene pathways that are not significantly altered
1155
+ after higher doses of radiotherapy.
1156
+ One limitation of the current study is the small sample
1157
+ size of the analyzed dataset (GSE43151). While it has been
1158
+ challenging to find large-scale human gene expression data
1159
+ under low-dose radiation exposure, should such data be
1160
+ available in the future, their analysis would shed further light
1161
+ onto the unique molecular signatures of low-dose radiation
1162
+ response. Furthermore, the pathway activity level inference
1163
+ scheme in (2) makes specific modeling assumptions, upon
1164
+ which the derived results depend. In fact, the adopted
1165
+ scheme [16] assumes that the gene expression levels of
1166
+ the member genes in a given pathway are conditionally
1167
+ independent given the class label (e.g., presence/absence of
1168
+ radiation exposure as was considered in the current study)
1169
+ and follow Gaussian distributions. Although we carried out
1170
+ some preliminary validation of this modeling assumption
1171
+ (e.g., see Fig. 6), it would be also worth validating the
1172
+ pathway analysis results using other methods [33], [34],
1173
+ which may be potentially pursued in our future studies.
1174
+ Acknowledgements
1175
+ This work is supported by the U.S. Department of
1176
+ Energy, Office of Science, RadBio program under Award
1177
+ KP1601011/FWP CC121.
1178
+ References
1179
+ [1]
1180
+ N. R. Council et al., “Health risks from exposure to low levels of
1181
+ ionizing radiation: Beir vii phase 2,” 2006.
1182
+ [2]
1183
+ W. M. Bonner, “Low-dose radiation: thresholds, bystander effects,
1184
+ and adaptive responses,” Proceedings of the National Academy of
1185
+ Sciences, vol. 100, no. 9, pp. 4973–4975, 2003.
1186
+ [3]
1187
+ L. Mullenders, M. Atkinson, H. Paretzke, L. Sabatier, and S. Bouf-
1188
+ fler, “Assessing cancer risks of low-dose radiation,” Nature Reviews
1189
+ Cancer, vol. 9, no. 8, pp. 596–604, 2009.
1190
+ [4]
1191
+ D. J. Brenner, R. Doll, D. T. Goodhead, E. J. Hall, C. E. Land,
1192
+ J. B. Little, J. H. Lubin, D. L. Preston, R. J. Preston, J. S. Puskin
1193
+ et al., “Cancer risks attributable to low doses of ionizing radiation:
1194
+ assessing what we really know,” Proceedings of the National Academy
1195
+ of Sciences, vol. 100, no. 24, pp. 13 761–13 766, 2003.
1196
+ [5]
1197
+ M. Tubiana, L. E. Feinendegen, C. Yang, and J. M. Kaminski, “The
1198
+ linear no-threshold relationship is inconsistent with radiation biologic
1199
+ and experimental data,” Radiology, vol. 251, no. 1, pp. 13–22, 2009.
1200
+
1201
+ 100
1202
+ Gaussian
1203
+ Non-Gaussian
1204
+ 80
1205
+ Percentage
1206
+ 60
1207
+ 40
1208
+ 20
1209
+ 0
1210
+ Low dose
1211
+ High doseNatural killer cell mediated cytotoxicity
1212
+ Name
1213
+ JAK-STAT signaling pathway
1214
+ Pathway
1215
+ Glycosphingolipid biosynthesis - lacto and neolacto series
1216
+ T cell receptor signaling pathway
1217
+ Human T-cell leukemia virus 1 infection
1218
+ 0
1219
+ 10
1220
+ 0
1221
+ 10
1222
+ Counts
1223
+ Aggregated t-test scoreGlycosphingolipid biosynthesis - lacto and neolacto series
1224
+ T cell receptor signaling pathway
1225
+ Cellular senescence
1226
+ Human T-cell leukemia virus 1 infection
1227
+ JAK-STAT signaling pathway
1228
+ 0
1229
+ 5
1230
+ 0
1231
+ 10
1232
+ Counts
1233
+ Aggregated t-test score[6]
1234
+ T. R. Golub, D. K. Slonim, P. Tamayo, C. Huard, M. Gaasenbeek,
1235
+ J. P. Mesirov, H. Coller, M. L. Loh, J. R. Downing, M. A. Caligiuri
1236
+ et al., “Molecular classification of cancer: class discovery and class
1237
+ prediction by gene expression monitoring,” science, vol. 286, no.
1238
+ 5439, pp. 531–537, 1999.
1239
+ [7]
1240
+ A. A. Alizadeh, M. B. Eisen, R. E. Davis, C. Ma, I. S. Lossos,
1241
+ A. Rosenwald, J. C. Boldrick, H. Sabet, T. Tran, X. Yu et al., “Distinct
1242
+ types of diffuse large b-cell lymphoma identified by gene expression
1243
+ profiling,” Nature, vol. 403, no. 6769, pp. 503–511, 2000.
