tensorkelechi/arxiv-cstext · Datasets at Hugging Face

text stringlengths 0 89.3k
while reducing computational resources
However existing medical LLMs do not fully leverage the
capabilities of RAG The description or medical terminology
provided by the users limited level of medical knowledge may
not accurately recall useful content Therefore research on
keyword analysis and rewriting of user queries is necessary
to improve knowledge retrieval accuracy and better utilize the
models medical knowledge Additionally healthcare organi
zations typically maintain electronic medical records for their
patients Investigating how to use this information to assist
medical LLMs in providing more personalized services to
patients is a valuable research direction
C Evaluation Platform for Medical Large Language Models
Currently several benchmarks or platforms exist for eval
uating general LLMs such as GLUE 143 SuperGLUE
144 and CLUE 145 These platforms provide consis
tent evaluation criteria and standardized training and testing
sets facilitating fair comparisons and advancements among
LLMs Existing medical LLMs are evaluated using incon
sistent datasets and ambiguous evaluation tasks resulting in
comparisons and improvements between models becoming
difficult Therefore it is particularly necessary to establish
a unified evaluation platform focused on medical LLMs to
foster the advancement of this field
Establishing a unified evaluation platform for medical LLMs
necessitates attention to several critical considerations Firstlythe platform should incorporate benchmarks across various
medical subdomains to ensure comprehensive evaluation of
model performance across diverse areas Secondly the evalu
ation tasks should be clearly defined encompassing medical
QA case diagnosis medical reasoning treatment recommen
dation and other pertinent aspects to accurately reflect the
models practical applicability Lastly the evaluation platform
should encompass a diverse range of metrics and dimensions
including traditional measures such as accuracy and BLEU
alongside considerations of dimensions such as proactive
helpful and safe
VI C ONCLUSION
This survey provides a guide on training customized medical
LLMs based on general LLMs Customized medical LLMs can
meet the personalized demands of individuals and healthcare
organizations We detail the training process at three levels
data methodology and evaluation We analyze the data level in
terms of corpus sources and processing methods the method
ology level in terms of training paradigms and the evaluation
level from both machine and human perspectives For each
level we offer recommendations and guidance Finally we
present the challenges and future research directions for each
of these three levels
ACKNOWLEDGMENTS
REFERENCES
1 J Achiam S Adler S Agarwal L Ahmad I Akkaya F L Aleman
D Almeida J Altenschmidt S Altman S Anadkat et al Gpt4
technical report arXiv preprint arXiv230308774 2023
2 Y Dan Z Lei Y Gu Y Li J Yin J Lin L Ye Z Tie Y Zhou
Y Wang et al Educhat A largescale language modelbased chatbot
system for intelligent education arXiv preprint arXiv230802773
2023
3 S Wang T Xu H Li C Zhang J Liang J Tang P S Yu and
Q Wen Large language models for education A survey and outlook
arXiv preprint arXiv240318105 2024
4 J Cui Z Li Y Yan B Chen and L Yuan Chatlaw Opensource
legal large language model with integrated external knowledge bases
arXiv preprint arXiv230616092 2023
5 P Colombo T P Pires M Boudiaf D Culver R Melo C Corro
A F Martins F Esposito V L Raposo S Morgado et al Saullm
7b A pioneering large language model for law arXiv preprint
arXiv240303883 2024
6 Z Zhang C Chen B Liu C Liao Z Gong H Yu J Li and R Wang
Unifying the perspectives of nlp and software engineering A survey
on language models for code arXiv preprint arXiv231107989 2023
7 K Du R Rui H Chai L Fu W Xia Y Wang R Tang Y Yu
and W Zhang Codegrag Extracting composed syntax graphs for
retrieval augmented crosslingual code generation arXiv preprint
arXiv240502355 2024
8 H Zhang J Chen F Jiang F Yu Z Chen J Li G Chen X Wu
Z Zhang Q Xiao et al Huatuogpt towards taming language model
to be a doctor arXiv preprint arXiv230515075 2023
9 H Gwon I Ahn H Jung B Kim Y H Kim and T J Jun Inmdx
Large language models for internal medicine doctors arXiv preprint
arXiv240211883 2024
10 Y Li S Wang H Ding and H Chen Large language models in
finance A survey in Proceedings of the Fourth ACM International
Conference on AI in Finance pp 374382 2023
11 J Lee N Stevens S C Han and M Song A survey of large language
models in finance finllms arXiv preprint arXiv240202315 2024
12 C Ling X Zhao J Lu C Deng C Zheng J Wang T Chowdhury
Y Li H Cui T Zhao et al Beyond onemodelfitsall A survey
of domain specialization for large language models arXiv preprint
arXiv230518703 2023JOURNAL OF L ATEX CLASS FILES VOL 14 NO 8 AUGUST 2021 18
13 W X Zhao K Zhou J Li T Tang X Wang Y Hou Y Min
B Zhang J Zhang Z Dong et al A survey of large language
models arXiv preprint arXiv230318223 2023
14 L Ouyang J Wu X Jiang D Almeida C Wainwright P Mishkin
C Zhang S Agarwal K Slama A Ray et al Training language
models to follow instructions with human feedback Advances in
neural information processing systems vol 35 pp 2773027744
2022
15 S Gururangan A Marasovi c S Swayamdipta K Lo I Beltagy
D Downey and N A Smith Dont stop pretraining Adapt language
models to domains and tasks in Proceedings of the 58th Annual

while reducing computational resources

However existing medical LLMs do not fully leverage the

capabilities of RAG The description or medical terminology

provided by the users limited level of medical knowledge may

not accurately recall useful content Therefore research on

keyword analysis and rewriting of user queries is necessary

to improve knowledge retrieval accuracy and better utilize the

models medical knowledge Additionally healthcare organi

zations typically maintain electronic medical records for their

patients Investigating how to use this information to assist

medical LLMs in providing more personalized services to

patients is a valuable research direction

C Evaluation Platform for Medical Large Language Models

Currently several benchmarks or platforms exist for eval

uating general LLMs such as GLUE 143 SuperGLUE

144 and CLUE 145 These platforms provide consis

tent evaluation criteria and standardized training and testing

sets facilitating fair comparisons and advancements among

LLMs Existing medical LLMs are evaluated using incon

sistent datasets and ambiguous evaluation tasks resulting in

comparisons and improvements between models becoming

difficult Therefore it is particularly necessary to establish

a unified evaluation platform focused on medical LLMs to

foster the advancement of this field

Establishing a unified evaluation platform for medical LLMs

necessitates attention to several critical considerations Firstlythe platform should incorporate benchmarks across various

medical subdomains to ensure comprehensive evaluation of

model performance across diverse areas Secondly the evalu

ation tasks should be clearly defined encompassing medical

QA case diagnosis medical reasoning treatment recommen

dation and other pertinent aspects to accurately reflect the

models practical applicability Lastly the evaluation platform

should encompass a diverse range of metrics and dimensions

including traditional measures such as accuracy and BLEU

alongside considerations of dimensions such as proactive

helpful and safe

VI C ONCLUSION

This survey provides a guide on training customized medical

LLMs based on general LLMs Customized medical LLMs can

meet the personalized demands of individuals and healthcare

organizations We detail the training process at three levels

data methodology and evaluation We analyze the data level in

terms of corpus sources and processing methods the method

ology level in terms of training paradigms and the evaluation

level from both machine and human perspectives For each

level we offer recommendations and guidance Finally we

present the challenges and future research directions for each

of these three levels

ACKNOWLEDGMENTS

REFERENCES

1 J Achiam S Adler S Agarwal L Ahmad I Akkaya F L Aleman

D Almeida J Altenschmidt S Altman S Anadkat et al Gpt4

technical report arXiv preprint arXiv230308774 2023

2 Y Dan Z Lei Y Gu Y Li J Yin J Lin L Ye Z Tie Y Zhou

Y Wang et al Educhat A largescale language modelbased chatbot

system for intelligent education arXiv preprint arXiv230802773

2023

3 S Wang T Xu H Li C Zhang J Liang J Tang P S Yu and

Q Wen Large language models for education A survey and outlook

arXiv preprint arXiv240318105 2024

4 J Cui Z Li Y Yan B Chen and L Yuan Chatlaw Opensource

legal large language model with integrated external knowledge bases

arXiv preprint arXiv230616092 2023

5 P Colombo T P Pires M Boudiaf D Culver R Melo C Corro

A F Martins F Esposito V L Raposo S Morgado et al Saullm

7b A pioneering large language model for law arXiv preprint

arXiv240303883 2024

6 Z Zhang C Chen B Liu C Liao Z Gong H Yu J Li and R Wang

Unifying the perspectives of nlp and software engineering A survey

on language models for code arXiv preprint arXiv231107989 2023

7 K Du R Rui H Chai L Fu W Xia Y Wang R Tang Y Yu

and W Zhang Codegrag Extracting composed syntax graphs for

retrieval augmented crosslingual code generation arXiv preprint

arXiv240502355 2024

8 H Zhang J Chen F Jiang F Yu Z Chen J Li G Chen X Wu

Z Zhang Q Xiao et al Huatuogpt towards taming language model

to be a doctor arXiv preprint arXiv230515075 2023

9 H Gwon I Ahn H Jung B Kim Y H Kim and T J Jun Inmdx

Large language models for internal medicine doctors arXiv preprint

arXiv240211883 2024

10 Y Li S Wang H Ding and H Chen Large language models in

finance A survey in Proceedings of the Fourth ACM International

Conference on AI in Finance pp 374382 2023

11 J Lee N Stevens S C Han and M Song A survey of large language

models in finance finllms arXiv preprint arXiv240202315 2024

12 C Ling X Zhao J Lu C Deng C Zheng J Wang T Chowdhury

Y Li H Cui T Zhao et al Beyond onemodelfitsall A survey

of domain specialization for large language models arXiv preprint

arXiv230518703 2023JOURNAL OF L ATEX CLASS FILES VOL 14 NO 8 AUGUST 2021 18

13 W X Zhao K Zhou J Li T Tang X Wang Y Hou Y Min

B Zhang J Zhang Z Dong et al A survey of large language

models arXiv preprint arXiv230318223 2023

14 L Ouyang J Wu X Jiang D Almeida C Wainwright P Mishkin

C Zhang S Agarwal K Slama A Ray et al Training language

models to follow instructions with human feedback Advances in

neural information processing systems vol 35 pp 2773027744

2022

15 S Gururangan A Marasovi c S Swayamdipta K Lo I Beltagy

D Downey and N A Smith Dont stop pretraining Adapt language

models to domains and tasks in Proceedings of the 58th Annual