1244
+ [8]
1245
+ T. Ideker, V. Thorsson, A. F. Siegel, and L. E. Hood, “Testing for
1246
+ differentially-expressed genes by maximum-likelihood analysis of
1247
+ microarray data,” Journal of computational biology, vol. 7, no. 6,
1248
+ pp. 805–817, 2000.
1249
+ [9]
1250
+ P. Baldi and A. D. Long, “A bayesian framework for the analysis of
1251
+ microarray expression data: regularized t-test and statistical inferences
1252
+ of gene changes,” Bioinformatics, vol. 17, no. 6, pp. 509–519, 2001.
1253
+ [10] U. M. Braga-Neto and E. R. Dougherty, “Is cross-validation valid
1254
+ for small-sample microarray classification?” Bioinformatics, vol. 20,
1255
+ no. 3, pp. 374–380, 2004.
1256
+ [11] S. Michiels, S. Koscielny, and C. Hill, “Prediction of cancer outcome
1257
+ with microarrays: a multiple random validation strategy,” The Lancet,
1258
+ vol. 365, no. 9458, pp. 488–492, 2005.
1259
+ [12] L. Ein-Dor, O. Zuk, and E. Domany, “Thousands of samples are
1260
+ needed to generate a robust gene list for predicting outcome in
1261
+ cancer,” Proceedings of the National Academy of Sciences, vol. 103,
1262
+ no. 15, pp. 5923–5928, 2006.
1263
+ [13] P. Khatri, M. Sirota, and A. J. Butte, “Ten years of pathway analysis:
1264
+ current approaches and outstanding challenges,” PLoS computational
1265
+ biology, vol. 8, no. 2, p. e1002375, 2012.
1266
+ [14] E. Lee, H.-Y. Chuang, J.-W. Kim, T. Ideker, and D. Lee, “Inferring
1267
+ pathway activity toward precise disease classification,” PLoS compu-
1268
+ tational biology, vol. 4, no. 11, p. e1000217, 2008.
1269
+ [15] M. L. Gatza, J. E. Lucas, W. T. Barry, J. W. Kim, Q. Wang, M. D.
1270
+ Crawford, M. B. Datto, M. Kelley, B. Mathey-Prevot, A. Potti et al.,
1271
+ “A pathway-based classification of human breast cancer,” Proceedings
1272
+ of the National Academy of Sciences, vol. 107, no. 15, pp. 6994–6999,
1273
+ 2010.
1274
+ [16] J. Su, B.-J. Yoon, and E. R. Dougherty, “Accurate and reliable cancer
1275
+ classification based on probabilistic inference of pathway activity,”
1276
+ PLOS ONE, vol. 4, no. 12, p. e8161, 2009.
1277
+ [17] N. Khunlertgit and B.-J. Yoon, “Incorporating topological information
1278
+ for predicting robust cancer subnetwork markers in human protein-
1279
+ protein interaction network,” in BMC bioinformatics, vol. 17, no. 13.
1280
+ Springer, 2016, pp. 143–152.
1281
+ [18] I. Nosel, A. Vaurijoux, J.-F. Barquinero, and G. Gruel, “Characteriza-
1282
+ tion of gene expression profiles at low and very low doses of ionizing
1283
+ radiation,” DNA repair, vol. 12, no. 7, pp. 508–517, 2013.
1284
+ [19] W. Luo, M. S. Friedman, K. Shedden, K. D. Hankenson, and P. J.
1285
+ Woolf, “Gage: generally applicable gene set enrichment for pathway
1286
+ analysis,” BMC bioinformatics, vol. 10, no. 1, pp. 1–17, 2009.
1287
+ [20] M. Kanehisa and S. Goto, “Kegg: kyoto encyclopedia of genes and
1288
+ genomes,” Nucleic acids research, vol. 28, no. 1, pp. 27–30, 2000.
1289
+ [21] L. Tian, S. A. Greenberg, S. W. Kong, J. Altschuler, I. S. Kohane, and
1290
+ P. J. Park, “Discovering statistically significant pathways in expression
1291
+ profiling studies,” Proceedings of the National Academy of Sciences,
1292
+ vol. 102, no. 38, pp. 13 544–13 549, 2005.
1293
+ [22] G. Yang, Q. Kong, G. Wang, H. Jin, L. Zhou, D. Yu, C. Niu,
1294
+ W. Han, W. Li, and J. Cui, “Low-dose ionizing radiation induces
1295
+ direct activation of natural killer cells and provides a novel approach
1296
+ for adoptive cellular immunotherapy,” Cancer Biotherapy and Radio-
1297
+ pharmaceuticals, vol. 29, no. 10, pp. 428–434, 2014.
1298
+ [23] J.-U. Jeong, T. N. T. Uong, W.-K. Chung, T.-K. Nam, S.-J. Ahn, J.-Y.
1299
+ Song, S.-K. Kim, D.-J. Shin, E. Cho, K. W. Kim et al., “Effect of
1300
+ irradiation-induced intercellular adhesion molecule-1 expression on
1301
+ natural killer cell-mediated cytotoxicity toward human cancer cells,”
1302
+ Cytotherapy, vol. 20, no. 5, pp. 715–727, 2018.
1303
+ [24] N. S´andor, F. R. Walter, A. Bocsik, P. S´antha, B. Schilling-T´oth,
1304
+ V. L´ener, Z. Varga, Z. Kah´an, M. A. Deli, G. S´afr´any et al., “Low
1305
+ dose cranial irradiation-induced cerebrovascular damage is reversible
1306
+ in mice,” PloS one, vol. 9, no. 11, p. e112397, 2014.
1307
+ [25] P. K. Shukla, R. Gangwar, B. Manda, A. S. Meena, N. Yadav,
1308
+ E. Szabo, A. Balogh, S. C. Lee, G. Tigyi, and R. Rao, “Rapid
1309
+ disruption of intestinal epithelial tight junction and barrier dysfunction
1310
+ by ionizing radiation in mouse colon in vivo: protection by n-acetyl-l-
1311
+ cysteine,” American Journal of Physiology-Gastrointestinal and Liver
1312
+ Physiology, vol. 310, no. 9, pp. G705–G715, 2016.
1313
+ [26] C. Shao, M. Folkard, B. D. Michael, and K. M. Prise, “Targeted
1314
+ cytoplasmic irradiation induces bystander responses,” Proceedings of
1315
+ the National Academy of Sciences, vol. 101, no. 37, pp. 13 495–
1316
+ 13 500, 2004.
1317
+ [27] J.-P. Truman, M. Garc´ıa-Barros, L. M. Obeid, and Y. A. Hannun,
1318
+ “Evolving concepts in cancer therapy through targeting sphingolipid
1319
+ metabolism,” Biochimica et Biophysica Acta (BBA)-Molecular and
1320
+ Cell Biology of Lipids, vol. 1841, no. 8, pp. 1174–1188, 2014.
1321
+ [28] S. Y. Lai, E. E. Childs, S. Xi, F. M. Coppelli, W. E. Gooding, A. Wells,
1322
+ R. L. Ferris, and J. R. Grandis, “Erythropoietin-mediated activation
1323
+ of jak-stat signaling contributes to cellular invasion in head and neck
1324
+ squamous cell carcinoma,” Oncogene, vol. 24, no. 27, pp. 4442–4449,
1325
+ 2005.
1326
+ [29] M. Aureli, V. Murdica, N. Loberto, M. Samarani, A. Prinetti, R. Bassi,
1327
+ and S. Sonnino, “Exploring the link between ceramide and ionizing
1328
+ radiation,” Glycoconjugate journal, vol. 31, no. 6, pp. 449–459, 2014.
1329
+ [30] G. R. Hill and J. L. Ferrara, “The primacy of the gastrointestinal tract
1330
+ as a target organ of acute graft-versus-host disease: rationale for the
1331
+ use of cytokine shields in allogeneic bone marrow transplantation,”
1332
+ Blood, The Journal of the American Society of Hematology, vol. 95,
1333
+ no. 9, pp. 2754–2759, 2000.
1334
+ [31] B. R. Blazar, W. J. Murphy, and M. Abedi, “Advances in graft-
1335
+ versus-host disease biology and therapy,” Nature reviews immunology,
1336
+ vol. 12, no. 6, pp. 443–458, 2012.
1337
+ [32] M. R. Van den Brink and S. J. Burakoff, “Cytolytic pathways in
1338
+ haematopoietic stem-cell transplantation,” Nature Reviews Immunol-
1339
+ ogy, vol. 2, no. 4, pp. 273–281, 2002.
1340
+ [33] J. Han, X. Shi, Y. Zhang, Y. Xu, Y. Jiang, C. Zhang, L. Feng, H. Yang,
1341
+ D. Shang, Z. Sun et al., “Esea: discovering the dysregulated pathways
1342
+ based on edge set enrichment analysis,” Scientific reports, vol. 5,
1343
+ no. 1, pp. 1–15, 2015.
1344
+ [34] M. Rydenfelt, B. Klinger, M. Kl¨unemann, and N. Bl¨uthgen, “Speed2:
1345
+ inferring upstream pathway activity from differential gene expres-
1346
+ sion,” Nucleic acids research, vol. 48, no. W1, pp. W307–W312,
1347
+ 2020.
1348
+
OdAzT4oBgHgl3EQfzf5C/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff