id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 6
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-05-16 00:00:00
| authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.05178 | Grigory Volovik | G.E. Volovik | Extended Tsallis-Cirto entropy for black and white holes | 5 pages, no figures | null | null | null | gr-qc cond-mat.other | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In reference [1] we considered the black hole thermodynamics with the
non-extensive entropy. This entropy obeys the composition rule which coincides
with the composition rule in the non-extensive Tsallis-Cirto $\delta=2$
statistics. Here we extend this approach to the thermodynamics of white holes.
The entropy of the white hole is negative as follows from the rate of
macroscopic quantum tunneling from black hole to white hole. The white hole
entropy is with the minus sign the entropy of the black hole with the same
mass, $S_{\rm WH}(M)=-S_{\rm BH}(M)$. This reflects the anti-symmetry with
respect to time reversal, at which the shift vector in the
Arnowitt-Deser-Misner formalism changes sign. This symmetry allows one to
extend the Tsallis-Chirto entropy by adding a minus sign to the Tsallis-Chirto
formula applied to white hole. As a result, the composition rule remains the
same, with the only difference being that instead of entropy it contains the
entropy modulus. The same non-extensive composition rule is obtained for the
entropy of the Reissner-Nordstr\"om black hole. This entropy is formed by the
positive entropy of the outer horizon and the negative entropy of the inner
horizon. The model of the black hole formed by "black hole atoms" with
Planck-scale mass is also extended to include the negative entropy of white
holes.
| [
{
"version": "v1",
"created": "Thu, 8 May 2025 12:28:21 GMT"
},
{
"version": "v2",
"created": "Mon, 12 May 2025 13:22:11 GMT"
},
{
"version": "v3",
"created": "Wed, 14 May 2025 16:34:14 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Volovik",
"G. E.",
""
]
] |
2505.05272 | Andr\'e Milagre | Andr\'e Milagre, Darius Jur\v{c}iukonis, Lu\'is Lavoura | Vacuum stability conditions for new $SU(2)$ multiplets | 33 pages, 2 figures; v2: added references and footnotes, corrected
the Acknowledgments | null | null | null | hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | We consider the addition to the Standard Model of a scalar $SU(2)$ multiplet
$\Delta_n$ with dimension $n$ going from $1$ to $6$. The multiplet $\Delta_n$
is assumed to have null vacuum expectation value and an arbitrary (free)
hypercharge. We determine the shape of the phase space for the new terms that
appear in the scalar potential (SP); we observe in particular that, in the case
of a 6-plet, the phase space is slightly concave along one of its boundaries.
We determine the bounded-from-below and vacuum stability conditions on the SP
for each value of $n$.
| [
{
"version": "v1",
"created": "Thu, 8 May 2025 14:18:09 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 14:09:47 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Milagre",
"André",
""
],
[
"Jurčiukonis",
"Darius",
""
],
[
"Lavoura",
"Luís",
""
]
] |
2505.05364 | Hao Liu | Yanbin Zhao, Hao Liu, Zhihua Deng, Tong Li, Haoyi Jiang, Zhenfei Ling,
Xingkai Wang, Lei Zhang, Xiaoping Ouyang | Machine learning bridging battery field data and laboratory data | 73 pages, 21 figures | null | null | null | stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Aiming at the dilemma that most laboratory data-driven diagnostic and
prognostic methods cannot be applied to field batteries in passenger cars and
energy storage systems, this paper proposes a method to bridge field data and
laboratory data using machine learning. Only two field real impedances
corresponding to a medium frequency and a high frequency are needed to predict
laboratory real impedance curve, laboratory charge/discharge curve, and
laboratory relaxation curve. Based on the predicted laboratory data, laboratory
data-driven methods can be used for field battery diagnosis and prognosis.
Compared with the field data-driven methods based on massive historical field
data, the proposed method has the advantages of higher accuracy, lower cost,
faster speed, readily available, and no use of private data. The proposed
method is tested using two open-source datasets containing 249 NMC cells. For a
test set containing 76 cells, the mean absolute percentage errors of laboratory
real impedance curve, charge curve, and discharge curve prediction results are
0.85%, 4.72%, and 2.69%, respectively. This work fills the gap between
laboratory data-driven diagnostic and prognostic methods and field battery
applications, making all laboratory data-driven methods applicable to field
battery diagnosis and prognosis. Furthermore, this work overturns the fixed
path of developing field battery diagnostic and prognostic methods based on
massive field historical data, opening up new research and breakthrough
directions for field battery diagnosis and prognosis.
| [
{
"version": "v1",
"created": "Thu, 8 May 2025 15:58:57 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 01:51:09 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Zhao",
"Yanbin",
""
],
[
"Liu",
"Hao",
""
],
[
"Deng",
"Zhihua",
""
],
[
"Li",
"Tong",
""
],
[
"Jiang",
"Haoyi",
""
],
[
"Ling",
"Zhenfei",
""
],
[
"Wang",
"Xingkai",
""
],
[
"Zhang",
"Lei",
""
],
[
"Ouyang",
"Xiaoping",
""
]
] |
2505.05507 | Igor Alentev | Igor Alentev, Lev Kozlov, Ivan Domrachev, Simeon Nedelchev, Jee-Hwan
Ryu | VIMPPI: Enhancing Model Predictive Path Integral Control with
Variational Integration for Underactuated Systems | null | null | null | null | eess.SY cs.RO cs.SY | http://creativecommons.org/licenses/by/4.0/ | This paper presents VIMPPI, a novel control approach for underactuated double
pendulum systems developed for the AI Olympics competition. We enhance the
Model Predictive Path Integral framework by incorporating variational
integration techniques, enabling longer planning horizons without additional
computational cost. Operating at 500-700 Hz with control interpolation and
disturbance detection mechanisms, VIMPPI substantially outperforms both
baseline methods and alternative MPPI implementations
| [
{
"version": "v1",
"created": "Wed, 7 May 2025 08:18:54 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 12:13:27 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Alentev",
"Igor",
""
],
[
"Kozlov",
"Lev",
""
],
[
"Domrachev",
"Ivan",
""
],
[
"Nedelchev",
"Simeon",
""
],
[
"Ryu",
"Jee-Hwan",
""
]
] |
2505.05776 | Zachary Lipel | Zachary G. Lipel, Yannick A. D. Omar, Dimitrios Fraggedakis | Finite Membrane Thickness Influences Hydrodynamics on the Nanoscale | 7 pages, 2 figures | null | null | null | cond-mat.soft physics.bio-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many lipid membrane-mediated transport processes--such as mechanically-gated
channel activation and solute transport--involve structural and dynamical
features on membrane thickness length scales. Most existing membrane models,
however, tend to adopt (quasi-)two-dimensional descriptions that neglect
thickness-dependent phenomena relevant to internal membrane mechanics, and thus
do not fully account for the complex coupling of lipid membranes with their
surrounding fluid media. Therefore, explicitly incorporating membrane thickness
effects in lipid membrane models will enable a more accurate description of the
influence of membrane/fluid coupling on transport phenomena in the vicinity of
the bilayer surfaces. Here, we present a continuum model for membrane
fluctuations that accounts for finite membrane thickness and resolves
hydrodynamic interactions between the bilayer and its surrounding fluid. By
applying linear response analysis, we observe that membrane thickness-mediated
effects, such as bending-induced lipid reorientations, can generate shear flows
close to the membrane surface that slow down the relaxation of nanometer scale
shape fluctuations. Additionally, we reveal the emergence of pressure inversion
and flow reversal near the membrane interfaces, accompanied by localized
stagnation points. Among these, extensional stagnation points give rise to a
novel mode of bulk dissipation, originating from bending-induced compression
and expansion of the membrane surfaces and their coupling to shear stresses in
the fluid. Our findings identify membrane thickness as a key factor in
nanoscale hydrodynamics and suggest that its effects may be detectable in
fluctuation spectra and can be relevant to interfacial processes such as solute
permeability and contact with solid boundaries or other membranes.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 04:37:55 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 15:33:35 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Lipel",
"Zachary G.",
""
],
[
"Omar",
"Yannick A. D.",
""
],
[
"Fraggedakis",
"Dimitrios",
""
]
] |
2505.05871 | Joan Bachs-Esteban | Joan Bachs-Esteban, Il\'idio Lopes, Javier Rubio | Screening Mechanisms on White Dwarfs: Symmetron & Dilaton | Corrected references. Matches published version | Universe 2025, 11(5), 158 | 10.3390/universe11050158 | null | gr-qc astro-ph.HE astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | This work provides the first comparison of the symmetron and dilaton fields
in white dwarfs. We show how these screening mechanisms behave inside {such
stars} and their impact on stellar properties. Employing a custom-developed
shooting method, we solve the scalar-tensor equilibrium equations in the
Newtonian approximation. We consider a Chandrasekhar equation of state and
examine a range of potential mass scales and coupling strengths for both
fields. Both fields enhance the pressure drop in low-density white dwarfs,
leading to smaller stellar masses, radii, and luminosities. Unlike chameleon
models, their effects are suppressed in more massive stars, with symmetron
fields fully decoupling and dilaton fields weakening but not vanishing.
Consequently, no mass-radius curve for screened white dwarfs exceeds the
Newtonian prediction in any of these three mechanisms. The mass-radius
deviations are generally more pronounced at lower densities, depending on model
parameters. Due to their common runaway potential, we confirm that dilaton and
chameleon fields display similar field and gradient profiles. In contrast, due
to their environment-dependent coupling, the dilaton and symmetron mechanisms
exhibit stronger density-dependent screening effects. These findings highlight
both phenomenological differences and theoretical similarities among these
mechanisms, motivating asteroseismology studies to constrain the symmetron and
dilaton parameter spaces.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 08:15:59 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 14:32:40 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Bachs-Esteban",
"Joan",
""
],
[
"Lopes",
"Ilídio",
""
],
[
"Rubio",
"Javier",
""
]
] |
2505.05925 | Yu Hao | Chang Li, Yangxiang Lu, Hao Yu | Infinite combinatorial Ricci flow in spherical background geometry | null | null | null | null | math.GT math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Since the fundamental work of Chow-Luo \cite{CL03}, Ge \cite{Ge12,Ge17} et
al., the combinatorial curvature flow methods became a basic technique in the
study of circle pattern theory. In this paper, we investigate the combinatorial
Ricci flow with prescribed total geodesic curvatures in spherical background
geometry. For infinite cellular decompositions, we establish the existence of a
solution to the flow equation for all time. Furthermore, under an additional
condition, we prove that the solution converges as time tends to infinity. To
the best of our knowledge, this is the first study of an infinite combinatorial
curvature flow in spherical background geometry.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 09:57:58 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 00:11:26 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Li",
"Chang",
""
],
[
"Lu",
"Yangxiang",
""
],
[
"Yu",
"Hao",
""
]
] |
2505.06006 | Heng Xu | Jiangwei Xu, Heng Xu, Yanjun Guo, Jinchen Jiang, Bojun Wang, Zihan
Xue, Yunpeng Men, Kejia Lee, Bing Zhang, Weiwei Zhu, Jinlin Han | Periodic variation of magnetoionic environment of a fast radio burst
source | Submitted, comments are welcome | null | null | null | astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fast radio bursts (FRBs) are luminous, dispersed millisecond-duration radio
bursts whose origin is poorly known. Recent observations suggest that some FRBs
may reside in binary systems, even though conclusive evidence remains elusive.
Here we report the detection of a 26.24$\pm$0.02 day periodicity in Faraday
rotation measure (RM) of an actively repeating source named FRB 20201124A. The
detection was made from 3,106 bursts collected with the Five-hundred-meter
Aperture Spherical radio Telescope (FAST) over ~365 days. The RM periodicity is
coherently phase-connected across ~14 cycles over a 1-year duration. Our
detection of RM periodicity corresponds to a natural logarithmic Bayesian
factor of 1,168. The detection significances vary between 5.9-34 {\sigma} under
different assumptions. Such a periodicity provides evidence for the binary
nature of FRB 20201124A, where the periodic RM variations arise from the
orbital motion of the FRB source within the magnetoionic environment of the
system. Together with previous observations, our result suggests that being in
binary systems may be a common feature for actively repeating FRB sources.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 12:35:51 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 03:22:34 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Xu",
"Jiangwei",
""
],
[
"Xu",
"Heng",
""
],
[
"Guo",
"Yanjun",
""
],
[
"Jiang",
"Jinchen",
""
],
[
"Wang",
"Bojun",
""
],
[
"Xue",
"Zihan",
""
],
[
"Men",
"Yunpeng",
""
],
[
"Lee",
"Kejia",
""
],
[
"Zhang",
"Bing",
""
],
[
"Zhu",
"Weiwei",
""
],
[
"Han",
"Jinlin",
""
]
] |
2505.06058 | Anna Fino | Beatrice Brienza, Anna Fino, Gueo Grantcharov and Misha Verbitsky | On the structure of compact strong HKT manifolds | 33 pages | null | null | null | math.DG hep-th | http://creativecommons.org/licenses/by/4.0/ | We study the geometry of compact strong HKT and, more generally, compact BHE
manifolds. We prove that any compact BHE manifold with full holonomy must be
K\"ahler and we establish a similar result for strong HKT manifolds.
Additionally, we demonstrate a rigidity theorem for strong HKT structures on
solvmanifolds and we completely classify those with parallel Bismut torsion.
Finally, we introduce the Ricci foliation for hypercomplex manifolds and
analyze its properties for compact, simply connected, 8-dimensional strong HKT
manifolds, proving that they are always Hopf fibrations over a compact
$4$-dimensional orbifold.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 13:53:03 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Brienza",
"Beatrice",
""
],
[
"Fino",
"Anna",
""
],
[
"Grantcharov",
"Gueo",
""
],
[
"Verbitsky",
"Misha",
""
]
] |
2505.06137 | Xianguo Lu | L. Alvarez Ruso, W. Chang, J. Franklin, P.R. Hobson, P.B. Jurj, R.
Kamath, P. Kyberd, X. Lu, K. Long, D. Pasari, S. Ricciardi, J. Turner, A.
Vogiatzi (on behalf of the nuSTORM collaboration) | Neutrinos from Stored Muons (nuSTORM) | Submitted to the 2026 Update of the European Strategy for Particle
Physics (https://indico.cern.ch/event/1439855/contributions/6461582/).
Contact: K. Long. 17 pages, 7 figures. v2: fixed top matter | null | null | null | hep-ex nucl-ex physics.acc-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The Neutrinos from Stored Muons, nuSTORM, facility has been designed to
deliver a definitive neutrino-nucleus scattering programme using beams of
,{\nu}e- and ,{\nu}{\mu}- from the decay of muons confined within a storage
ring. The facility is unique, it will be capable of storing {\mu} beams with a
central momentum of between 1 GeV/c and 6 GeV/c and a momentum spread of 16%.
This specification will allow neutrino-scattering measurements to be made over
the kinematic range of interest to the DUNE and Hyper-K collaborations. At
nuSTORM, the flavour composition of the beam and the neutrino-energy spectrum
are both precisely known. The storage-ring instrumentation will allow the
neutrino flux to be determined to a precision of 1% or better. With its
existing proton-beam infrastructure, CERN is uniquely well-placed to implement
nuSTORM. A summary of the proposed implementation of nuSTORM at CERN is
presented below.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 15:43:42 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 14:59:25 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Ruso",
"L. Alvarez",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Chang",
"W.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Franklin",
"J.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Hobson",
"P. R.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Jurj",
"P. B.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Kamath",
"R.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Kyberd",
"P.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Lu",
"X.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Long",
"K.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Pasari",
"D.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Ricciardi",
"S.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Turner",
"J.",
"",
"on behalf of the nuSTORM collaboration"
],
[
"Vogiatzi",
"A.",
"",
"on behalf of the nuSTORM collaboration"
]
] |
2505.06140 | Ramesh Sharma | Ramesh Sharma | A Conformal Quasi Einstein Characterization Of The Round Sphere | None | null | null | null | math.DG | http://creativecommons.org/licenses/by/4.0/ | We extend the following result of Cochran ``A closed $m$-quasi Einstein
manifold ($M,g,X$) with $m \ne -2$ has constant scalar curvature if and only if
$X$ is Killing" covering the missing accidental case $m=-2$ and generalize it
showing that $X$ is Killing if the integral of the Lie derivative of the scalar
curvature along $X$ is non-positive. For a closed $m$-quasi Einstein manifold
of dimension $n \ge 2$, if $X$ is conformal, then it is Killing; and in
addition, if $M$ admits a non-Killing conformal vector field $V$, then it is
globally isometric to a sphere and $V$ is gradient for $n > 2$. Finally, we
derive an integral identity for a vector field on a closed Riemannian manifold,
which provides a direct proof of the Bourguignon-Ezin conservation identity.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 15:49:16 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 17:47:31 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Sharma",
"Ramesh",
""
]
] |
2505.06372 | Przemyslaw Ignaciuk | Naohisa Otsuka, Daiki Kakehi, Przemys{\l}aw Ignaciuk | Interval reduced-order switched positive observers for uncertain
switched positive linear systems | 11 pages, 4 figures | Int. J. Control Autom. Syst. 22, 1105-1115, 2024 | 10.1007/s12555-023-0103-6 | null | eess.SY cs.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, existence conditions and a design procedure of reduced-order
switched positive observers for continuous- and discrete-time switched positive
linear systems with uncertainty are established. In the analyzed class,
arbitrary switching is permitted, whereas the uncertainty expressed via matrix
inequalities concerns both the initial state and system parameters. Positive
lower and positive upper interval switched observers are obtained. The proposed
observers are of (n - p) order, where n is the dimension of the state vector
and p is the rank of the output matrix, i.e., p-dimensional measurement
information. Moreover, as a special case, existence conditions and a design
procedure of reduced-order positive observers for uncertain positive linear
systems without switching are provided. The theoretical findings are
illustrated by two numerical examples for continuous- and discrete-time
systems.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 18:27:43 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 07:08:35 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Otsuka",
"Naohisa",
""
],
[
"Kakehi",
"Daiki",
""
],
[
"Ignaciuk",
"Przemysław",
""
]
] |
2505.06625 | Tianhao Cai | Tianhao Cai, Liang Wang, Limin Xiao, Meng Han, Zeyu Wang, Lin Sun and
Xiaojian Liao | CaMDN: Enhancing Cache Efficiency for Multi-tenant DNNs on Integrated
NPUs | 7 pages, 9 figures. This paper has been accepted to the 2025 Design
Automation Conference (DAC) | null | null | null | cs.AR cs.AI cs.OS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid development of DNN applications, multi-tenant execution, where
multiple DNNs are co-located on a single SoC, is becoming a prevailing trend.
Although many methods are proposed in prior works to improve multi-tenant
performance, the impact of shared cache is not well studied. This paper
proposes CaMDN, an architecture-scheduling co-design to enhance cache
efficiency for multi-tenant DNNs on integrated NPUs. Specifically, a
lightweight architecture is proposed to support model-exclusive, NPU-controlled
regions inside shared cache to eliminate unexpected cache contention. Moreover,
a cache scheduling method is proposed to improve shared cache utilization. In
particular, it includes a cache-aware mapping method for adaptability to the
varying available cache capacity and a dynamic allocation algorithm to adjust
the usage among co-located DNNs at runtime. Compared to prior works, CaMDN
reduces the memory access by 33.4% on average and achieves a model speedup of
up to 2.56$\times$ (1.88$\times$ on average).
| [
{
"version": "v1",
"created": "Sat, 10 May 2025 12:16:50 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Cai",
"Tianhao",
""
],
[
"Wang",
"Liang",
""
],
[
"Xiao",
"Limin",
""
],
[
"Han",
"Meng",
""
],
[
"Wang",
"Zeyu",
""
],
[
"Sun",
"Lin",
""
],
[
"Liao",
"Xiaojian",
""
]
] |
2505.06738 | Zibo Gao | Zibo Gao, Junjie Hu, Feng Guo, Yixin Zhang, Yinglong Han, Siyuan Liu,
Haiyang Li and Zhiqiang Lv | I Know What You Said: Unveiling Hardware Cache Side-Channels in Local
Large Language Model Inference | Submitted for review in January 22, 2025 | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) that can be deployed locally have recently
gained popularity for privacy-sensitive tasks, with companies such as Meta,
Google, and Intel playing significant roles in their development. However, the
security of local LLMs through the lens of hardware cache side-channels remains
unexplored. In this paper, we unveil novel side-channel vulnerabilities in
local LLM inference: token value and token position leakage, which can expose
both the victim's input and output text, thereby compromising user privacy.
Specifically, we found that adversaries can infer the token values from the
cache access patterns of the token embedding operation, and deduce the token
positions from the timing of autoregressive decoding phases. To demonstrate the
potential of these leaks, we design a novel eavesdropping attack framework
targeting both open-source and proprietary LLM inference systems. The attack
framework does not directly interact with the victim's LLM and can be executed
without privilege.
We evaluate the attack on a range of practical local LLM deployments (e.g.,
Llama, Falcon, and Gemma), and the results show that our attack achieves
promising accuracy. The restored output and input text have an average edit
distance of 5.2% and 17.3% to the ground truth, respectively. Furthermore, the
reconstructed texts achieve average cosine similarity scores of 98.7% (input)
and 98.0% (output).
| [
{
"version": "v1",
"created": "Sat, 10 May 2025 19:06:37 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 16:04:57 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Gao",
"Zibo",
""
],
[
"Hu",
"Junjie",
""
],
[
"Guo",
"Feng",
""
],
[
"Zhang",
"Yixin",
""
],
[
"Han",
"Yinglong",
""
],
[
"Liu",
"Siyuan",
""
],
[
"Li",
"Haiyang",
""
],
[
"Lv",
"Zhiqiang",
""
]
] |
2505.06795 | Abhijit Gupta PhD | Abhijit Gupta | Decoding Futures Price Dynamics: A Regularized Sparse Autoencoder for
Interpretable Multi-Horizon Forecasting and Factor Discovery | null | null | null | null | cs.LG cs.AI cs.CE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Commodity price volatility creates economic challenges, necessitating
accurate multi-horizon forecasting. Predicting prices for commodities like
copper and crude oil is complicated by diverse interacting factors
(macroeconomic, supply/demand, geopolitical, etc.). Current models often lack
transparency, limiting strategic use. This paper presents a Regularized Sparse
Autoencoder (RSAE), a deep learning framework for simultaneous multi-horizon
commodity price prediction and discovery of interpretable latent market
drivers. The RSAE forecasts prices at multiple horizons (e.g., 1-day, 1-week,
1-month) using multivariate time series. Crucially, L1 regularization
($\|\mathbf{z}\|_1$) on its latent vector $\mathbf{z}$ enforces sparsity,
promoting parsimonious explanations of market dynamics through learned factors
representing underlying drivers (e.g., demand, supply shocks). Drawing from
energy-based models and sparse coding, the RSAE optimizes predictive accuracy
while learning sparse representations. Evaluated on historical Copper and Crude
Oil data with numerous indicators, our findings indicate the RSAE offers
competitive multi-horizon forecasting accuracy and data-driven insights into
price dynamics via its interpretable latent space, a key advantage over
traditional black-box approaches.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 00:21:53 GMT"
},
{
"version": "v2",
"created": "Tue, 13 May 2025 06:44:43 GMT"
},
{
"version": "v3",
"created": "Wed, 14 May 2025 17:49:51 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Gupta",
"Abhijit",
""
]
] |
2505.06977 | WenJu Sun | Wenju Sun and Qingyong Li and Yangli-ao Geng and Boyang Li | CAT Merging: A Training-Free Approach for Resolving Conflicts in Model
Merging | null | null | null | null | cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-task model merging offers a promising paradigm for integrating multiple
expert models into a unified model without additional training. Existing
state-of-the-art techniques, such as Task Arithmetic and its variants, merge
models by accumulating task vectors -- the parameter differences between
pretrained and finetuned models. However, task vector accumulation is often
hindered by knowledge conflicts, leading to performance degradation. To address
this challenge, we propose Conflict-Aware Task Merging (CAT Merging), a novel
training-free framework that selectively trims conflict-prone components from
the task vectors. CAT Merging introduces several parameter-specific strategies,
including projection for linear weights and masking for scaling and shifting
parameters in normalization layers. Extensive experiments on vision, language,
and vision-language tasks demonstrate that CAT Merging effectively suppresses
knowledge conflicts, achieving average accuracy improvements of up to 2.5%
(ViT-B/32) and 2.0% (ViT-L/14) over state-of-the-art methods.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 13:24:09 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 14:11:52 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Sun",
"Wenju",
""
],
[
"Li",
"Qingyong",
""
],
[
"Geng",
"Yangli-ao",
""
],
[
"Li",
"Boyang",
""
]
] |
2505.07051 | Shannon Starr | Shannon Starr | Some Observations about the "Generalized Abundancy Index" | added 1 reference | null | null | null | math.CO math.NT math.PR | http://creativecommons.org/licenses/by/4.0/ | Let $\mathcal{A}(\ell,n) \subset S_n^{\ell}$ denote the set of all
$\ell$-tuples $(\pi_1,\dots,\pi_{\ell})$, for $\pi_1,\dots,\pi_{\ell} \in S_n$
satisfying: $\forall i<j$ we have $\pi_i\pi_j=\pi_j\pi_i$. Considering the
action of $S_n$ on $[n]=\{1,\dots,n\}$, let $\kappa(\pi_1,\dots,\pi_{\ell})$ be
equal to the number of orbits of the action of the subgroup $\langle
\pi_1,\dots,\pi_{\ell} \rangle \subset S_n$. There has been interest in the
study of the combinatorial numbers $A(\ell,n,k)$ equal to the cardinalities
$|\{(\pi_1,\dots,\pi_{\ell}) \in \mathcal{A}(\ell,n)\, :\,
\kappa(\pi_1,\dots\pi_{\ell})=k\}|$. If one defines
$B(\ell,n)=A(\ell,n,1)/(n-1)!$, then it is known that $B(\ell,n) =
\sum_{(f_1,\dots,f_{\ell}) \in \mathbb{N}^{\ell}} \mathbf{1}_{\{n\}}(f_1\cdots
f_{\ell}) \prod_{r=1}^{\ell-1} f_r^{\ell-r}$. A special case, $\ell=2$, is
$B(2,n) = \sum_{d|n} d = \sigma_1(n)$ the sum-of-divisors function. Then
$A(2,n,1)/n!=B(2,n)/n$ is called the abundancy index: $\sigma_1(n)/n$. We call
$B(\ell,n) n^{-\ell+1}$ the ``generalized abundancy index.'' Building on work
of Abdesselam, using the probability model, we prove that $\lim_{N \to \infty}
N^{-1} \sum_{n=1}^{N} B(\ell,n) n^{-\ell+1}$ equals $\zeta(2)\cdots
\zeta(\ell)$. Motivated by this we state a more precise conjecture for the
asymptotics of $-\zeta(2) + N^{-1}\sum_{n=1}^{N} (B(2,n)/n)$.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 16:49:28 GMT"
},
{
"version": "v2",
"created": "Tue, 13 May 2025 22:23:07 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Starr",
"Shannon",
""
]
] |
2505.07086 | Pranam Chatterjee | Tong Chen, Yinuo Zhang, Sophia Tang, Pranam Chatterjee | Multi-Objective-Guided Discrete Flow Matching for Controllable
Biological Sequence Design | null | null | null | null | cs.LG q-bio.BM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Designing biological sequences that satisfy multiple, often conflicting,
functional and biophysical criteria remains a central challenge in biomolecule
engineering. While discrete flow matching models have recently shown promise
for efficient sampling in high-dimensional sequence spaces, existing approaches
address only single objectives or require continuous embeddings that can
distort discrete distributions. We present Multi-Objective-Guided Discrete Flow
Matching (MOG-DFM), a general framework to steer any pretrained discrete flow
matching generator toward Pareto-efficient trade-offs across multiple scalar
objectives. At each sampling step, MOG-DFM computes a hybrid rank-directional
score for candidate transitions and applies an adaptive hypercone filter to
enforce consistent multi-objective progression. We also trained two
unconditional discrete flow matching models, PepDFM for diverse peptide
generation and EnhancerDFM for functional enhancer DNA generation, as base
generation models for MOG-DFM. We demonstrate MOG-DFM's effectiveness in
generating peptide binders optimized across five properties (hemolysis,
non-fouling, solubility, half-life, and binding affinity), and in designing DNA
sequences with specific enhancer classes and DNA shapes. In total, MOG-DFM
proves to be a powerful tool for multi-property-guided biomolecule sequence
design.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 18:17:44 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 16:19:40 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Chen",
"Tong",
""
],
[
"Zhang",
"Yinuo",
""
],
[
"Tang",
"Sophia",
""
],
[
"Chatterjee",
"Pranam",
""
]
] |
2505.07089 | Yuanliang Li | Hanzheng Dai and Yuanliang Li and Zhibo Zhang and Jun Yan | RefPentester: A Knowledge-Informed Self-Reflective Penetration Testing
Framework Based on Large Language Models | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Automated penetration testing (AutoPT) powered by large language models
(LLMs) has gained attention for its ability to automate ethical hacking
processes and identify vulnerabilities in target systems by leveraging the
intrinsic knowledge of LLMs. However, existing LLM-based AutoPT frameworks
often underperform compared to human experts in challenging tasks for several
reasons: the imbalanced knowledge used in LLM training, short-sighted planning
in the planning process, and hallucinations during command generation. In
addition, the penetration testing (PT) process, with its trial-and-error
nature, is limited by existing frameworks that lack mechanisms to learn from
previous failed operations, restricting adaptive improvement of PT strategies.
To address these limitations, we propose a knowledge-informed self-reflective
PT framework powered by LLMs, called RefPentester, which is an AutoPT framework
designed to assist human operators in identifying the current stage of the PT
process, selecting appropriate tactic and technique for the stage, choosing
suggested action, providing step-by-step operational guidance, and learning
from previous failed operations. We also modeled the PT process as a
seven-state Stage Machine to integrate the proposed framework effectively. The
evaluation shows that RefPentester can successfully reveal credentials on Hack
The Box's Sau machine, outperforming the baseline GPT-4o model by 16.7%. Across
PT stages, RefPentester also demonstrates superior success rates on PT stage
transitions.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 18:38:00 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 00:44:05 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Dai",
"Hanzheng",
""
],
[
"Li",
"Yuanliang",
""
],
[
"Zhang",
"Zhibo",
""
],
[
"Yan",
"Jun",
""
]
] |
2505.07356 | Smarajit Karmakar Dr. | Rashmi Priya and Smarajit Karmakar | Inverse Bauschinger Effect in Active Ultrastable Glasses | null | null | null | null | cond-mat.soft cond-mat.stat-mech physics.bio-ph | http://creativecommons.org/licenses/by/4.0/ | Memory effects in amorphous materials have been widely studied because of
their possible widespread future applications. We show here that ultrastable
glasses can exhibit a transient reversible memory effect when subjected to both
a local driving force via Run-and-tumble active particles and global shear. We
investigate the system's response across different yielding regimes by
selectively switching the shear direction at different strains. We analyze how
changes in shear direction influence yielding, post-yield behavior, and
structural evolution in active amorphous solids. Our model active system
exhibits an enhanced anisotropic response, displaying both conventional and
inverse Bauschinger effects, depending on the deformation history. The results
indicate that activity-induced shear band networks create structural memory,
enabling the system to heal upon shear reversal due to the transient nature of
this phenomenon. Additionally, we observe that shear softening under cyclic
loading produces an irreversible, stable, and less branched network structure
with increasing cycles. These findings provide novel insights into how activity
and shear collectively contribute to mechanical response, including memory
formation in ultrastable disordered systems.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 08:48:14 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 04:07:35 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Priya",
"Rashmi",
""
],
[
"Karmakar",
"Smarajit",
""
]
] |
2505.07363 | Serge Massar | Serge Massar | Equilibrium Propagation for Learning in Lagrangian Dynamical Systems | 8 pages, 1 figure | null | null | null | nlin.CD cs.LG physics.data-an | http://creativecommons.org/licenses/by/4.0/ | We propose a method for training dynamical systems governed by Lagrangian
mechanics using Equilibrium Propagation. Our approach extends Equilibrium
Propagation -- initially developed for energy-based models -- to dynamical
trajectories by leveraging the principle of action extremization. Training is
achieved by gently nudging trajectories toward desired targets and measuring
how the variables conjugate to the parameters to be trained respond. This
method is particularly suited to systems with periodic boundary conditions or
fixed initial and final states, enabling efficient parameter updates without
requiring explicit backpropagation through time. In the case of periodic
boundary conditions, this approach yields the semiclassical limit of Quantum
Equilibrium Propagation. Applications to systems with dissipation are also
discussed.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 08:59:17 GMT"
},
{
"version": "v2",
"created": "Tue, 13 May 2025 07:06:52 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Massar",
"Serge",
""
]
] |
2505.07392 | Spencer Bryngelson | Benjamin Wilfong and Anand Radhakrishnan and Henry Le Berre and
Nikolaos Tselepidis and Benedikt Dorschner and Reuben Budiardja and Brian
Cornille and Stephen Abbott and Florian Sch\"afer and Spencer H. Bryngelson | Simulating many-engine spacecraft: Exceeding 100 trillion grid points
via information geometric regularization and the MFC flow solver | 10 pages, 7 figures | null | null | null | physics.comp-ph cs.CE physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | This work proposes a method and optimized implementation for exascale
simulations of high-speed compressible fluid flows, enabling the simulation of
multi-engine rocket craft at an unprecedented scale. We significantly improve
upon the state-of-the-art in terms of computational cost and memory footprint
through a carefully crafted implementation of the recently proposed information
geometric regularization, which eliminates the need for numerical shock
capturing. Unified addressing on tightly coupled CPU--GPU platforms increases
the total problem size with negligible performance hit. Despite linear stencil
algorithms being memory-bound, we achieve wall clock times that are four times
faster than optimized baseline numerics. This enables the execution of CFD
simulations at more than 100 trillion grid points, surpassing the largest
state-of-the-art publicly available simulations by an order of magnitude. Ideal
weak scaling is demonstrated on OLCF Frontier and CSCS Alps using the full
system, entailing 37.8K AMD MI250X GPUs (Frontier) or 9.2K NVIDIA GH200
superchips (Alps).
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 09:35:38 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 08:58:41 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wilfong",
"Benjamin",
""
],
[
"Radhakrishnan",
"Anand",
""
],
[
"Berre",
"Henry Le",
""
],
[
"Tselepidis",
"Nikolaos",
""
],
[
"Dorschner",
"Benedikt",
""
],
[
"Budiardja",
"Reuben",
""
],
[
"Cornille",
"Brian",
""
],
[
"Abbott",
"Stephen",
""
],
[
"Schäfer",
"Florian",
""
],
[
"Bryngelson",
"Spencer H.",
""
]
] |
2505.07407 | Asif Manzoor Hasan | Asif Manzoor Hasan, Pedro Costa, Johan Larsson and Rene Pecnik | Scaling of wall pressure and the peak of streamwise turbulence intensity
in compressible wall flows | 10 pages, 2 figures | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | This paper develops scaling laws for wall-pressure root-mean-square (r.m.s.)
and the peak of streamwise turbulence intensity, accounting for both
variable-property and intrinsic compressibility effects -- those associated
with changes in fluid volume due to pressure variations. To develop such
scaling laws, we express the target quantities as an expansion series in powers
of an appropriately defined Mach number. The leading-order term is represented
using the scaling relations developed for incompressible flows, but with an
effective Reynolds number. Higher-order terms capture intrinsic compressibility
effects and are modeled as constant coefficients, calibrated using flow cases
specifically designed to isolate these effects. The resulting scaling relations
are shown to be accurate for a wide range of turbulent channel flows and
boundary layers.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 09:59:43 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 00:17:32 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Hasan",
"Asif Manzoor",
""
],
[
"Costa",
"Pedro",
""
],
[
"Larsson",
"Johan",
""
],
[
"Pecnik",
"Rene",
""
]
] |
2505.07412 | Tabish Qureshi | Rashi Adhikari, Tabish Qureshi | Characterizing the Generalized Einstein-Podolsky-Rosen State | 5 pages. A new section added on experimental measurement of
entanglement | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In their seminal paper, Einstein Podolsky and Rosen (EPR) had introduced a
momentum entangled state for two particles. That state, referred to as the EPR
state, has been widely used in studies on entangled particles with continuous
degrees of freedom. Later that state was generalized to a form that allows
varying degree of entanglement, known as the generalized EPR state. In a
suitable limit it reduces to the EPR state. The generalized EPR state is
theoretically analyzed here and its entanglement quantified in terms of a
recently introduced generalized entanglement measure. This state can also be
applied to entangled photons produced from spontaneous parametric down
conversion (SPDC). The present analysis is then used in quantifying the
entanglement of photons produced from the SPDC process, in terms of certain
experimental parameters. A comparison is also made with the Schmidt number,
which is normally used as an entanglement measure in such situations. A
procedure for experimentally determining the entanglement of SPDC photons has
also been described.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 10:07:36 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 17:22:02 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Adhikari",
"Rashi",
""
],
[
"Qureshi",
"Tabish",
""
]
] |
2505.07434 | Giovanni Viglietta | Giovanni Viglietta and Yasuyuki Kachi | Efficient Lifting of Discrete Logarithms Modulo Prime Powers | 15 pages | null | null | null | math.NT cs.DM | http://creativecommons.org/licenses/by/4.0/ | We present a deterministic algorithm that, given a prime $p$ and a solution
$x \in \mathbb Z$ to the discrete logarithm problem $a^x \equiv b \pmod p$ with
$p\nmid a$, efficiently lifts it to a solution modulo $p^k$, i.e., $a^x \equiv
b \pmod {p^k}$, for any fixed $k \geq 1$.
The algorithm performs $k(\lceil \log_2 p\rceil +2)+O(\log p)$
multiplications modulo $p^k$ in the worst case, improving upon prior lifting
methods by at least a factor of 8.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 10:52:19 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 15:55:10 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Viglietta",
"Giovanni",
""
],
[
"Kachi",
"Yasuyuki",
""
]
] |
2505.07574 | Vinod P | Matteo Brosolo, Asmitha K. A., Rafidha Rehiman K. A., Muhammed Shafi
K. P., Serena Nicolazzo, Antonino Nocera, Vinod P | Security through the Eyes of AI: How Visualization is Shaping Malware
Detection | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Malware, a persistent cybersecurity threat, increasingly targets
interconnected digital systems such as desktop, mobile, and IoT platforms
through sophisticated attack vectors. By exploiting these vulnerabilities,
attackers compromise the integrity and resilience of modern digital ecosystems.
To address this risk, security experts actively employ Machine Learning or Deep
Learning-based strategies, integrating static, dynamic, or hybrid approaches to
categorize malware instances. Despite their advantages, these methods have
inherent drawbacks and malware variants persistently evolve with increased
sophistication, necessitating advancements in detection strategies.
Visualization-based techniques are emerging as scalable and interpretable
solutions for detecting and understanding malicious behaviors across diverse
platforms including desktop, mobile, IoT, and distributed systems as well as
through analysis of network packet capture files. In this comprehensive survey
of more than 100 high-quality research articles, we evaluate existing
visualization-based approaches applied to malware detection and classification.
As a first contribution, we propose a new all-encompassing framework to study
the landscape of visualization-based malware detection techniques. Within this
framework, we systematically analyze state-of-the-art approaches across the
critical stages of the malware detection pipeline. By analyzing not only the
single techniques but also how they are combined to produce the final solution,
we shed light on the main challenges in visualization-based approaches and
provide insights into the advancements and potential future directions in this
critical field.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 13:53:56 GMT"
},
{
"version": "v2",
"created": "Tue, 13 May 2025 23:35:12 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Brosolo",
"Matteo",
""
],
[
"A.",
"Asmitha K.",
""
],
[
"A.",
"Rafidha Rehiman K.",
""
],
[
"P.",
"Muhammed Shafi K.",
""
],
[
"Nicolazzo",
"Serena",
""
],
[
"Nocera",
"Antonino",
""
],
[
"P",
"Vinod",
""
]
] |
2505.07634 | Jian Liu | Jian Liu, Xiongtao Shi, Thai Duy Nguyen, Haitian Zhang, Tianxiang
Zhang, Wei Sun, Yanjie Li, Athanasios V. Vasilakos, Giovanni Iacca, Arshad
Ali Khan, Arvind Kumar, Jae Won Cho, Ajmal Mian, Lihua Xie, Erik Cambria, Lin
Wang | Neural Brain: A Neuroscience-inspired Framework for Embodied Agents | 51 pages, 17 figures, 9 tables | null | null | null | cs.RO cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid evolution of artificial intelligence (AI) has shifted from static,
data-driven models to dynamic systems capable of perceiving and interacting
with real-world environments. Despite advancements in pattern recognition and
symbolic reasoning, current AI systems, such as large language models, remain
disembodied, unable to physically engage with the world. This limitation has
driven the rise of embodied AI, where autonomous agents, such as humanoid
robots, must navigate and manipulate unstructured environments with human-like
adaptability. At the core of this challenge lies the concept of Neural Brain, a
central intelligence system designed to drive embodied agents with human-like
adaptability. A Neural Brain must seamlessly integrate multimodal sensing and
perception with cognitive capabilities. Achieving this also requires an
adaptive memory system and energy-efficient hardware-software co-design,
enabling real-time action in dynamic environments. This paper introduces a
unified framework for the Neural Brain of embodied agents, addressing two
fundamental challenges: (1) defining the core components of Neural Brain and
(2) bridging the gap between static AI models and the dynamic adaptability
required for real-world deployment. To this end, we propose a biologically
inspired architecture that integrates multimodal active sensing,
perception-cognition-action function, neuroplasticity-based memory storage and
updating, and neuromorphic hardware/software optimization. Furthermore, we also
review the latest research on embodied agents across these four aspects and
analyze the gap between current AI systems and human intelligence. By
synthesizing insights from neuroscience, we outline a roadmap towards the
development of generalizable, autonomous agents capable of human-level
intelligence in real-world scenarios.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 15:05:34 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 12:56:45 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Liu",
"Jian",
""
],
[
"Shi",
"Xiongtao",
""
],
[
"Nguyen",
"Thai Duy",
""
],
[
"Zhang",
"Haitian",
""
],
[
"Zhang",
"Tianxiang",
""
],
[
"Sun",
"Wei",
""
],
[
"Li",
"Yanjie",
""
],
[
"Vasilakos",
"Athanasios V.",
""
],
[
"Iacca",
"Giovanni",
""
],
[
"Khan",
"Arshad Ali",
""
],
[
"Kumar",
"Arvind",
""
],
[
"Cho",
"Jae Won",
""
],
[
"Mian",
"Ajmal",
""
],
[
"Xie",
"Lihua",
""
],
[
"Cambria",
"Erik",
""
],
[
"Wang",
"Lin",
""
]
] |
2505.07773 | Xinji Mai | Xinji Mai, Haotian Xu, Xing W, Weinong Wang, Yingying Zhang, Wenqiang
Zhang | Agent RL Scaling Law: Agent RL with Spontaneous Code Execution for
Mathematical Problem Solving | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) often struggle with mathematical reasoning tasks
requiring precise, verifiable computation. While Reinforcement Learning (RL)
from outcome-based rewards enhances text-based reasoning, understanding how
agents autonomously learn to leverage external tools like code execution
remains crucial. We investigate RL from outcome-based rewards for
Tool-Integrated Reasoning, ZeroTIR, training base LLMs to spontaneously
generate and execute Python code for mathematical problems without supervised
tool-use examples. Our central contribution is we demonstrate that as RL
training progresses, key metrics scale predictably. Specifically, we observe
strong positive correlations where increased training steps lead to increases
in the spontaneous code execution frequency, the average response length, and,
critically, the final task accuracy. This suggests a quantifiable relationship
between computational effort invested in training and the emergence of
effective, tool-augmented reasoning strategies. We implement a robust framework
featuring a decoupled code execution environment and validate our findings
across standard RL algorithms and frameworks. Experiments show ZeroTIR
significantly surpasses non-tool ZeroRL baselines on challenging math
benchmarks. Our findings provide a foundational understanding of how autonomous
tool use is acquired and scales within Agent RL, offering a reproducible
benchmark for future studies. Code is released at
\href{https://github.com/yyht/openrlhf_async_pipline}{https://github.com/yyht/openrlhf\_async\_pipline}.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 17:23:34 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 04:15:06 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Mai",
"Xinji",
""
],
[
"Xu",
"Haotian",
""
],
[
"W",
"Xing",
""
],
[
"Wang",
"Weinong",
""
],
[
"Zhang",
"Yingying",
""
],
[
"Zhang",
"Wenqiang",
""
]
] |
2505.07835 | Alex C. Y. Wong | Alex C. Y. Wong, Duncan McFarlane, C. Ellarby, M. Lee, and M. Kuok | Intelligent Product 3.0: Decentralised AI Agents and Web3 Intelligence
Standards | 18 pages, 1 Figure, 3 Tables; Corrected typo in Section 3.4 heading | null | null | null | cs.NI cs.AI cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Twenty-five years ago, the specification of the Intelligent Product was
established, envisaging real-time connectivity that not only enables products
to gather accurate data about themselves but also allows them to assess and
influence their own destiny. Early work by the Auto-ID project focused on
creating a single, open-standard repository for storing and retrieving product
information, laying a foundation for scalable connectivity. A decade later, the
approach was revisited in light of low-cost RFID systems that promised a
low-cost link between physical goods and networked information environments.
Since then, advances in blockchain, Web3, and artificial intelligence have
introduced unprecedented levels of resilience, consensus, and autonomy. By
leveraging decentralised identity, blockchain-based product information and
history, and intelligent AI-to-AI collaboration, this paper examines these
developments and outlines a new specification for the Intelligent Product 3.0,
illustrating how decentralised and AI-driven capabilities facilitate seamless
interaction between physical AI and everyday products.
| [
{
"version": "v1",
"created": "Fri, 2 May 2025 22:01:17 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 11:10:49 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wong",
"Alex C. Y.",
""
],
[
"McFarlane",
"Duncan",
""
],
[
"Ellarby",
"C.",
""
],
[
"Lee",
"M.",
""
],
[
"Kuok",
"M.",
""
]
] |
2505.07887 | Songyin Wu | Songyin Wu, Zhaoyang Lv, Yufeng Zhu, Duncan Frost, Zhengqin Li,
Ling-Qi Yan, Carl Ren, Richard Newcombe, Zhao Dong | Monocular Online Reconstruction with Enhanced Detail Preservation | Accepted to SIGGRAPH 2025 (Conference Track). Project page:
https://poiw.github.io/MODP | null | null | null | cs.GR cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We propose an online 3D Gaussian-based dense mapping framework for
photorealistic details reconstruction from a monocular image stream. Our
approach addresses two key challenges in monocular online reconstruction:
distributing Gaussians without relying on depth maps and ensuring both local
and global consistency in the reconstructed maps. To achieve this, we introduce
two key modules: the Hierarchical Gaussian Management Module for effective
Gaussian distribution and the Global Consistency Optimization Module for
maintaining alignment and coherence at all scales. In addition, we present the
Multi-level Occupancy Hash Voxels (MOHV), a structure that regularizes
Gaussians for capturing details across multiple levels of granularity. MOHV
ensures accurate reconstruction of both fine and coarse geometries and
textures, preserving intricate details while maintaining overall structural
integrity. Compared to state-of-the-art RGB-only and even RGB-D methods, our
framework achieves superior reconstruction quality with high computational
efficiency. Moreover, it integrates seamlessly with various tracking systems,
ensuring generality and scalability.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 01:49:09 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 03:15:06 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wu",
"Songyin",
""
],
[
"Lv",
"Zhaoyang",
""
],
[
"Zhu",
"Yufeng",
""
],
[
"Frost",
"Duncan",
""
],
[
"Li",
"Zhengqin",
""
],
[
"Yan",
"Ling-Qi",
""
],
[
"Ren",
"Carl",
""
],
[
"Newcombe",
"Richard",
""
],
[
"Dong",
"Zhao",
""
]
] |
2505.07890 | Kutay Ert\"urk | Kutay Ert\"urk, Furkan Alt{\i}n{\i}\c{s}{\i}k, \.Irem Sar{\i}alt{\i}n,
\"Omer Nezih Gerek | TSLFormer: A Lightweight Transformer Model for Turkish Sign Language
Recognition Using Skeletal Landmarks | null | null | null | null | cs.CL eess.IV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This study presents TSLFormer, a light and robust word-level Turkish Sign
Language (TSL) recognition model that treats sign gestures as ordered,
string-like language. Instead of using raw RGB or depth videos, our method only
works with 3D joint positions - articulation points - extracted using Google's
Mediapipe library, which focuses on the hand and torso skeletal locations. This
creates efficient input dimensionality reduction while preserving important
semantic gesture information.
Our approach revisits sign language recognition as sequence-to-sequence
translation, inspired by the linguistic nature of sign languages and the
success of transformers in natural language processing. Since TSLFormer uses
the self-attention mechanism, it effectively captures temporal co-occurrence
within gesture sequences and highlights meaningful motion patterns as words
unfold.
Evaluated on the AUTSL dataset with over 36,000 samples and 227 different
words, TSLFormer achieves competitive performance with minimal computational
cost. These results show that joint-based input is sufficient for enabling
real-time, mobile, and assistive communication systems for hearing-impaired
individuals.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 14:30:56 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 16:43:25 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Ertürk",
"Kutay",
""
],
[
"Altınışık",
"Furkan",
""
],
[
"Sarıaltın",
"İrem",
""
],
[
"Gerek",
"Ömer Nezih",
""
]
] |
2505.07961 | Xuechen Zhang | Xuechen Zhang, Zijian Huang, Chenshun Ni, Ziyang Xiong, Jiasi Chen,
Samet Oymak | Making Small Language Models Efficient Reasoners: Intervention,
Supervision, Reinforcement | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent research enhances language model reasoning by scaling test-time
compute via longer chain-of-thought traces. This often improves accuracy but
also introduces redundancy and high computational cost, especially for small
language models distilled with supervised fine-tuning (SFT). In this work, we
propose new algorithms to improve token-efficient reasoning with small-scale
models by effectively trading off accuracy and computation. We first show that
the post-SFT model fails to determine the optimal stopping point of the
reasoning process, resulting in verbose and repetitive outputs. Verbosity also
significantly varies across wrong vs correct responses. To address these
issues, we propose two solutions: (1) Temperature scaling (TS) to control the
stopping point for the thinking phase and thereby trace length, and (2) TLDR: a
length-regularized reinforcement learning method based on GRPO that facilitates
multi-level trace length control (e.g. short, medium, long reasoning).
Experiments on four reasoning benchmarks, MATH500, AMC, AIME24 and
OlympiadBench, demonstrate that TS is highly effective compared to s1's budget
forcing approach and TLDR significantly improves token efficiency by about 50%
with minimal to no accuracy loss over the SFT baseline. Moreover, TLDR also
facilitates flexible control over the response length, offering a practical and
effective solution for token-efficient reasoning in small models. Ultimately,
our work reveals the importance of stopping time control, highlights
shortcomings of pure SFT, and provides effective algorithmic recipes.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 18:04:39 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 01:42:08 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Zhang",
"Xuechen",
""
],
[
"Huang",
"Zijian",
""
],
[
"Ni",
"Chenshun",
""
],
[
"Xiong",
"Ziyang",
""
],
[
"Chen",
"Jiasi",
""
],
[
"Oymak",
"Samet",
""
]
] |
2505.08020 | Daniel Cranston | Stijn Cambie, Wouter Cames van Batenburg, Daniel W. Cranston, Jan van
den Heuvel, Ross J. Kang | Reconfiguration of List Colourings | 27 pages, 4 figures | null | null | null | math.CO cs.DM cs.DS | http://creativecommons.org/licenses/by/4.0/ | Given a proper (list) colouring of a graph $G$, a recolouring step changes
the colour at a single vertex to another colour (in its list) that is currently
unused on its neighbours, hence maintaining a proper colouring. Suppose that
each vertex $v$ has its own private list $L(v)$ of allowed colours such that
$|L(v)|\ge \mbox{deg}(v)+1$. We prove that if $G$ is connected and its maximum
degree $\Delta$ is at least $3$, then for any two proper $L$-colourings in
which at least one vertex can be recoloured, one can be transformed to the
other by a sequence of $O(|V(G)|^2)$ recolouring steps. We also show that
reducing the list-size of a single vertex $w$ to $\mbox{deg}(w)$ can lead to
situations where the space of proper $L$-colourings is `shattered'. Our results
can be interpreted as showing a sharp phase transition in the Glauber dynamics
of proper $L$-colourings of graphs. This constitutes a `local' strengthening
and generalisation of a result of Feghali, Johnson, and Paulusma, which
considered the situation where the lists are all identical to
$\{1,\ldots,\Delta+1\}$.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 19:41:48 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Cambie",
"Stijn",
""
],
[
"van Batenburg",
"Wouter Cames",
""
],
[
"Cranston",
"Daniel W.",
""
],
[
"Heuvel",
"Jan van den",
""
],
[
"Kang",
"Ross J.",
""
]
] |
2505.08037 | Yutong Liu | Yutong Liu, Feng Xiao, Ziyue Zhang, Yongbin Yu, Cheng Huang, Fan Gao,
Xiangxiang Wang, Ma-bao Ban, Manping Fan, Thupten Tsering, Cheng Huang,
Gadeng Luosang, Renzeng Duojie, Nyima Tashi | TiSpell: A Semi-Masked Methodology for Tibetan Spelling Correction
covering Multi-Level Error with Data Augmentation | 14 pages, 7 figures | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-level Tibetan spelling correction addresses errors at both the
character and syllable levels within a unified model. Existing methods focus
mainly on single-level correction and lack effective integration of both
levels. Moreover, there are no open-source datasets or augmentation methods
tailored for this task in Tibetan. To tackle this, we propose a data
augmentation approach using unlabeled text to generate multi-level corruptions,
and introduce TiSpell, a semi-masked model capable of correcting both
character- and syllable-level errors. Although syllable-level correction is
more challenging due to its reliance on global context, our semi-masked
strategy simplifies this process. We synthesize nine types of corruptions on
clean sentences to create a robust training set. Experiments on both simulated
and real-world data demonstrate that TiSpell, trained on our dataset,
outperforms baseline models and matches the performance of state-of-the-art
approaches, confirming its effectiveness.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 20:08:05 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 04:04:08 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Liu",
"Yutong",
""
],
[
"Xiao",
"Feng",
""
],
[
"Zhang",
"Ziyue",
""
],
[
"Yu",
"Yongbin",
""
],
[
"Huang",
"Cheng",
""
],
[
"Gao",
"Fan",
""
],
[
"Wang",
"Xiangxiang",
""
],
[
"Ban",
"Ma-bao",
""
],
[
"Fan",
"Manping",
""
],
[
"Tsering",
"Thupten",
""
],
[
"Huang",
"Cheng",
""
],
[
"Luosang",
"Gadeng",
""
],
[
"Duojie",
"Renzeng",
""
],
[
"Tashi",
"Nyima",
""
]
] |
2505.08048 | Sian Lee | Sian Lee, Haeseung Seo, Aiping Xiong, Dongwon Lee | Partisan Fact-Checkers' Warnings Can Effectively Correct Individuals'
Misbeliefs About Political Misinformation | To appear in International AAAI Conference on Web and Social Media
(ICWSM) 2025 | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Political misinformation, particularly harmful when it aligns with
individuals' preexisting beliefs and political ideologies, has become
widespread on social media platforms. In response, platforms like Facebook and
X introduced warning messages leveraging fact-checking results from third-party
fact-checkers to alert users against false content. However, concerns persist
about the effectiveness of these fact-checks, especially when fact-checkers are
perceived as politically biased. To address these concerns, this study presents
findings from an online human-subject experiment (N=216) investigating how the
political stances of fact-checkers influence their effectiveness in correcting
misbeliefs about political misinformation. Our findings demonstrate that
partisan fact-checkers can decrease the perceived accuracy of political
misinformation and correct misbeliefs without triggering backfire effects. This
correction is even more pronounced when the misinformation aligns with
individuals' political ideologies. Notably, while previous research suggests
that fact-checking warnings are less effective for conservatives than liberals,
our results suggest that explicitly labeled partisan fact-checkers, positioned
as political counterparts to conservatives, are particularly effective in
reducing conservatives' misbeliefs toward pro-liberal misinformation.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 20:34:50 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 15:19:43 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Lee",
"Sian",
""
],
[
"Seo",
"Haeseung",
""
],
[
"Xiong",
"Aiping",
""
],
[
"Lee",
"Dongwon",
""
]
] |
2505.08090 | Luc Ta | Luc Ta | Equivalences of racks, Legendrian racks, and symmetric racks | 13 pages, 1 figure, 1 table; comments welcome. Simplified the
notation in Ex. 2.16, fixed a typo in Prop. 5.9, and added clarification to
the proofs of Prop. 4.2 and Thm. 1.2 | null | null | null | math.GT math.GR math.QA | http://creativecommons.org/licenses/by/4.0/ | Racks and Legendrian racks are nonassociative algebraic structures based on
the framed and Legendrian Reidemeister moves, respectively. Motivated by the
classification problem for Legendrian knots, we construct an equivalence of
categories between racks and Legendrian racks (and, hence, GL-quandles). We
deduce equivalences between kink-involutory racks and Legendrian quandles,
involutory racks and Legendrian kei, and the respective pairs of full
subcategories whose objects are medial.
As applications, we classify objects in these categories up to order 8 and
classify several families of symmetric racks; these results are likely to be of
independent interest. In particular, the categories of kei with good
involutions, Legendrian kei, and involutory racks are all equivalent.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 21:52:32 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 14:41:15 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Ta",
"Luc",
""
]
] |
2505.08167 | Ruilin Liu | Ruilin Liu, Zhixiao Zhao, Jieqiong Li, Chang Liu, Dongbo Wang | Fusing Bidirectional Chains of Thought and Reward Mechanisms A Method
for Enhancing Question-Answering Capabilities of Large Language Models for
Chinese Intangible Cultural Heritage | 22 pages, 5 figures | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid development of large language models (LLMs) has provided
significant support and opportunities for the advancement of domain-specific
LLMs. However, fine-tuning these large models using Intangible Cultural
Heritage (ICH) data inevitably faces challenges such as bias, incorrect
knowledge inheritance, and catastrophic forgetting. To address these issues, we
propose a novel training method that integrates a bidirectional chains of
thought and a reward mechanism. This method is built upon ICH-Qwen, a large
language model specifically designed for the field of intangible cultural
heritage. The proposed method enables the model to not only perform forward
reasoning but also enhances the accuracy of the generated answers by utilizing
reverse questioning and reverse reasoning to activate the model's latent
knowledge. Additionally, a reward mechanism is introduced during training to
optimize the decision-making process. This mechanism improves the quality of
the model's outputs through structural and content evaluations with different
weighting schemes. We conduct comparative experiments on ICH-Qwen, with results
demonstrating that our method outperforms 0-shot, step-by-step reasoning,
knowledge distillation, and question augmentation methods in terms of accuracy,
Bleu-4, and Rouge-L scores on the question-answering task. Furthermore, the
paper highlights the effectiveness of combining the bidirectional chains of
thought and reward mechanism through ablation experiments. In addition, a
series of generalizability experiments are conducted, with results showing that
the proposed method yields improvements on various domain-specific datasets and
advanced models in areas such as Finance, Wikidata, and StrategyQA. This
demonstrates that the method is adaptable to multiple domains and provides a
valuable approach for model training in future applications across diverse
fields.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 02:05:25 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 01:35:33 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Liu",
"Ruilin",
""
],
[
"Zhao",
"Zhixiao",
""
],
[
"Li",
"Jieqiong",
""
],
[
"Liu",
"Chang",
""
],
[
"Wang",
"Dongbo",
""
]
] |
2505.08175 | Zachary Novack | Zachary Novack, Zach Evans, Zack Zukowski, Josiah Taylor, CJ Carr,
Julian Parker, Adnan Al-Sinan, Gian Marco Iodice, Julian McAuley, Taylor
Berg-Kirkpatrick, Jordi Pons | Fast Text-to-Audio Generation with Adversarial Post-Training | null | null | null | null | cs.SD cs.AI cs.LG cs.MM eess.AS | http://creativecommons.org/licenses/by/4.0/ | Text-to-audio systems, while increasingly performant, are slow at inference
time, thus making their latency unpractical for many creative applications. We
present Adversarial Relativistic-Contrastive (ARC) post-training, the first
adversarial acceleration algorithm for diffusion/flow models not based on
distillation. While past adversarial post-training methods have struggled to
compare against their expensive distillation counterparts, ARC post-training is
a simple procedure that (1) extends a recent relativistic adversarial
formulation to diffusion/flow post-training and (2) combines it with a novel
contrastive discriminator objective to encourage better prompt adherence. We
pair ARC post-training with a number optimizations to Stable Audio Open and
build a model capable of generating $\approx$12s of 44.1kHz stereo audio in
$\approx$75ms on an H100, and $\approx$7s on a mobile edge-device, the fastest
text-to-audio model to our knowledge.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 02:25:47 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 06:07:26 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Novack",
"Zachary",
""
],
[
"Evans",
"Zach",
""
],
[
"Zukowski",
"Zack",
""
],
[
"Taylor",
"Josiah",
""
],
[
"Carr",
"CJ",
""
],
[
"Parker",
"Julian",
""
],
[
"Al-Sinan",
"Adnan",
""
],
[
"Iodice",
"Gian Marco",
""
],
[
"McAuley",
"Julian",
""
],
[
"Berg-Kirkpatrick",
"Taylor",
""
],
[
"Pons",
"Jordi",
""
]
] |
2505.08221 | Dongsheng Sui | Dongsheng Sui, Cunhua Pan, Hong Ren, Jiahua Wan, Liuchang Zhuo, Jing
Jin, Qixing Wang and Jiangzhou Wang | Performance Analysis of Cooperative Integrated Sensing and
Communications for 6G Networks | null | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we aim to effectively characterize the performance of
cooperative integrated sensing and communication (ISAC) networks and to reveal
how performance metrics relate to network parameters. To this end, we introduce
a generalized stochastic geometry framework to model the cooperative ISAC
networks, which approximates the spatial randomness of the network deployment.
Based on this framework, we derive analytical expressions for key performance
metrics in both communication and sensing domains, with a particular focus on
communication coverage probability and radar information rate. The analytical
expressions derived explicitly highlight how performance metrics depend on
network parameters, thereby offering valuable insights into the deployment and
design of cooperative ISAC networks. In the end, we validate the theoretical
performance analysis through Monte Carlo simulation results. Our results
demonstrate that increasing the number of cooperative base stations (BSs)
significantly improves both metrics, while increasing the BS deployment density
has a limited impact on communication coverage probability but substantially
enhances the radar information rate. Additionally, increasing the number of
transmit antennas is effective when the total number of transmit antennas is
relatively small. The incremental performance gain reduces with the increase of
the number of transmit antennas, suggesting that indiscriminately increasing
antennas is not an efficient strategy to improve the performance of the system
in cooperative ISAC networks.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 04:37:56 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 01:44:34 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Sui",
"Dongsheng",
""
],
[
"Pan",
"Cunhua",
""
],
[
"Ren",
"Hong",
""
],
[
"Wan",
"Jiahua",
""
],
[
"Zhuo",
"Liuchang",
""
],
[
"Jin",
"Jing",
""
],
[
"Wang",
"Qixing",
""
],
[
"Wang",
"Jiangzhou",
""
]
] |
2505.08233 | Santhoshkumar Peddi | Santhoshkumar Peddi and Soham Bandyopadhyay and Debasis Samanta | G-MSGINet: A Grouped Multi-Scale Graph-Involution Network for
Contactless Fingerprint Recognition | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper presents G-MSGINet, a unified and efficient framework for robust
contactless fingerprint recognition that jointly performs minutiae localization
and identity embedding directly from raw input images. Existing approaches rely
on multi-branch architectures, orientation labels, or complex preprocessing
steps, which limit scalability and generalization across real-world acquisition
scenarios. In contrast, the proposed architecture introduces the GMSGI layer, a
novel computational module that integrates grouped pixel-level involution,
dynamic multi-scale kernel generation, and graph-based relational modelling
into a single processing unit. Stacked GMSGI layers progressively refine both
local minutiae-sensitive features and global topological representations
through end-to-end optimization. The architecture eliminates explicit
orientation supervision and adapts graph connectivity directly from learned
kernel descriptors, thereby capturing meaningful structural relationships among
fingerprint regions without fixed heuristics. Extensive experiments on three
benchmark datasets, namely PolyU, CFPose, and Benchmark 2D/3D, demonstrate that
G-MSGINet consistently achieves minutiae F1-scores in the range of
$0.83\pm0.02$ and Rank-1 identification accuracies between 97.0% and 99.1%,
while maintaining an Equal Error Rate (EER) as low as 0.5%. These results
correspond to improvements of up to 4.8% in F1-score and 1.4% in Rank-1
accuracy when compared to prior methods, using only 0.38 million parameters and
6.63 giga floating-point operations, which represents up to ten times fewer
parameters than competitive baselines. This highlights the scalability and
effectiveness of G-MSGINet in real-world contactless biometric recognition
scenarios.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 05:24:24 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 06:16:41 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Peddi",
"Santhoshkumar",
""
],
[
"Bandyopadhyay",
"Soham",
""
],
[
"Samanta",
"Debasis",
""
]
] |
2505.08254 | Zhuo-Cheng Xiao | Zhongyi Wang, Louis Tao, Zhuo-Cheng Xiao | Efficient, simulation-free estimators of firing rates with Markovian
surrogates | 9 pages, 5 figures | null | null | null | q-bio.NC math.PR | http://creativecommons.org/licenses/by/4.0/ | Spiking neural networks (SNNs) are powerful mathematical models that
integrate the biological details of neural systems, but their complexity often
makes them computationally expensive and analytically untractable. The firing
rate of an SNN is a crucial first-order statistic to characterize network
activity. However, estimating firing rates analytically from even simplified
SNN models is challenging due to 1) the intricate dependence between the
nonlinear network dynamics and parameters, and 2) the singularity and
irreversibility of spikes. In this Letter, we propose a class of
computationally efficient, simulation-free estimators of firing rates. This is
based on a hierarchy of Markovian approximations that reduces the complexity of
SNN dynamics. We show that while considering firing rates alone is insufficient
for accurate estimations of themselves, the information of spiking synchrony
dramatically improves the estimator's accuracy. This approach provides a
practical tool for brain modelers, directly mapping biological parameters to
firing rate.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 06:04:02 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 15:52:14 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wang",
"Zhongyi",
""
],
[
"Tao",
"Louis",
""
],
[
"Xiao",
"Zhuo-Cheng",
""
]
] |
2505.08301 | Sanghun Lee | Jihyeon Lee and Sanghun Lee | Modified Hawking mass and rigidity of three-manifolds with boundary | null | null | null | null | math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we prove a rigidity result for three-dimensional Riemannian
manifolds with boundary, under the assumption that a free boundary minimal
two-disk, which locally maximizes a modified Hawking mass, is embedded in a
$3$-dimensional Riemannian manifold with negative scalar curvature and mean
convex boundary. First, we establish area estimates for free boundary strictly
stable two-disks. Finally, we show that the $3$-dimensional Riemannian manifold
with boundary is locally isometric to the half anti-de Sitter-Schwarzschild
manifold.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 07:25:27 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Lee",
"Jihyeon",
""
],
[
"Lee",
"Sanghun",
""
]
] |
2505.08320 | YoonHyuk Choi | Yoonhyuk Choi, Chong-Kwon Kim | SpecSphere: Dual-Pass Spectral-Spatial Graph Neural Networks with
Certified Robustness | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We introduce SpecSphere, the first dual-pass spectral-spatial GNN that
certifies every prediction against both $\ell\_{0}$ edge flips and
$\ell\_{\infty}$ feature perturbations, adapts to the full
homophily-heterophily spectrum, and surpasses the expressive power of
1-Weisfeiler-Lehman while retaining linear-time complexity. Our model couples a
Chebyshev-polynomial spectral branch with an attention-gated spatial branch and
fuses their representations through a lightweight MLP trained in a
cooperative-adversarial min-max game. We further establish (i) a uniform
Chebyshev approximation theorem, (ii) minimax-optimal risk across the
homophily-heterophily spectrum, (iii) closed-form robustness certificates, and
(iv) universal approximation strictly beyond 1-WL. SpecSphere achieves
state-of-the-art node-classification accuracy and delivers tighter certified
robustness guarantees on real-world benchmarks. These results demonstrate that
high expressivity, heterophily adaptation, and provable robustness can coexist
within a single, scalable architecture.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 08:00:16 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 17:07:37 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Choi",
"Yoonhyuk",
""
],
[
"Kim",
"Chong-Kwon",
""
]
] |
2505.08401 | Tarun Dalal | Francesc Bars, Tarun Dalal | The modular automorphisms of quotient modular curves | null | null | null | null | math.NT math.AG math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We obtain the modular automorphism group of any quotient modular curve of
level $N$, with $4,9\nmid N$. In particular, we obtain some non-expected
automorphisms of order 3 that appear for the quotient modular curves when the
Atkin-Lehner involution $w_{25}$ belongs to the quotient modular group, such
automorphisms are not necessarily defined over $\mathbb{Q}$. As a consequence
of the results, we obtain the full automorphism group of the quotient modular
curve $X_0^*(N^2)$, for sufficiently large $N$.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 09:56:14 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Bars",
"Francesc",
""
],
[
"Dalal",
"Tarun",
""
]
] |
2505.08435 | Mehran Sarmadi | Mehran Sarmadi, Morteza Alikhani, Erfan Zinvandi, Zahra Pourbahman | Hakim: Farsi Text Embedding Model | null | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in text embedding have significantly improved natural
language understanding across many languages, yet Persian remains notably
underrepresented in large-scale embedding research. In this paper, we present
Hakim, a novel state-of-the-art Persian text embedding model that achieves a
8.5% performance improvement over existing approaches on the FaMTEB benchmark,
outperforming all previously developed Persian language models. As part of this
work, we introduce three new datasets - Corpesia, Pairsia-sup, and
Pairsia-unsup - to support supervised and unsupervised training scenarios.
Additionally, Hakim is designed for applications in chatbots and
retrieval-augmented generation (RAG) systems, particularly addressing retrieval
tasks that require incorporating message history within these systems. We also
propose a new baseline model built on the BERT architecture. Our language model
consistently achieves higher accuracy across various Persian NLP tasks, while
the RetroMAE-based model proves particularly effective for textual information
retrieval applications. Together, these contributions establish a new
foundation for advancing Persian language understanding.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 10:57:32 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 13:47:12 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Sarmadi",
"Mehran",
""
],
[
"Alikhani",
"Morteza",
""
],
[
"Zinvandi",
"Erfan",
""
],
[
"Pourbahman",
"Zahra",
""
]
] |
2505.08477 | Chenhui Xie | Runyu Lei, Chen-Hui Xie, Jiayi Liu, Zhong Liu, Xin Liu, Yu Gao, Sichun
Sun, Jinxing Zhang | Emerging axion detection in artificial magnetoelectric materials | 17 pages, 5 figures | null | null | null | hep-ph cond-mat.mes-hall cond-mat.mtrl-sci cond-mat.str-el hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The origin of dark matter is a fundamental problem in physics. Axions are
considered a key component of dark matter, characterized by very weak
Chern-Simons couplings to electromagnetism, gravity, and fermions. We propose a
novel symmetry-breaking detection mechanism in magnetoelectric materials,
allowing for a linear axionic coupling between magnetism and ferroelectric
polarization. We focus on strain gradient Sr2IrO4, where the breaking of
space-inversion symmetry results in an emergent polar phase and out-of-plane
magnetic moment, exhibiting a flexomagnetoelectric effect. In this material,
the linear P||M enables direct coupling between the external axion field and
the intrinsic axion-like field within the material. This mechanism amplifies
the weak electromagnetic signals induced by axions, paving the way for
pioneering axion detection. These signals can be detected by monitoring changes
in macroscopic physical quantities, such as the magnetoelectric and magnetic
responses. In contrast to conventional detection techniques, this mechanism
significantly enhances the sensitivity of the axion-electron and axion-photon
coupling, providing a novel platform for axion detection and advancing the
study of dark matter through the magnetoelectric effect.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 12:04:30 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Lei",
"Runyu",
""
],
[
"Xie",
"Chen-Hui",
""
],
[
"Liu",
"Jiayi",
""
],
[
"Liu",
"Zhong",
""
],
[
"Liu",
"Xin",
""
],
[
"Gao",
"Yu",
""
],
[
"Sun",
"Sichun",
""
],
[
"Zhang",
"Jinxing",
""
]
] |
2505.08526 | Wuzhe Xu | Wuzhe Xu, Yulong Lu, Sifan Wang and Tong-Rui Liu | Improving Data Fidelity via Diffusion Model-based Correction and
Super-Resolution | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a unified diffusion model-based correction and super-resolution
method to enhance the fidelity and resolution of diverse low-quality data
through a two-step pipeline. First, the correction step employs a novel
enhanced stochastic differential editing technique based on an imbalanced
perturbation and denoising process, ensuring robust and effective bias
correction at the low-resolution level. The robustness and effectiveness of
this approach are validated theoretically and experimentally. Next, the
super-resolution step leverages cascaded conditional diffusion models to
iteratively refine the corrected data to high-resolution. Numerical experiments
on three PDE problems and a climate dataset demonstrate that the proposed
method effectively enhances low-fidelity, low-resolution data by correcting
numerical errors and noise while simultaneously improving resolution to recover
fine-scale structures.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 12:59:07 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 13:32:25 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Xu",
"Wuzhe",
""
],
[
"Lu",
"Yulong",
""
],
[
"Wang",
"Sifan",
""
],
[
"Liu",
"Tong-Rui",
""
]
] |
2505.08527 | Zheang Huai | Zheang Huai, Hui Tang, Yi Li, Zhuangzhuang Chen, Xiaomeng Li | Leveraging Segment Anything Model for Source-Free Domain Adaptation via
Dual Feature Guided Auto-Prompting | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Source-free domain adaptation (SFDA) for segmentation aims at adapting a
model trained in the source domain to perform well in the target domain with
only the source model and unlabeled target data.Inspired by the recent success
of Segment Anything Model (SAM) which exhibits the generality of segmenting
images of various modalities and in different domains given human-annotated
prompts like bounding boxes or points, we for the first time explore the
potentials of Segment Anything Model for SFDA via automatedly finding an
accurate bounding box prompt. We find that the bounding boxes directly
generated with existing SFDA approaches are defective due to the domain gap.To
tackle this issue, we propose a novel Dual Feature Guided (DFG) auto-prompting
approach to search for the box prompt. Specifically, the source model is first
trained in a feature aggregation phase, which not only preliminarily adapts the
source model to the target domain but also builds a feature distribution
well-prepared for box prompt search. In the second phase, based on two feature
distribution observations, we gradually expand the box prompt with the guidance
of the target model feature and the SAM feature to handle the class-wise
clustered target features and the class-wise dispersed target features,
respectively. To remove the potentially enlarged false positive regions caused
by the over-confident prediction of the target model, the refined pseudo-labels
produced by SAM are further postprocessed based on connectivity analysis.
Experiments on 3D and 2D datasets indicate that our approach yields superior
performance compared to conventional methods. Code is available at
https://github.com/xmed-lab/DFG.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 13:00:48 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 02:43:13 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Huai",
"Zheang",
""
],
[
"Tang",
"Hui",
""
],
[
"Li",
"Yi",
""
],
[
"Chen",
"Zhuangzhuang",
""
],
[
"Li",
"Xiaomeng",
""
]
] |
2505.08536 | Jiafu Hao | Jiafu Hao, Chentao Yue, Hao Chang, Branka Vucetic, and Yonghui Li | Short Wins Long: Short Codes with Language Model Semantic Correction
Outperform Long Codes | 6 pages, 3 figures | null | null | null | eess.SP cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a novel semantic-enhanced decoding scheme for
transmitting natural language sentences with multiple short block codes over
noisy wireless channels. After ASCII source coding, the natural language
sentence message is divided into segments, where each is encoded with short
block channel codes independently before transmission. At the receiver, each
short block of codewords is decoded in parallel, followed by a semantic error
correction (SEC) model to reconstruct corrupted segments semantically. We
design and train the SEC model based on Bidirectional and Auto-Regressive
Transformers (BART). Simulations demonstrate that the proposed scheme can
significantly outperform encoding the sentence with one conventional long LDPC
code, in terms of block error rate (BLER), semantic metrics, and decoding
latency. Finally, we proposed a semantic hybrid automatic repeat request (HARQ)
scheme to further enhance the error performance, which selectively requests
retransmission depends on semantic uncertainty.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 13:07:01 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Hao",
"Jiafu",
""
],
[
"Yue",
"Chentao",
""
],
[
"Chang",
"Hao",
""
],
[
"Vucetic",
"Branka",
""
],
[
"Li",
"Yonghui",
""
]
] |
2505.08537 | Mohamed Lamine Mekhalfi | Mohamed Lamine Mekhalfi, Paul Chippendale, Fabio Poiesi, Samuele
Bonecher, Gilberto Osler, Nicola Zancanella | The RaspGrade Dataset: Towards Automatic Raspberry Ripeness Grading with
Deep Learning | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This research investigates the application of computer vision for rapid,
accurate, and non-invasive food quality assessment, focusing on the novel
challenge of real-time raspberry grading into five distinct classes within an
industrial environment as the fruits move along a conveyor belt. To address
this, a dedicated dataset of raspberries, namely RaspGrade, was acquired and
meticulously annotated. Instance segmentation experiments revealed that
accurate fruit-level masks can be obtained; however, the classification of
certain raspberry grades presents challenges due to color similarities and
occlusion, while others are more readily distinguishable based on color. The
acquired and annotated RaspGrade dataset is accessible on Hugging Face at:
https://huggingface.co/datasets/FBK-TeV/RaspGrade.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 13:07:29 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 10:39:44 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Mekhalfi",
"Mohamed Lamine",
""
],
[
"Chippendale",
"Paul",
""
],
[
"Poiesi",
"Fabio",
""
],
[
"Bonecher",
"Samuele",
""
],
[
"Osler",
"Gilberto",
""
],
[
"Zancanella",
"Nicola",
""
]
] |
2505.08550 | Wenzhen Yue | Wenzhen Yue, Yong Liu, Haoxuan Li, Hao Wang, Xianghua Ying, Ruohao
Guo, Bowei Xing, Ji Shi | OLinear: A Linear Model for Time Series Forecasting in Orthogonally
Transformed Domain | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | This paper presents $\mathbf{OLinear}$, a $\mathbf{linear}$-based
multivariate time series forecasting model that operates in an
$\mathbf{o}$rthogonally transformed domain. Recent forecasting models typically
adopt the temporal forecast (TF) paradigm, which directly encode and decode
time series in the time domain. However, the entangled step-wise dependencies
in series data can hinder the performance of TF. To address this, some
forecasters conduct encoding and decoding in the transformed domain using
fixed, dataset-independent bases (e.g., sine and cosine signals in the Fourier
transform). In contrast, we utilize $\mathbf{OrthoTrans}$, a data-adaptive
transformation based on an orthogonal matrix that diagonalizes the series'
temporal Pearson correlation matrix. This approach enables more effective
encoding and decoding in the decorrelated feature domain and can serve as a
plug-in module to enhance existing forecasters. To enhance the representation
learning for multivariate time series, we introduce a customized linear layer,
$\mathbf{NormLin}$, which employs a normalized weight matrix to capture
multivariate dependencies. Empirically, the NormLin module shows a surprising
performance advantage over multi-head self-attention, while requiring nearly
half the FLOPs. Extensive experiments on 24 benchmarks and 140 forecasting
tasks demonstrate that OLinear consistently achieves state-of-the-art
performance with high efficiency. Notably, as a plug-in replacement for
self-attention, the NormLin module consistently enhances Transformer-based
forecasters. The code and datasets are available at
https://anonymous.4open.science/r/OLinear
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 10:39:37 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 11:00:57 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Yue",
"Wenzhen",
""
],
[
"Liu",
"Yong",
""
],
[
"Li",
"Haoxuan",
""
],
[
"Wang",
"Hao",
""
],
[
"Ying",
"Xianghua",
""
],
[
"Guo",
"Ruohao",
""
],
[
"Xing",
"Bowei",
""
],
[
"Shi",
"Ji",
""
]
] |
2505.08568 | Xiao Ni | Xiao Ni, Carsten Kuehnel, and Xiaoyi Jiang | Thermal Detection of People with Mobility Restrictions for Barrier
Reduction at Traffic Lights Controlled Intersections | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Rapid advances in deep learning for computer vision have driven the adoption
of RGB camera-based adaptive traffic light systems to improve traffic safety
and pedestrian comfort. However, these systems often overlook the needs of
people with mobility restrictions. Moreover, the use of RGB cameras presents
significant challenges, including limited detection performance under adverse
weather or low-visibility conditions, as well as heightened privacy concerns.
To address these issues, we propose a fully automated, thermal detector-based
traffic light system that dynamically adjusts signal durations for individuals
with walking impairments or mobility burden and triggers the auditory signal
for visually impaired individuals, thereby advancing towards barrier-free
intersection for all users. To this end, we build the thermal dataset for
people with mobility restrictions (TD4PWMR), designed to capture diverse
pedestrian scenarios, particularly focusing on individuals with mobility aids
or mobility burden under varying environmental conditions, such as different
lighting, weather, and crowded urban settings. While thermal imaging offers
advantages in terms of privacy and robustness to adverse conditions, it also
introduces inherent hurdles for object detection due to its lack of color and
fine texture details and generally lower resolution of thermal images. To
overcome these limitations, we develop YOLO-Thermal, a novel variant of the
YOLO architecture that integrates advanced feature extraction and attention
mechanisms for enhanced detection accuracy and robustness in thermal imaging.
Experiments demonstrate that the proposed thermal detector outperforms existing
detectors, while the proposed traffic light system effectively enhances
barrier-free intersection. The source codes and dataset are available at
https://github.com/leon2014dresden/YOLO-THERMAL.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 13:44:21 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 13:01:48 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Ni",
"Xiao",
""
],
[
"Kuehnel",
"Carsten",
""
],
[
"Jiang",
"Xiaoyi",
""
]
] |
2505.08592 | Kunpeng Zhang | Kunpeng Zhang, Lei Xu, Xinlei Yi, Guanghui Wen, Ming Cao, Karl H.
Johansson, Tianyou Chai, and Tao Yang | Communication-Efficient Distributed Online Nonconvex Optimization with
Time-Varying Constraints | 56 pages, 3 figures. arXiv admin note: substantial text overlap with
arXiv:2503.22410 | null | null | null | math.OC cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper considers distributed online nonconvex optimization with
time-varying inequality constraints over a network of agents, where the
nonconvex local loss and convex local constraint functions can vary arbitrarily
across iterations, and the information of them is privately revealed to each
agent at each iteration. For a uniformly jointly strongly connected
time-varying directed graph, we propose two distributed bandit online
primal--dual algorithm with compressed communication to efficiently utilize
communication resources in the one-point and two-point bandit feedback
settings, respectively. In nonconvex optimization, finding a globally optimal
decision is often NP-hard. As a result, the standard regret metric used in
online convex optimization becomes inapplicable. To measure the performance of
the proposed algorithms, we use a network regret metric grounded in the
first-order optimality condition associated with the variational inequality. We
show that the compressed algorithms establish sublinear network regret and
cumulative constraint violation bounds. Finally, a simulation example is
presented to validate the theoretical results.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 14:04:12 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 09:43:09 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Zhang",
"Kunpeng",
""
],
[
"Xu",
"Lei",
""
],
[
"Yi",
"Xinlei",
""
],
[
"Wen",
"Guanghui",
""
],
[
"Cao",
"Ming",
""
],
[
"Johansson",
"Karl H.",
""
],
[
"Chai",
"Tianyou",
""
],
[
"Yang",
"Tao",
""
]
] |
2505.08614 | Ziyuan He | Ziyuan He, Zhiqing Guo, Liejun Wang, Gaobo Yang, Yunfeng Diao, Dan Ma | WaveGuard: Robust Deepfake Detection and Source Tracing via Dual-Tree
Complex Wavelet and Graph Neural Networks | 11 pages, 5 figures, 4 tables | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deepfake technology poses increasing risks such as privacy invasion and
identity theft. To address these threats, we propose WaveGuard, a proactive
watermarking framework that enhances robustness and imperceptibility via
frequency-domain embedding and graph-based structural consistency.
Specifically, we embed watermarks into high-frequency sub-bands using Dual-Tree
Complex Wavelet Transform (DT-CWT) and employ a Structural Consistency Graph
Neural Network (SC-GNN) to preserve visual quality. We also design an attention
module to refine embedding precision. Experimental results on face swap and
reenactment tasks demonstrate that WaveGuard outperforms state-of-the-art
methods in both robustness and visual quality. Code is available at
https://github.com/vpsg-research/WaveGuard.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 14:31:42 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 03:27:59 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"He",
"Ziyuan",
""
],
[
"Guo",
"Zhiqing",
""
],
[
"Wang",
"Liejun",
""
],
[
"Yang",
"Gaobo",
""
],
[
"Diao",
"Yunfeng",
""
],
[
"Ma",
"Dan",
""
]
] |
2505.08660 | Pramath Anamby | Pramath Anamby and Soumya Das | Pullbacks of Saito-Kurokawa lifts of square-free levels, their
non-vanishing and the $L^2$-mass | 46 pages; Comments are most welcome | null | null | null | math.NT | http://creativecommons.org/licenses/by/4.0/ | We obtain the full spectral decomposition of the pullback of a Saito-Kurokawa
(SK) newform $F$ of odd, square-free level; and show that the projections onto
the elements $\mathbf g \otimes \mathbf g$ of an arithmetically orthogonalized
old-basis are either zero or whose squares are given by the certain
$\mathrm{GL}(3)\times \mathrm{GL}(2)$ central $L$-values $L(f\otimes
\mathrm{sym}^2 g, \frac{1}{2})$, where $F$ is the lift of the $\mathrm{GL}(2)$
newform $f$ and $g$ is the newform underlying $\mathbf g$. Based on this, we
work out a conjectural formula for the $L^2$-mass of the pullback of $F$ via
the CFKRS heuristics, which becomes a weighted average (over $g$) of the
central $L$-values. We show that on average over $f$, the main term predicted
by the above heuristics matches with the actual main term. We also provide
several results and sufficient conditions that ensure the non-vanishing of the
pullbacks.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 15:22:37 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 05:46:14 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Anamby",
"Pramath",
""
],
[
"Das",
"Soumya",
""
]
] |
2505.08699 | George Saon | George Saon, Avihu Dekel, Alexander Brooks, Tohru Nagano, Abraham
Daniels, Aharon Satt, Ashish Mittal, Brian Kingsbury, David Haws, Edmilson
Morais, Gakuto Kurata, Hagai Aronowitz, Ibrahim Ibrahim, Jeff Kuo, Kate
Soule, Luis Lastras, Masayuki Suzuki, Ron Hoory, Samuel Thomas, Sashi
Novitasari, Takashi Fukuda, Vishal Sunder, Xiaodong Cui, Zvi Kons | Granite-speech: open-source speech-aware LLMs with strong English ASR
capabilities | 7 pages, 9 figures | null | null | null | eess.AS | http://creativecommons.org/licenses/by/4.0/ | Granite-speech LLMs are compact and efficient speech language models
specifically designed for English ASR and automatic speech translation (AST).
The models were trained by modality aligning the 2B and 8B parameter variants
of granite-3.3-instruct to speech on publicly available open-source corpora
containing audio inputs and text targets consisting of either human transcripts
for ASR or automatically generated translations for AST. Comprehensive
benchmarking shows that on English ASR, which was our primary focus, they
outperform several competitors' models that were trained on orders of magnitude
more proprietary data, and they keep pace on English-to-X AST for major
European languages, Japanese, and Chinese. The speech-specific components are:
a conformer acoustic encoder using block attention and self-conditioning
trained with connectionist temporal classification, a windowed
query-transformer speech modality adapter used to do temporal downsampling of
the acoustic embeddings and map them to the LLM text embedding space, and LoRA
adapters to further fine-tune the text LLM. Granite-speech-3.3 operates in two
modes: in speech mode, it performs ASR and AST by activating the encoder,
projector, and LoRA adapters; in text mode, it calls the underlying
granite-3.3-instruct model directly (without LoRA), essentially preserving all
the text LLM capabilities and safety. Both models are freely available on
HuggingFace (https://huggingface.co/ibm-granite/granite-speech-3.3-2b and
https://huggingface.co/ibm-granite/granite-speech-3.3-8b) and can be used for
both research and commercial purposes under a permissive Apache 2.0 license.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 15:58:57 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 02:10:29 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Saon",
"George",
""
],
[
"Dekel",
"Avihu",
""
],
[
"Brooks",
"Alexander",
""
],
[
"Nagano",
"Tohru",
""
],
[
"Daniels",
"Abraham",
""
],
[
"Satt",
"Aharon",
""
],
[
"Mittal",
"Ashish",
""
],
[
"Kingsbury",
"Brian",
""
],
[
"Haws",
"David",
""
],
[
"Morais",
"Edmilson",
""
],
[
"Kurata",
"Gakuto",
""
],
[
"Aronowitz",
"Hagai",
""
],
[
"Ibrahim",
"Ibrahim",
""
],
[
"Kuo",
"Jeff",
""
],
[
"Soule",
"Kate",
""
],
[
"Lastras",
"Luis",
""
],
[
"Suzuki",
"Masayuki",
""
],
[
"Hoory",
"Ron",
""
],
[
"Thomas",
"Samuel",
""
],
[
"Novitasari",
"Sashi",
""
],
[
"Fukuda",
"Takashi",
""
],
[
"Sunder",
"Vishal",
""
],
[
"Cui",
"Xiaodong",
""
],
[
"Kons",
"Zvi",
""
]
] |
2505.08740 | Abdolmehdi Behroozi | Abdolmehdi Behroozi, Chaopeng Shen and, Daniel Kifer | Sensitivity-Constrained Fourier Neural Operators for Forward and Inverse
Problems in Parametric Differential Equations | null | null | null | null | cs.LG cs.CE | http://creativecommons.org/licenses/by/4.0/ | Parametric differential equations of the form du/dt = f(u, x, t, p) are
fundamental in science and engineering. While deep learning frameworks such as
the Fourier Neural Operator (FNO) can efficiently approximate solutions, they
struggle with inverse problems, sensitivity estimation (du/dp), and concept
drift. We address these limitations by introducing a sensitivity-based
regularization strategy, called Sensitivity-Constrained Fourier Neural
Operators (SC-FNO). SC-FNO achieves high accuracy in predicting solution paths
and consistently outperforms standard FNO and FNO with physics-informed
regularization. It improves performance in parameter inversion tasks, scales to
high-dimensional parameter spaces (tested with up to 82 parameters), and
reduces both data and training requirements. These gains are achieved with a
modest increase in training time (30% to 130% per epoch) and generalize across
various types of differential equations and neural operators. Code and selected
experiments are available at: https://github.com/AMBehroozi/SC_Neural_Operators
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 16:54:10 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 15:24:15 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Behroozi",
"Abdolmehdi",
""
],
[
"and",
"Chaopeng Shen",
""
],
[
"Kifer",
"Daniel",
""
]
] |
2505.08765 | Zhengqiu Zhu | Yatai Ji, Zhengqiu Zhu, Yong Zhao, Beidan Liu, Chen Gao, Yihao Zhao,
Sihang Qiu, Yue Hu, Quanjun Yin and Yong Li | Towards Autonomous UAV Visual Object Search in City Space: Benchmark and
Agentic Methodology | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Aerial Visual Object Search (AVOS) tasks in urban environments require
Unmanned Aerial Vehicles (UAVs) to autonomously search for and identify target
objects using visual and textual cues without external guidance. Existing
approaches struggle in complex urban environments due to redundant semantic
processing, similar object distinction, and the exploration-exploitation
dilemma. To bridge this gap and support the AVOS task, we introduce CityAVOS,
the first benchmark dataset for autonomous search of common urban objects. This
dataset comprises 2,420 tasks across six object categories with varying
difficulty levels, enabling comprehensive evaluation of UAV agents' search
capabilities. To solve the AVOS tasks, we also propose PRPSearcher
(Perception-Reasoning-Planning Searcher), a novel agentic method powered by
multi-modal large language models (MLLMs) that mimics human three-tier
cognition. Specifically, PRPSearcher constructs three specialized maps: an
object-centric dynamic semantic map enhancing spatial perception, a 3D
cognitive map based on semantic attraction values for target reasoning, and a
3D uncertainty map for balanced exploration-exploitation search. Also, our
approach incorporates a denoising mechanism to mitigate interference from
similar objects and utilizes an Inspiration Promote Thought (IPT) prompting
mechanism for adaptive action planning. Experimental results on CityAVOS
demonstrate that PRPSearcher surpasses existing baselines in both success rate
and search efficiency (on average: +37.69% SR, +28.96% SPL, -30.69% MSS, and
-46.40% NE). While promising, the performance gap compared to humans highlights
the need for better semantic reasoning and spatial exploration capabilities in
AVOS tasks. This work establishes a foundation for future advances in embodied
target search. Dataset and source code are available at
https://anonymous.4open.science/r/CityAVOS-3DF8.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 17:34:54 GMT"
},
{
"version": "v2",
"created": "Wed, 14 May 2025 01:30:03 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Ji",
"Yatai",
""
],
[
"Zhu",
"Zhengqiu",
""
],
[
"Zhao",
"Yong",
""
],
[
"Liu",
"Beidan",
""
],
[
"Gao",
"Chen",
""
],
[
"Zhao",
"Yihao",
""
],
[
"Qiu",
"Sihang",
""
],
[
"Hu",
"Yue",
""
],
[
"Yin",
"Quanjun",
""
],
[
"Li",
"Yong",
""
]
] |
2505.08790 | Octavio Miramontes | Leticia R. Paiva, Sidiney G. Alves, Og DeSouza and Octavio Miramontes | Emergent dynamical phases and collective motion in termites | 10 figures | null | null | null | q-bio.PE nlin.AO | http://creativecommons.org/licenses/by/4.0/ | Termites which are able to forage in the open can be often seen, in the field
or in the lab: (i) wandering around, forming no observable pattern, or (ii)
clustering themselves in a dense and almost immobile pack, or (iii) milling
about in a circular movement. Despite been well reported patterns, they are
normally regarded as independent phenomena whose specific traits have never
been properly quantified. Evidence, however, favours the hypothesis that these
are interdependent patterns, arisen from self-organised interactions and
movement among workers. After all, termites are a form of active matter where
blind cooperative individuals are self-propelled and lack the possibility of
visual cues to spatially orientate and align. It follows that their non-trivial
close-contact patterns could generate motion-collision induced phase
separations. This would then trigger the emergence of these three patterns
(disorder, clustering, milling) as parts of the same continuum. By inspecting
termite groups confined in arenas, we could quantitatively describe each one of
these patterns in detail. We identified disorder, clustering and milling
spatial patterns. These phases and their transitions are characterised aiming
to offer refinements in the understanding of these aspects of self-propelled
particles in active matter where close-range contacts and collisions are
important.
| [
{
"version": "v1",
"created": "Mon, 28 Apr 2025 21:36:30 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Paiva",
"Leticia R.",
""
],
[
"Alves",
"Sidiney G.",
""
],
[
"DeSouza",
"Og",
""
],
[
"Miramontes",
"Octavio",
""
]
] |
2505.08791 | Alexander Meyer | Alexander Meyer | Post-Quantum Cryptography: An Analysis of Code-Based and Lattice-Based
Cryptosystems | null | null | null | null | cs.CR cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | Most modern cryptographic systems, such as RSA and the Diffie-Hellman Key
Exchange, rely on "trapdoor" mathematical functions that are presumed to be
computationally difficult with existing tools. However, quantum computers will
be able to break these systems using Shor's Algorithm, necessitating the
development of quantum-resistant alternatives. We first examine the McEliece
cryptosystem, a code-based scheme believed to be secure against quantum attacks
due to the hardness of decoding arbitrary linear codes. We then explore NTRU, a
lattice-based system grounded in the difficulty of solving the Shortest Vector
Problem. Finally, we establish connections between the structural foundations
and security of the two systems.
| [
{
"version": "v1",
"created": "Tue, 6 May 2025 03:42:38 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Meyer",
"Alexander",
""
]
] |
2505.08792 | Brittany Johnson-Matthews | Michelle Nashla Turcios, Alicia E. Boyd, Angela D.R. Smith, Brittany
Johnson | A Preliminary Framework for Intersectionality in ML Pipelines | Accepted for the 1st International Intersectionality and Software
Engineering Workshop, colocated with FSE 2025 | null | 10.1145/3696630.3728692 | null | cs.LG cs.CY | http://creativecommons.org/licenses/by/4.0/ | Machine learning (ML) has become a go-to solution for improving how we use,
experience, and interact with technology (and the world around us).
Unfortunately, studies have repeatedly shown that machine learning technologies
may not provide adequate support for societal identities and experiences.
Intersectionality is a sociological framework that provides a mechanism for
explicitly considering complex social identities, focusing on social justice
and power. While the framework of intersectionality can support the development
of technologies that acknowledge and support all members of society, it has
been adopted and adapted in ways that are not always true to its foundations,
thereby weakening its potential for impact. To support the appropriate adoption
and use of intersectionality for more equitable technological outcomes, we
amplify the foundational intersectionality scholarship--Crenshaw, Combahee, and
Collins (three C's), to create a socially relevant preliminary framework in
developing machine-learning solutions. We use this framework to evaluate and
report on the (mis)alignments of intersectionality application in machine
learning literature.
| [
{
"version": "v1",
"created": "Tue, 6 May 2025 16:57:56 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Turcios",
"Michelle Nashla",
""
],
[
"Boyd",
"Alicia E.",
""
],
[
"Smith",
"Angela D. R.",
""
],
[
"Johnson",
"Brittany",
""
]
] |
2505.08793 | Monirul Islam Pavel | Monirul Islam Pavel, Siyi Hu, Mahardhika Pratama and Ryszard Kowalczyk | Onboard Optimization and Learning: A Survey | 36 pages, 5 figures, 3 tables | null | null | null | cs.LG cs.AR | http://creativecommons.org/licenses/by/4.0/ | Onboard learning is a transformative approach in edge AI, enabling real-time
data processing, decision-making, and adaptive model training directly on
resource-constrained devices without relying on centralized servers. This
paradigm is crucial for applications demanding low latency, enhanced privacy,
and energy efficiency. However, onboard learning faces challenges such as
limited computational resources, high inference costs, and security
vulnerabilities. This survey explores a comprehensive range of methodologies
that address these challenges, focusing on techniques that optimize model
efficiency, accelerate inference, and support collaborative learning across
distributed devices. Approaches for reducing model complexity, improving
inference speed, and ensuring privacy-preserving computation are examined
alongside emerging strategies that enhance scalability and adaptability in
dynamic environments. By bridging advancements in hardware-software co-design,
model compression, and decentralized learning, this survey provides insights
into the current state of onboard learning to enable robust, efficient, and
secure AI deployment at the edge.
| [
{
"version": "v1",
"created": "Wed, 7 May 2025 07:47:14 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Pavel",
"Monirul Islam",
""
],
[
"Hu",
"Siyi",
""
],
[
"Pratama",
"Mahardhika",
""
],
[
"Kowalczyk",
"Ryszard",
""
]
] |
2505.08794 | Martin Rivas | Juan Barandiaran and Martin Rivas | Classical Dirac particle I | 32 pages, 13 figures, 2 Mathematica Notebooks with 16 bookmarks | null | null | null | physics.class-ph math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | In this work we produce a classical Lagrangian description of an elementary
spinning particle which satisfies Dirac equation when quantized. We call this
particle a classical Dirac particle. We analyze in detail the way we arrive to
this model and how the different observables and constants of the motion can be
expressed in terms of the degrees of freedom and their derivatives, by making
use of Noether's theorem. The main feature is that the particle has a center of
charge r, moving at the speed of light, that satisfies fourth-order diferential
equations and all observables can be expressed only in terms of this point and
their time derivatives. The particle has also a center of mass q, that is a
different point than the center of charge. This implies that two different spin
observables can be defined, one S with respect to the point r and another SCM
with respect to the point q, that satisfy different dynamical equations. The
spin S satisfies the same dynamical equation than Dirac's spin operator. The
fourth-order differential equations for the point r can be transformed into a
system of second-order ordinary differential equations for the center of charge
and center of mass. The dynamics can be described in terms of dimensionless
variables. The possible interaction Lagrangians are described and we devote the
main part of the work to the electromagnetic interaction of the Dirac particle
with uniform and oscillating electric and magnetic fields. The numerical
integrations of the dynamical equations are performed with different
Mathematica notebooks that are available for the interested reader.
| [
{
"version": "v1",
"created": "Wed, 7 May 2025 10:58:09 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Barandiaran",
"Juan",
""
],
[
"Rivas",
"Martin",
""
]
] |
2505.08795 | Andres Anabalon | Andres Anabalon, Hugo Garces, Julio Oliva and Jose Cifuentes | The Geometry of Meaning: Perfect Spacetime Representations of
Hierarchical Structures | 7 pages, 3 figures | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | We show that there is a fast algorithm that embeds hierarchical structures in
three-dimensional Minkowski spacetime. The correlation of data ends up purely
encoded in the causal structure. Our model relies solely on oriented token
pairs -- local hierarchical signals -- with no access to global symbolic
structure. We apply our method to the corpus of \textit{WordNet}. We provide a
perfect embedding of the mammal sub-tree including ambiguities (more than one
hierarchy per node) in such a way that the hierarchical structures get
completely codified in the geometry and exactly reproduce the ground-truth. We
extend this to a perfect embedding of the maximal unambiguous subset of the
\textit{WordNet} with 82{,}115 noun tokens and a single hierarchy per token. We
introduce a novel retrieval mechanism in which causality, not distance, governs
hierarchical access. Our results seem to indicate that all discrete data has a
perfect geometrical representation that is three-dimensional. The resulting
embeddings are nearly conformally invariant, indicating deep connections with
general relativity and field theory. These results suggest that concepts,
categories, and their interrelations, namely hierarchical meaning itself, is
geometric.
| [
{
"version": "v1",
"created": "Wed, 7 May 2025 20:41:06 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Anabalon",
"Andres",
""
],
[
"Garces",
"Hugo",
""
],
[
"Oliva",
"Julio",
""
],
[
"Cifuentes",
"Jose",
""
]
] |
2505.08796 | Robbe De Prins | Robbe De Prins, Jacob Lamers, Peter Bienstman, Guy Van der Sande, Guy
Verschaffelt, Thomas Van Vaerenbergh | How to Incorporate External Fields in Analog Ising Machines | 30 pages, 17 figures, and 2 tables, including Supplementary Material.
The first two authors contributed equally | null | null | null | cond-mat.stat-mech cs.ET quant-ph | http://creativecommons.org/licenses/by/4.0/ | Ising machines (IMs) are specialized devices designed to efficiently solve
combinatorial optimization problems (COPs). They consist of artificial spins
that evolve towards a low-energy configuration representing a problem's
solution. Most realistic COPs require both spin-spin couplings and external
fields. In IMs with analog spins, these interactions scale differently with the
continuous spin amplitudes, leading to imbalances that affect performance.
Various techniques have been proposed to mitigate this issue, but their
performance has not been benchmarked. We address this gap through a numerical
analysis. We evaluate the time-to-solution of these methods across three
distinct problem classes with up to 500 spins. Our results show that the most
effective way to incorporate external fields is through an approach where the
spin interactions are proportional to the spin signs, rather than their
continuous amplitudes.
| [
{
"version": "v1",
"created": "Thu, 8 May 2025 18:01:02 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"De Prins",
"Robbe",
""
],
[
"Lamers",
"Jacob",
""
],
[
"Bienstman",
"Peter",
""
],
[
"Van der Sande",
"Guy",
""
],
[
"Verschaffelt",
"Guy",
""
],
[
"Van Vaerenbergh",
"Thomas",
""
]
] |
2505.08797 | Annaki Fouad | F. Annaki, S. Ouassou and S. Igamane | Visibility and Influence in Digital Social Relations: Towards a New
Symbolic Capital? | null | null | 10.5281/zenodo.15364096 | null | cs.SI cs.CY econ.GN q-fin.EC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This study explores the dynamics of visibility and influence in digital
social relations, examining their implications for the emergence of a new
symbolic capital. Using a mixedmethods design, the research combined
semi-structured interviews with 20 digitally active individuals and
quantitative social media data analysis to identify key predictors of digital
symbolic capital. Findings reveal that visibility is influenced by content
quality, network size, and engagement strategies, while influence depends on
credibility, authority, and trust. The study identifies a new form of symbolic
capital based on online visibility, influence, and reputation, distinct from
traditional forms. The research discusses the ethical implications of these
dynamics and suggests future research directions, emphasizing the need to
update social theories to account for digital transformations.
| [
{
"version": "v1",
"created": "Thu, 8 May 2025 19:10:47 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Annaki",
"F.",
""
],
[
"Ouassou",
"S.",
""
],
[
"Igamane",
"S.",
""
]
] |
2505.08798 | Mobina Shrestha | Mobina Shrestha, Bishwas Mandal, Vishal Mandal, Asis Shrestha | In-Context Learning for Label-Efficient Cancer Image Classification in
Oncology | null | null | null | null | eess.IV cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | The application of AI in oncology has been limited by its reliance on large,
annotated datasets and the need for retraining models for domain-specific
diagnostic tasks. Taking heed of these limitations, we investigated in-context
learning as a pragmatic alternative to model retraining by allowing models to
adapt to new diagnostic tasks using only a few labeled examples at inference,
without the need for retraining. Using four vision-language models
(VLMs)-Paligemma, CLIP, ALIGN and GPT-4o, we evaluated the performance across
three oncology datasets: MHIST, PatchCamelyon and HAM10000. To the best of our
knowledge, this is the first study to compare the performance of multiple VLMs
on different oncology classification tasks. Without any parameter updates, all
models showed significant gains with few-shot prompting, with GPT-4o reaching
an F1 score of 0.81 in binary classification and 0.60 in multi-class
classification settings. While these results remain below the ceiling of fully
fine-tuned systems, they highlight the potential of ICL to approximate
task-specific behavior using only a handful of examples, reflecting how
clinicians often reason from prior cases. Notably, open-source models like
Paligemma and CLIP demonstrated competitive gains despite their smaller size,
suggesting feasibility for deployment in computing constrained clinical
environments. Overall, these findings highlight the potential of ICL as a
practical solution in oncology, particularly for rare cancers and
resource-limited contexts where fine-tuning is infeasible and annotated data is
difficult to obtain.
| [
{
"version": "v1",
"created": "Thu, 8 May 2025 20:49:01 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Shrestha",
"Mobina",
""
],
[
"Mandal",
"Bishwas",
""
],
[
"Mandal",
"Vishal",
""
],
[
"Shrestha",
"Asis",
""
]
] |
2505.08799 | Loay Abdelrazek | Loay Abdelrazek, Rim ElMalki, Filippo Rebecchi, Daniel Cho | Measuring Security in 5G and Future Networks | Accepted and presented in IEEE Future Networks World Forum 2024
conference, This is a pre-print version | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | In today's increasingly interconnected and fast-paced digital ecosystem,
mobile networks, such as 5G and future generations such as 6G, play a pivotal
role and must be considered as critical infrastructures. Ensuring their
security is paramount to safeguard both individual users and the industries
that depend on these networks. An essential condition for maintaining and
improving the security posture of a system is the ability to effectively
measure and monitor its security state. In this work we address the need for an
objective measurement of the security state of 5G and future networks. We
introduce a state machine model designed to capture the security life cycle of
network functions and the transitions between different states within the life
cycle. Such a model can be computed locally at each node, or hierarchically, by
aggregating measurements into security domains or the whole network. We
identify three essential security metrics -- attack surface exposure, impact of
system vulnerabilities, and effectiveness of applied security controls -- that
collectively form the basis for calculating the overall security score. With
this approach, it is possible to provide a holistic understanding of the
security posture, laying the foundation for effective security management in
the expected dynamic threat landscape of 6G networks. Through practical
examples, we illustrate the real-world application of our proposed methodology,
offering valuable insights for developing risk management and informed
decision-making strategies in 5G and 6G security operations and laying the
foundation for effective security management in the expected dynamic threat
landscape of 6G networks.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 04:24:17 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Abdelrazek",
"Loay",
""
],
[
"ElMalki",
"Rim",
""
],
[
"Rebecchi",
"Filippo",
""
],
[
"Cho",
"Daniel",
""
]
] |
2505.08800 | Olivia Nocentini | Olivia Nocentini, Marta Lagomarsino, Gokhan Solak, Younggeol Cho, Qiyi
Tong, Marta Lorenzini, and Arash Ajoudani | Graph-based Online Monitoring of Train Driver States via Facial and
Skeletal Features | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Driver fatigue poses a significant challenge to railway safety, with
traditional systems like the dead-man switch offering limited and basic
alertness checks. This study presents an online behavior-based monitoring
system utilizing a customised Directed-Graph Neural Network (DGNN) to classify
train driver's states into three categories: alert, not alert, and
pathological. To optimize input representations for the model, an ablation
study was performed, comparing three feature configurations: skeletal-only,
facial-only, and a combination of both. Experimental results show that
combining facial and skeletal features yields the highest accuracy (80.88%) in
the three-class model, outperforming models using only facial or skeletal
features. Furthermore, this combination achieves over 99% accuracy in the
binary alertness classification. Additionally, we introduced a novel dataset
that, for the first time, incorporates simulated pathological conditions into
train driver monitoring, broadening the scope for assessing risks related to
fatigue and health. This work represents a step forward in enhancing railway
safety through advanced online monitoring using vision-based technologies.
| [
{
"version": "v1",
"created": "Fri, 9 May 2025 15:06:19 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Nocentini",
"Olivia",
""
],
[
"Lagomarsino",
"Marta",
""
],
[
"Solak",
"Gokhan",
""
],
[
"Cho",
"Younggeol",
""
],
[
"Tong",
"Qiyi",
""
],
[
"Lorenzini",
"Marta",
""
],
[
"Ajoudani",
"Arash",
""
]
] |
2505.08801 | Bishowjit Paul Jit | Md. Sakib Hassan Chowdhury, Md. Hafiz Ahamed, Bishowjit Paul, Sarafat
Hussain Abhi, Abu Bakar Siddique, Md. Robius Sany | OptiGait-LGBM: An Efficient Approach of Gait-based Person
Re-identification in Non-Overlapping Regions | 12 pages, 17 figures | null | null | null | cs.CV cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gait recognition, known for its ability to identify individuals from a
distance, has gained significant attention in recent times due to its
non-intrusive verification. While video-based gait identification systems
perform well on large public datasets, their performance drops when applied to
real-world, unconstrained gait data due to various factors. Among these,
uncontrolled outdoor environments, non-overlapping camera views, varying
illumination, and computational efficiency are core challenges in gait-based
authentication. Currently, no dataset addresses all these challenges
simultaneously. In this paper, we propose an OptiGait-LGBM model capable of
recognizing person re-identification under these constraints using a skeletal
model approach, which helps mitigate inconsistencies in a person's appearance.
The model constructs a dataset from landmark positions, minimizing memory usage
by using non-sequential data. A benchmark dataset, RUET-GAIT, is introduced to
represent uncontrolled gait sequences in complex outdoor environments. The
process involves extracting skeletal joint landmarks, generating numerical
datasets, and developing an OptiGait-LGBM gait classification model. Our aim is
to address the aforementioned challenges with minimal computational cost
compared to existing methods. A comparative analysis with ensemble techniques
such as Random Forest and CatBoost demonstrates that the proposed approach
outperforms them in terms of accuracy, memory usage, and training time. This
method provides a novel, low-cost, and memory-efficient video-based gait
recognition solution for real-world scenarios.
| [
{
"version": "v1",
"created": "Sat, 10 May 2025 08:28:57 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Chowdhury",
"Md. Sakib Hassan",
""
],
[
"Ahamed",
"Md. Hafiz",
""
],
[
"Paul",
"Bishowjit",
""
],
[
"Abhi",
"Sarafat Hussain",
""
],
[
"Siddique",
"Abu Bakar",
""
],
[
"Sany",
"Md. Robius",
""
]
] |
2505.08802 | Hyunseo Choi Mr. | Sanghyeon Park, Yuseung Hong, Hyunseo Choi | Enhanced Performance of Highly Activated Carbon and Surface-Treated
Porous Polymers as Physical Adsorbents for Chemical Warfare Agents | 14 Pages, 7 figures, This manuscript is formatted according to
journal submission standards and is posted on arXiv as a preprint for
dissemination and research record purposes. The study presents
adsorption-driven protective material strategies against chemical warfare
agents using activated carbon and functionalized porous organic polymers | null | null | null | physics.chem-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | The use of chemical warfare agents (CWAs) in modern warfare cannot be
disregarded due to their ease of use and potential for large-scale
incapacitation. An effective countermeasure involves the physical adsorption of
these agents, preventing their entry through the respiratory tract by
non-specific adsorption. In this study, we investigate the physical interaction
between potential adsorbents and model gases mimicking CWAs, thereby
identifying sufficient conditions for higher physical adsorption performance.
Our findings reveal that the physical adsorption capacity is highly sensitive
to the surface properties of the adsorbents, with uniform development of
micropores, rather than solely high surface area, emerging as a critical
factor. Additionally, we identified the potential of porous organic polymers as
promising alternatives to conventional activated carbon-based adsorbents.
Through a facile introduction of polar sulfone functional groups on the polymer
surface, we demonstrated that these polar surface polymers exhibit physical
adsorption capabilities for formaldehyde under ambient conditions comparable to
high-performance activated carbons. Notably, the superior activated carbon
possessed a high BET surface area of 2400 m^2/g and an exceptionally uniform
micropore structure with an average pore size of approximately 11 Angstroms.
This research paves the way for designing adsorbents with high physical
adsorption capacities tailored for CWAs protection, offering a significant
advancement in developing next-generation protective materials.
| [
{
"version": "v1",
"created": "Sat, 10 May 2025 17:20:06 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Park",
"Sanghyeon",
""
],
[
"Hong",
"Yuseung",
""
],
[
"Choi",
"Hyunseo",
""
]
] |
2505.08803 | Zizhao Hu | Zizhao Hu, Mohammad Rostami, Jesse Thomason | Multi-modal Synthetic Data Training and Model Collapse: Insights from
VLMs and Diffusion Models | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent research has highlighted the risk of generative model collapse, where
performance progressively degrades when continually trained on self-generated
data. However, existing exploration on model collapse is limited to single,
unimodal models, limiting our understanding in more realistic scenarios, such
as diverse multi-modal AI agents interacting autonomously through synthetic
data and continually evolving. We expand the synthetic data training and model
collapse study to multi-modal vision-language generative systems, such as
vision-language models (VLMs) and text-to-image diffusion models, as well as
recursive generate-train loops with multiple models. We find that model
collapse, previously observed in single-modality generative models, exhibits
distinct characteristics in the multi-modal context, such as improved
vision-language alignment and increased variance in VLM image-captioning task.
Additionally, we find that general approaches such as increased decoding
budgets, greater model diversity, and relabeling with frozen models can
effectively mitigate model collapse. Our findings provide initial insights and
practical guidelines for reducing the risk of model collapse in self-improving
multi-agent AI systems and curating robust multi-modal synthetic datasets.
| [
{
"version": "v1",
"created": "Sat, 10 May 2025 22:42:29 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Hu",
"Zizhao",
""
],
[
"Rostami",
"Mohammad",
""
],
[
"Thomason",
"Jesse",
""
]
] |
2505.08804 | Longtian Wang | Longtian Wang, Xiaofei Xie, Tianlin Li, Yuhan Zhi, Chao Shen | TokenProber: Jailbreaking Text-to-image Models via Fine-grained Word
Impact Analysis | 13 pages, 5 figures | null | null | null | cs.CR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Text-to-image (T2I) models have significantly advanced in producing
high-quality images. However, such models have the ability to generate images
containing not-safe-for-work (NSFW) content, such as pornography, violence,
political content, and discrimination. To mitigate the risk of generating NSFW
content, refusal mechanisms, i.e., safety checkers, have been developed to
check potential NSFW content. Adversarial prompting techniques have been
developed to evaluate the robustness of the refusal mechanisms. The key
challenge remains to subtly modify the prompt in a way that preserves its
sensitive nature while bypassing the refusal mechanisms. In this paper, we
introduce TokenProber, a method designed for sensitivity-aware differential
testing, aimed at evaluating the robustness of the refusal mechanisms in T2I
models by generating adversarial prompts. Our approach is based on the key
observation that adversarial prompts often succeed by exploiting discrepancies
in how T2I models and safety checkers interpret sensitive content. Thus, we
conduct a fine-grained analysis of the impact of specific words within prompts,
distinguishing between dirty words that are essential for NSFW content
generation and discrepant words that highlight the different sensitivity
assessments between T2I models and safety checkers. Through the
sensitivity-aware mutation, TokenProber generates adversarial prompts, striking
a balance between maintaining NSFW content generation and evading detection.
Our evaluation of TokenProber against 5 safety checkers on 3 popular T2I
models, using 324 NSFW prompts, demonstrates its superior effectiveness in
bypassing safety filters compared to existing methods (e.g., 54%+ increase on
average), highlighting TokenProber's ability to uncover robustness issues in
the existing refusal mechanisms.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 06:32:33 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wang",
"Longtian",
""
],
[
"Xie",
"Xiaofei",
""
],
[
"Li",
"Tianlin",
""
],
[
"Zhi",
"Yuhan",
""
],
[
"Shen",
"Chao",
""
]
] |
2505.08805 | Anastasia Konik | Anastasia Konik and Laurent Desbat | Range conditions on distributions and their possible application to
geometric calibration in 2D parallel and fan-beam geometries | null | null | null | null | stat.AP eess.IV | http://creativecommons.org/licenses/by/4.0/ | In tomography, range conditions or data consistency conditions (DCCs) on
functions have proven useful for geometric self-calibration, which involves
identifying geometric parameters of acquisition systems based only on acquired
radiographic images. These self-calibration methods using range conditions on
functions typically require non-truncated data. In this work, we derive range
conditions on distributions and demonstrate their application in addressing
data truncation issues during the calibration process. We propose a novel
approach based on range conditions on distributions, employing Dirac
distributions to model markers within the field-of-view of an X-ray system. Our
calibration methods are based on the local geometric information from
non-truncated projections of a marker set. By applying range conditions to
projections of sums of Dirac distributions, combined with specific calibration
marker sets, we derive analytical formulas that enable the identification of
geometric calibration parameters. We aim to present DCCs on distributions in
tomography and explore the potential of DCCs on distributions as a possible
tool in calibration. This approach represents one possible application,
demonstrating how DCCs on distributions can effectively address challenges such
as data truncation and incomplete marker set information. We present results
for the 2D parallel geometry (Radon transform) and the 2D fan-beam geometry
with sources on a line.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 14:06:45 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Konik",
"Anastasia",
""
],
[
"Desbat",
"Laurent",
""
]
] |
2505.08806 | Amaury Cruz Alvarez Dr. | A.C. Alvarez, H. Bruining, D. Marchesin | Estimation of oil recovery due to wettability changes in carbonate
reservoirs | null | null | null | null | cond-mat.mtrl-sci physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | Low salinity waterflooding (LSWF) enhances oil recovery at low cost in
carbonate reservoirs, but its effectiveness requires precise control of
injected water chemistry and interaction with reservoir minerals. This study
develops an integrated reactive transport model coupling geochemical surface
complexation modeling (SCM) with multiphase compositional dynamics to quantify
wettability alteration during LSWF. The framework combines PHREEQC-based
equilibrium calculations of the Total Bond Product (TBP), a wettability
indicator derived from oil-calcite ionic bridging, with Corey-type relative
permeability interpolation, resolved via COMSOL Multiphysics. Core flooding
simulations, compared with experimental data from calcite systems at 100
degrees Celsius and 220 bar, reveal that magnesium and sulfate concentrations
modulate TBP, reducing oil-rock adhesion under controlled low-salinity
conditions. Parametric analysis demonstrates that acidic crude oils (TAN higher
than 1 mg KOH per gram) exhibit TBP values approximately 2.5 times higher than
sweet crudes, due to carboxylate-calcite bridging, while pH elevation (above
7.5) amplifies wettability shifts by promoting deprotonated carboxyl
interactions. The model further identifies synergistic effects between
magnesium (50 to 200 millimoles per kilogram of water) and sulfate (above 500
millimoles per kilogram of water), which reduce calcium-mediated oil adhesion
through competitive mineral surface binding. By correlating TBP with fractional
flow dynamics, this framework could support the optimization of injection
strategies in carbonate reservoirs, suggesting that ion-specific adjustments
are more effective than bulk salinity reduction.
| [
{
"version": "v1",
"created": "Sun, 11 May 2025 22:11:15 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Alvarez",
"A. C.",
""
],
[
"Bruining",
"H.",
""
],
[
"Marchesin",
"D.",
""
]
] |
2505.08807 | Yuntao Wang | Yuntao Wang, Yanghe Pan, Shaolong Guo, and Zhou Su | Security of Internet of Agents: Attacks and Countermeasures | 11 pages, 5 figures, 3 tables, submitted to IEEE OJCS | null | null | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rise of large language and vision-language models, AI agents have
evolved into autonomous, interactive systems capable of perception, reasoning,
and decision-making. As they proliferate across virtual and physical domains,
the Internet of Agents (IoA) has emerged as a key infrastructure for enabling
scalable and secure coordination among heterogeneous agents. This survey offers
a comprehensive examination of the security and privacy landscape in IoA
systems. We begin by outlining the IoA architecture and its distinct
vulnerabilities compared to traditional networks, focusing on four critical
aspects: identity authentication threats, cross-agent trust issues, embodied
security, and privacy risks. We then review existing and emerging defense
mechanisms and highlight persistent challenges. Finally, we identify open
research directions to advance the development of resilient and
privacy-preserving IoA ecosystems.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 02:04:57 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wang",
"Yuntao",
""
],
[
"Pan",
"Yanghe",
""
],
[
"Guo",
"Shaolong",
""
],
[
"Su",
"Zhou",
""
]
] |
2505.08808 | Anqing Jiang | Anqing Jiang, Jinhao Chai, Yu Gao, Yiru Wang, Yuwen Heng, Zhigang Sun,
Hao Sun, Zezhong Zhao, Li Sun, Jian Zhou, Lijuan Zhu, Shugong Xu, Hao Zhao | SparseMeXT Unlocking the Potential of Sparse Representations for HD Map
Construction | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in high-definition \emph{HD} map construction have
demonstrated the effectiveness of dense representations, which heavily rely on
computationally intensive bird's-eye view \emph{BEV} features. While sparse
representations offer a more efficient alternative by avoiding dense BEV
processing, existing methods often lag behind due to the lack of tailored
designs. These limitations have hindered the competitiveness of sparse
representations in online HD map construction. In this work, we systematically
revisit and enhance sparse representation techniques, identifying key
architectural and algorithmic improvements that bridge the gap with--and
ultimately surpass--dense approaches. We introduce a dedicated network
architecture optimized for sparse map feature extraction, a sparse-dense
segmentation auxiliary task to better leverage geometric and semantic cues, and
a denoising module guided by physical priors to refine predictions. Through
these enhancements, our method achieves state-of-the-art performance on the
nuScenes dataset, significantly advancing HD map construction and centerline
detection. Specifically, SparseMeXt-Tiny reaches a mean average precision
\emph{mAP} of 55.5% at 32 frames per second \emph{fps}, while SparseMeXt-Base
attains 65.2% mAP. Scaling the backbone and decoder further, SparseMeXt-Large
achieves an mAP of 68.9% at over 20 fps, establishing a new benchmark for
sparse representations in HD map construction. These results underscore the
untapped potential of sparse methods, challenging the conventional reliance on
dense representations and redefining efficiency-performance trade-offs in the
field.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 02:26:58 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Jiang",
"Anqing",
""
],
[
"Chai",
"Jinhao",
""
],
[
"Gao",
"Yu",
""
],
[
"Wang",
"Yiru",
""
],
[
"Heng",
"Yuwen",
""
],
[
"Sun",
"Zhigang",
""
],
[
"Sun",
"Hao",
""
],
[
"Zhao",
"Zezhong",
""
],
[
"Sun",
"Li",
""
],
[
"Zhou",
"Jian",
""
],
[
"Zhu",
"Lijuan",
""
],
[
"Xu",
"Shugong",
""
],
[
"Zhao",
"Hao",
""
]
] |
2505.08809 | Shixi Qin | Shixi Qin, Zhiyong Yang, Shilong Bao, Shi Wang, Qianqian Xu, Qingming
Huang | MixBridge: Heterogeneous Image-to-Image Backdoor Attack through Mixture
of Schr\"odinger Bridges | null | null | null | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper focuses on implanting multiple heterogeneous backdoor triggers in
bridge-based diffusion models designed for complex and arbitrary input
distributions. Existing backdoor formulations mainly address single-attack
scenarios and are limited to Gaussian noise input models. To fill this gap, we
propose MixBridge, a novel diffusion Schr\"odinger bridge (DSB) framework to
cater to arbitrary input distributions (taking I2I tasks as special cases).
Beyond this trait, we demonstrate that backdoor triggers can be injected into
MixBridge by directly training with poisoned image pairs. This eliminates the
need for the cumbersome modifications to stochastic differential equations
required in previous studies, providing a flexible tool to study backdoor
behavior for bridge models. However, a key question arises: can a single DSB
model train multiple backdoor triggers? Unfortunately, our theory shows that
when attempting this, the model ends up following the geometric mean of benign
and backdoored distributions, leading to performance conflict across backdoor
tasks. To overcome this, we propose a Divide-and-Merge strategy to mix
different bridges, where models are independently pre-trained for each specific
objective (Divide) and then integrated into a unified model (Merge). In
addition, a Weight Reallocation Scheme (WRS) is also designed to enhance the
stealthiness of MixBridge. Empirical studies across diverse generation tasks
speak to the efficacy of MixBridge.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 06:40:23 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Qin",
"Shixi",
""
],
[
"Yang",
"Zhiyong",
""
],
[
"Bao",
"Shilong",
""
],
[
"Wang",
"Shi",
""
],
[
"Xu",
"Qianqian",
""
],
[
"Huang",
"Qingming",
""
]
] |
2505.08810 | Bappa Muktar | Bappa Muktar, Vincent Fono and Adama Nouboukpo | Machine Learning-Based Detection of DDoS Attacks in VANETs for Emergency
Vehicle Communication | null | null | null | null | cs.CR cs.AI | http://creativecommons.org/licenses/by/4.0/ | Vehicular Ad Hoc Networks (VANETs) play a key role in Intelligent
Transportation Systems (ITS), particularly in enabling real-time communication
for emergency vehicles. However, Distributed Denial of Service (DDoS) attacks,
which interfere with safety-critical communication channels, can severely
impair their reliability. This study introduces a robust and scalable framework
to detect DDoS attacks in highway-based VANET environments. A synthetic dataset
was constructed using Network Simulator 3 (NS-3) in conjunction with the
Simulation of Urban Mobility (SUMO) and further enriched with real-world
mobility traces from Germany's A81 highway, extracted via OpenStreetMap (OSM).
Three traffic categories were simulated: DDoS, VoIP, and TCP-based video
streaming (VideoTCP). The data preprocessing pipeline included normalization,
signal-to-noise ratio (SNR) feature engineering, missing value imputation, and
class balancing using the Synthetic Minority Over-sampling Technique (SMOTE).
Feature importance was assessed using SHapley Additive exPlanations (SHAP).
Eleven classifiers were benchmarked, among them XGBoost (XGB), CatBoost (CB),
AdaBoost (AB), GradientBoosting (GB), and an Artificial Neural Network (ANN).
XGB and CB achieved the best performance, each attaining an F1-score of 96%.
These results highlight the robustness of the proposed framework and its
potential for real-time deployment in VANETs to secure critical emergency
communications.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 07:00:04 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Muktar",
"Bappa",
""
],
[
"Fono",
"Vincent",
""
],
[
"Nouboukpo",
"Adama",
""
]
] |
2505.08811 | Shijie Lian | Shijie Lian and Ziyi Zhang and Laurence Tianruo Yang and, Mengyu Ren
and Debin Liu and Hua Li | TUGS: Physics-based Compact Representation of Underwater Scenes by
Tensorized Gaussian | null | null | null | null | cs.CV cs.RO | http://creativecommons.org/publicdomain/zero/1.0/ | Underwater 3D scene reconstruction is crucial for undewater robotic
perception and navigation. However, the task is significantly challenged by the
complex interplay between light propagation, water medium, and object surfaces,
with existing methods unable to model their interactions accurately.
Additionally, expensive training and rendering costs limit their practical
application in underwater robotic systems. Therefore, we propose Tensorized
Underwater Gaussian Splatting (TUGS), which can effectively solve the modeling
challenges of the complex interactions between object geometries and water
media while achieving significant parameter reduction. TUGS employs lightweight
tensorized higher-order Gaussians with a physics-based underwater Adaptive
Medium Estimation (AME) module, enabling accurate simulation of both light
attenuation and backscatter effects in underwater environments. Compared to
other NeRF-based and GS-based methods designed for underwater, TUGS is able to
render high-quality underwater images with faster rendering speeds and less
memory usage. Extensive experiments on real-world underwater datasets have
demonstrated that TUGS can efficiently achieve superior reconstruction quality
using a limited number of parameters, making it particularly suitable for
memory-constrained underwater UAV applications
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 07:09:35 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Lian",
"Shijie",
""
],
[
"Zhang",
"Ziyi",
""
],
[
"and",
"Laurence Tianruo Yang",
""
],
[
"Ren",
"Mengyu",
""
],
[
"Liu",
"Debin",
""
],
[
"Li",
"Hua",
""
]
] |
2505.08812 | Nicolas Ressayre | Micha\"el Bulois (ICJ, AGL), Roland Denis (INSMI-CNRS, CNRS, ICJ),
Nicolas Ressayre (ICJ, AGL) | An Algorithm to compute the Kronecker cone and other moment cones | null | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe an algorithm that computes the minimal list of inequalities for
the moment cone of any representation of a complex reductive group. We give
more details for the Kronecker cone (that is the asymptotic support of the
Kronecker coefficients) and the fermionic cone. These two cases correspond to
the action of $GL_{d_1}(\mathbb C)\times\cdots\times GL_{d_s}(\mathbb C)$ on
$\mathbb C^{d_1}\otimes\cdots\otimes \mathbb C^{d_s}$ and $GL_d(\mathbb C)$ on
$\bigwedge^r\mathbb C^d$, respectively. These two cases are implemented in
Python-Sage. Available at https://ea-icj.github.io/, our algorithm improves in
several directions, the one given by Vergne-Walter in Inequalities for Moment
Cones of Finite-Dimensional Representations. For example, it can compute the
minimal list of 463 inequalities for the Kronecker cone $(5,5,5)$ in 20
minutes. An important ingredient is an algorithm that decides whether an
application is birational or not. It could be of independent interest in
effective algebraic geometry.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 08:03:36 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Bulois",
"Michaël",
"",
"ICJ, AGL"
],
[
"Denis",
"Roland",
"",
"INSMI-CNRS, CNRS, ICJ"
],
[
"Ressayre",
"Nicolas",
"",
"ICJ, AGL"
]
] |
2505.08813 | Francesco Russo | Carlo Ciccarella, Francesco Russo (OC, ENSTA) | $C^{ 0,1}$ -It{\^o} chain rules and generalized solutions of parabolic
PDEs | arXiv admin note: substantial text overlap with arXiv:2407.06243 | null | null | null | math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we first establish an It\^o formula for a finite quadratic
variation process $X$ expanding $f(t,X_t),$ when $f$ is of class $C^2$ in space
and is absolutely continuous in time. Second, via a Fukushima-Dirichlet
decomposition we obtain an explicit chain rule for $f(t,X_t)$, when $X$ is a
continuous semimartingale and $f$ is a ``quasi-strong solution'' (in the sense
of approximation of classical solutions) of a parabolic PDE.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 08:13:51 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Ciccarella",
"Carlo",
"",
"OC, ENSTA"
],
[
"Russo",
"Francesco",
"",
"OC, ENSTA"
]
] |
2505.08814 | Wenkai Li | Wenkai Li, Xiaoqi Li, Yingjie Mao, Yishun Wang | Towards Understanding Deep Learning Model in Image Recognition via
Coverage Test | null | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks (DNNs) play a crucial role in the field of artificial
intelligence, and their security-related testing has been a prominent research
focus. By inputting test cases, the behavior of models is examined for
anomalies, and coverage metrics are utilized to determine the extent of neurons
covered by these test cases. With the widespread application and advancement of
DNNs, different types of neural behaviors have garnered attention, leading to
the emergence of various coverage metrics for neural networks. However, there
is currently a lack of empirical research on these coverage metrics,
specifically in analyzing the relationships and patterns between model depth,
configuration information, and neural network coverage. This paper aims to
investigate the relationships and patterns of four coverage metrics: primary
functionality, boundary, hierarchy, and structural coverage. A series of
empirical experiments were conducted, selecting LeNet, VGG, and ResNet as
different DNN architectures, along with 10 models of varying depths ranging
from 5 to 54 layers, to compare and study the relationships between different
depths, configuration information, and various neural network coverage metrics.
Additionally, an investigation was carried out on the relationships between
modified decision/condition coverage and dataset size. Finally, three potential
future directions are proposed to further contribute to the security testing of
DNN Models.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 08:25:55 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Li",
"Wenkai",
""
],
[
"Li",
"Xiaoqi",
""
],
[
"Mao",
"Yingjie",
""
],
[
"Wang",
"Yishun",
""
]
] |
2505.08815 | Yuejiang Shi | Yuejiang Shi | Achievement of 1 MA Discharges in Hydrogen-Boron Plasmas on EXL-50U | null | null | null | null | physics.plasm-ph | http://creativecommons.org/licenses/by/4.0/ | One mega ampere (MA) plasma discharges were achieved on the EXL-50U spherical
torus (ST) in hydrogen-boron (p-B) plasmas with toroidal fields up to 1 T at
major radii up to 0.6m. A key innovation in these experiments was the use of a
boron-rich fueling strategy, incorporating a high-concentration
boron-containing gas mixture and real-time boron powder injection during the
discharges. The boron content in the fueling reached 10%, representing the
first publicly reported MA-class hydrogen-boron plasma with such a high boron
concentration. Non-inductive plasma startup was achieved using electron
cyclotron resonance heating (ECRH), and a rapid current ramp-up to 1 MA was
realized through the synergistic use of ECRH and the central solenoid (CS).
With 800 kW of ECRH power, core electron temperatures of up to 3 keV were
attained. These results demonstrate the feasibility of producing
high-performance hydrogen-boron plasmas in a spherical torus configuration and
offer important physics and engineering insights for future reactor-scale
applications - particularly in the areas of low loop voltage current start-up
and real-time boronization techniques.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 09:55:58 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Shi",
"Yuejiang",
""
]
] |
2505.08816 | Ilias Syrigos | Ippokratis Koukoulis, Ilias Syrigos, and Thanasis Korakis | Self-Supervised Transformer-based Contrastive Learning for Intrusion
Detection Systems | Accepted at IFIP Networking 2025. Code available at
https://github.com/koukipp/contrastive_transformers_ids | null | null | null | cs.CR cs.LG | http://creativecommons.org/licenses/by/4.0/ | As the digital landscape becomes more interconnected, the frequency and
severity of zero-day attacks, have significantly increased, leading to an
urgent need for innovative Intrusion Detection Systems (IDS). Machine
Learning-based IDS that learn from the network traffic characteristics and can
discern attack patterns from benign traffic offer an advanced solution to
traditional signature-based IDS. However, they heavily rely on labeled
datasets, and their ability to generalize when encountering unseen traffic
patterns remains a challenge. This paper proposes a novel self-supervised
contrastive learning approach based on transformer encoders, specifically
tailored for generalizable intrusion detection on raw packet sequences. Our
proposed learning scheme employs a packet-level data augmentation strategy
combined with a transformer-based architecture to extract and generate
meaningful representations of traffic flows. Unlike traditional methods reliant
on handcrafted statistical features (NetFlow), our approach automatically
learns comprehensive packet sequence representations, significantly enhancing
performance in anomaly identification tasks and supervised learning for
intrusion detection. Our transformer-based framework exhibits better
performance in comparison to existing NetFlow self-supervised methods.
Specifically, we achieve up to a 3% higher AUC in anomaly detection for
intra-dataset evaluation and up to 20% higher AUC scores in inter-dataset
evaluation. Moreover, our model provides a strong baseline for supervised
intrusion detection with limited labeled data, exhibiting an improvement over
self-supervised NetFlow models of up to 1.5% AUC when pretrained and evaluated
on the same dataset. Additionally, we show the adaptability of our pretrained
model when fine-tuned across different datasets, demonstrating strong
performance even when lacking benign data from the target domain.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 13:42:00 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Koukoulis",
"Ippokratis",
""
],
[
"Syrigos",
"Ilias",
""
],
[
"Korakis",
"Thanasis",
""
]
] |
2505.08817 | Camilo Carvajal Reyes | Camilo Carvajal Reyes, Joaqu\'in Fontbona and Felipe Tobar | Towards SFW sampling for diffusion models via external conditioning | Accepcted at IJCNN 2025 | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Score-based generative models (SBM), also known as diffusion models, are the
de facto state of the art for image synthesis. Despite their unparalleled
performance, SBMs have recently been in the spotlight for being tricked into
creating not-safe-for-work (NSFW) content, such as violent images and
non-consensual nudity. Current approaches that prevent unsafe generation are
based on the models' own knowledge, and the majority of them require
fine-tuning. This article explores the use of external sources for ensuring
safe outputs in SBMs. Our safe-for-work (SFW) sampler implements a Conditional
Trajectory Correction step that guides the samples away from undesired regions
in the ambient space using multimodal models as the source of conditioning.
Furthermore, using Contrastive Language Image Pre-training (CLIP), our method
admits user-defined NSFW classes, which can vary in different settings. Our
experiments on the text-to-image SBM Stable Diffusion validate that the
proposed SFW sampler effectively reduces the generation of explicit content
while being competitive with other fine-tuning-based approaches, as assessed
via independent NSFW detectors. Moreover, we evaluate the impact of the SFW
sampler on image quality and show that the proposed correction scheme comes at
a minor cost with negligible effect on samples not needing correction. Our
study confirms the suitability of the SFW sampler towards aligned SBM models
and the potential of using model-agnostic conditioning for the prevention of
unwanted images.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 17:27:40 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Reyes",
"Camilo Carvajal",
""
],
[
"Fontbona",
"Joaquín",
""
],
[
"Tobar",
"Felipe",
""
]
] |
2505.08818 | Amara Tariq | Amara Tariq, Rimita Lahiri, Charles Kahn, Imon Banerjee | Position: Restructuring of Categories and Implementation of Guidelines
Essential for VLM Adoption in Healthcare | 15 pages, 2, tables, 3 figures | null | null | null | cs.CY cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | The intricate and multifaceted nature of vision language model (VLM)
development, adaptation, and application necessitates the establishment of
clear and standardized reporting protocols, particularly within the high-stakes
context of healthcare. Defining these reporting standards is inherently
challenging due to the diverse nature of studies involving VLMs, which vary
significantly from the development of all new VLMs or finetuning for domain
alignment to off-the-shelf use of VLM for targeted diagnosis and prediction
tasks. In this position paper, we argue that traditional machine learning
reporting standards and evaluation guidelines must be restructured to
accommodate multiphase VLM studies; it also has to be organized for intuitive
understanding of developers while maintaining rigorous standards for
reproducibility. To facilitate community adoption, we propose a categorization
framework for VLM studies and outline corresponding reporting standards that
comprehensively address performance evaluation, data reporting protocols, and
recommendations for manuscript composition. These guidelines are organized
according to the proposed categorization scheme. Lastly, we present a checklist
that consolidates reporting standards, offering a standardized tool to ensure
consistency and quality in the publication of VLM-related research.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 18:39:54 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Tariq",
"Amara",
""
],
[
"Lahiri",
"Rimita",
""
],
[
"Kahn",
"Charles",
""
],
[
"Banerjee",
"Imon",
""
]
] |
2505.08819 | Tsuyoshi Okita | Asahi Miyazaki, Tsuyoshi Okita | Thoughts on Objectives of Sparse and Hierarchical Masked Image Model | 9 pages, 11 figures | null | null | null | eess.IV cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Masked image modeling is one of the most poplular objectives of training.
Recently, the SparK model has been proposed with superior performance among
self-supervised learning models. This paper proposes a new mask pattern for
this SparK model, proposing it as the Mesh Mask-ed SparK model. We report the
effect of the mask pattern used for image masking in pre-training on
performance.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 18:40:46 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Miyazaki",
"Asahi",
""
],
[
"Okita",
"Tsuyoshi",
""
]
] |
2505.08820 | Xi-Wang Luo | Xian-Hao Wei, Xi-Wang Luo, Guang-Can Guo and Zheng-Wei Zhou | Higher-order Topological Parity Anomaly and Half-integer Hall Effect in
High-dimensional Synthetic Lattices | 5 pages, 4 figures, with 15 page Supplementary information | null | null | null | cond-mat.mes-hall cond-mat.other physics.optics quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in constructing synthetic dimension provide a powerful tool
for exploring exotic topological states of matter in high dimensions. Here we
report that the parity anomaly and associated \textit{half-integer} quantized
Hall conductance, arising in 2$j$+1 (space-time) dimensions with a single or
odd number of Dirac cones, can be realized by the boundary states of $n$-th
order topological insulators in (2$j$+$n$)-dimensional synthetic lattices. We
establish a general bulk-boundary correspondence by integrating the ``nested"
Wilson loop theory with the time-reversal polarization at highly-symmetric
momenta, a set of $Z_2$ topological invariants are extracted which determines
the number of higher-order-boundary Dirac cones and their locations. We develop
a general construction procedure for Hamiltonians supporting such higher-order
topological parity anomaly. Moreover, we propose an experimental implementation
scheme based on photonic synthetic dimensions and provide a method for probing
the associated half-integer Hall conductance by the transmission spectra. Our
work offers the realization and characterization of parity anomaly in general
high-dimensional higher-order topological insulators and opens an avenue for
exploring fundamental physics and possible device applications enabled by
manipulating Dirac cones.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 18:44:54 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wei",
"Xian-Hao",
""
],
[
"Luo",
"Xi-Wang",
""
],
[
"Guo",
"Guang-Can",
""
],
[
"Zhou",
"Zheng-Wei",
""
]
] |
2505.08821 | Meryem Karagoz | Meryem Altin Karagoz, Marc D. Breton, Anas El Fathi | A Comparative Study of Transformer-Based Models for Multi-Horizon Blood
Glucose Prediction | 7 pages, 2 figures, 1 table, 1st IFAC Workshop on Engineering
Diabetes Technologies (EDT 2025) | null | null | null | q-bio.QM cs.AI stat.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Accurate blood glucose prediction can enable novel interventions for type 1
diabetes treatment, including personalized insulin and dietary adjustments.
Although recent advances in transformer-based architectures have demonstrated
the power of attention mechanisms in complex multivariate time series
prediction, their potential for blood glucose (BG) prediction remains
underexplored. We present a comparative analysis of transformer models for
multi-horizon BG prediction, examining forecasts up to 4 hours and input
history up to 1 week. The publicly available DCLP3 dataset (n=112) was split
(80%-10%-10%) for training, validation, and testing, and the OhioT1DM dataset
(n=12) served as an external test set. We trained networks with point-wise,
patch-wise, series-wise, and hybrid embeddings, using CGM, insulin, and meal
data. For short-term blood glucose prediction, Crossformer, a patch-wise
transformer architecture, achieved a superior 30-minute prediction of RMSE
(15.6 mg / dL on OhioT1DM). For longer-term predictions (1h, 2h, and 4h),
PatchTST, another path-wise transformer, prevailed with the lowest RMSE (24.6
mg/dL, 36.1 mg/dL, and 46.5 mg/dL on OhioT1DM). In general, models that used
tokenization through patches demonstrated improved accuracy with larger input
sizes, with the best results obtained with a one-week history. These findings
highlight the promise of transformer-based architectures for BG prediction by
capturing and leveraging seasonal patterns in multivariate time-series data to
improve accuracy.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 20:22:44 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Karagoz",
"Meryem Altin",
""
],
[
"Breton",
"Marc D.",
""
],
[
"Fathi",
"Anas El",
""
]
] |
2505.08822 | Kailai Wang | Yuhao Wang, Kailai Wang, Songhua Hu, Yunpeng (Jack) Zhang, Gino Lim,
and Pengyu Zhu | The Geography of Transportation Cybersecurity: Visitor Flows, Industry
Clusters, and Spatial Dynamics | null | null | null | null | cs.CY cs.LG physics.soc-ph | http://creativecommons.org/licenses/by/4.0/ | The rapid evolution of the transportation cybersecurity ecosystem,
encompassing cybersecurity, automotive, and transportation and logistics
sectors, will lead to the formation of distinct spatial clusters and visitor
flow patterns across the US. This study examines the spatiotemporal dynamics of
visitor flows, analyzing how socioeconomic factors shape industry clustering
and workforce distribution within these evolving sectors. To model and predict
visitor flow patterns, we develop a BiTransGCN framework, integrating an
attention-based Transformer architecture with a Graph Convolutional Network
backbone. By integrating AI-enabled forecasting techniques with spatial
analysis, this study improves our ability to track, interpret, and anticipate
changes in industry clustering and mobility trends, thereby supporting
strategic planning for a secure and resilient transportation network. It offers
a data-driven foundation for economic planning, workforce development, and
targeted investments in the transportation cybersecurity ecosystem.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 20:44:02 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Wang",
"Yuhao",
"",
"Jack"
],
[
"Wang",
"Kailai",
"",
"Jack"
],
[
"Hu",
"Songhua",
"",
"Jack"
],
[
"Yunpeng",
"",
"",
"Jack"
],
[
"Zhang",
"",
""
],
[
"Lim",
"Gino",
""
],
[
"Zhu",
"Pengyu",
""
]
] |
2505.08823 | Cody Steinmetz | Cody Steinmetz, Gavin Childress, Aaron Herbst, Gavin Jones, Jasdeep
Singh, Eli Vang, and Keagan Weinstock | An Extra RMSNorm is All You Need for Fine Tuning to 1.58 Bits | null | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have transformed natural-language processing,
yet their scale makes real-world deployment costly. Post-training quantization
reduces memory and computation but often degrades accuracy, while
quantization-aware training can recover performance at the cost of extra
training. Pushing quantization to the ternary (2-bit) regime yields even larger
savings but is notoriously unstable. Building on recent work showing that a
bias-free, RMS-normalized Transformer with straight-through estimation can
reach 1.58-bit precision, we demonstrate that simply inserting RMS
normalization before every linear projection and applying a gradual, layer-wise
quantization schedule stably fine-tunes full-precision checkpoints into ternary
LLMs. Our approach matches or surpasses more elaborate knowledge-distillation
pipelines on standard language-modeling benchmarks without adding model
complexity. These results indicate that careful normalization alone can close
much of the accuracy gap between ternary and full-precision LLMs, making
ultra-low-bit inference practical.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 21:14:29 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Steinmetz",
"Cody",
""
],
[
"Childress",
"Gavin",
""
],
[
"Herbst",
"Aaron",
""
],
[
"Jones",
"Gavin",
""
],
[
"Singh",
"Jasdeep",
""
],
[
"Vang",
"Eli",
""
],
[
"Weinstock",
"Keagan",
""
]
] |
2505.08824 | Swagat Saurav Mishra | Varun Sahni | Remembering Alexei Starobinsky -- the gentle giant of cosmology | 23 pages, 10 figures, Contribution to the book "Open Issues in
Gravitation and Cosmology - Original Contributions, Essays and Recollections
in Honor of Alexei Starobinsky", to be published by Springer, edited by
Andrei Barvinsky and Alexander Kamenshchik | null | null | null | gr-qc astro-ph.CO hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | I share fond memories of my former PhD advisor Alexei Starobinsky with whom I
was closely associated for nearly 45 years. I reflect upon my early years in
Moscow when I worked with him on my thesis, and touch upon the seminal work on
inflation which he did during that period. Alexei visited India often and
actively interacted with Indian scientists and students on issues relating to
inflation, large scale structure and dark energy. This extensive collaboration,
which lasted several decades, resulted in the publication of over a dozen
important papers, several PhD's, and the development of the Statefinder and Om
diagnostics, which I briefly discuss.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 21:30:31 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Sahni",
"Varun",
""
]
] |
2505.08825 | Pedro Antonio Alarcon Granadeno | Pedro Antonio Alarcon Granadeno, Theodore Chambers, Jane Cleland-Huang | Multi-source Plume Tracing via Multi-Agent Reinforcement Learning | 13 pages, 7 figures | null | null | null | cs.MA cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Industrial catastrophes like the Bhopal disaster (1984) and the Aliso Canyon
gas leak (2015) demonstrate the urgent need for rapid and reliable plume
tracing algorithms to protect public health and the environment. Traditional
methods, such as gradient-based or biologically inspired approaches, often fail
in realistic, turbulent conditions. To address these challenges, we present a
Multi-Agent Reinforcement Learning (MARL) algorithm designed for localizing
multiple airborne pollution sources using a swarm of small uncrewed aerial
systems (sUAS). Our method models the problem as a Partially Observable Markov
Game (POMG), employing a Long Short-Term Memory (LSTM)-based Action-specific
Double Deep Recurrent Q-Network (ADDRQN) that uses full sequences of historical
action-observation pairs, effectively approximating latent states. Unlike prior
work, we use a general-purpose simulation environment based on the Gaussian
Plume Model (GPM), incorporating realistic elements such as a three-dimensional
environment, sensor noise, multiple interacting agents, and multiple plume
sources. The incorporation of action histories as part of the inputs further
enhances the adaptability of our model in complex, partially observable
environments. Extensive simulations show that our algorithm significantly
outperforms conventional approaches. Specifically, our model allows agents to
explore only 1.29\% of the environment to successfully locate pollution
sources.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 21:33:15 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Granadeno",
"Pedro Antonio Alarcon",
""
],
[
"Chambers",
"Theodore",
""
],
[
"Cleland-Huang",
"Jane",
""
]
] |
2505.08826 | Phil Attard | Phil Attard | The Two-Fluid Theory for Superfluid Hydrodynamics and the Fountain
Pressure Paradox | 6 pages | null | null | null | cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | The two-fluid theory for superfluid hydrodynamics is derived from the
fountain pressure result that condensed bosons move at constant entropy, and
that therefore steady superfluid flow connects regions of equal chemical
potential. The paradox of the fountain pressure equation is resolved with a
quantitative prediction.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 21:57:44 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Attard",
"Phil",
""
]
] |
2505.08827 | Toby Simonds | Toby Simonds, Kevin Lopez, Akira Yoshiyama, Dominique Garmier | Self Rewarding Self Improving | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | We demonstrate that large language models can effectively self-improve
through self-judging without requiring reference solutions, leveraging the
inherent asymmetry between generating and verifying solutions. Our experiments
on Countdown puzzles and MIT Integration Bee problems show that models can
provide reliable reward signals without ground truth answers, enabling
reinforcement learning in domains previously not possible. By implementing
self-judging, we achieve significant performance gains maintaining alignment
with formal verification. When combined with synthetic question generation, we
establish a complete self-improvement loop where models generate practice
problems, solve them, and evaluate their own performance-achieving an 8%
improvement with Qwen 2.5 7B over baseline and surpassing GPT-4o performance on
integration tasks. Our findings demonstrate that LLM judges can provide
effective reward signals for training models, unlocking many reinforcement
learning environments previously limited by the difficulty of creating
programmatic rewards. This suggests a potential paradigm shift toward AI
systems that continuously improve through self-directed learning rather than
human-guided training, potentially accelerating progress in domains with scarce
training data or complex evaluation requirements.
| [
{
"version": "v1",
"created": "Mon, 12 May 2025 23:51:04 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Simonds",
"Toby",
""
],
[
"Lopez",
"Kevin",
""
],
[
"Yoshiyama",
"Akira",
""
],
[
"Garmier",
"Dominique",
""
]
] |
2505.08828 | Eduardo Oliveira | Eduardo Araujo Oliveira, Madhavi Mohoni, Sonsoles L\'opez-Pernas,
Mohammed Saqr | Human-AI Collaboration or Academic Misconduct? Measuring AI Use in
Student Writing Through Stylometric Evidence | 19 pages, 10 figures, 11 tables | null | null | null | cs.CL cs.AI cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As human-AI collaboration becomes increasingly prevalent in educational
contexts, understanding and measuring the extent and nature of such
interactions pose significant challenges. This research investigates the use of
authorship verification (AV) techniques not as a punitive measure, but as a
means to quantify AI assistance in academic writing, with a focus on promoting
transparency, interpretability, and student development. Building on prior
work, we structured our investigation into three stages: dataset selection and
expansion, AV method development, and systematic evaluation. Using three
datasets - including a public dataset (PAN-14) and two from University of
Melbourne students from various courses - we expanded the data to include
LLM-generated texts, totalling 1,889 documents and 540 authorship problems from
506 students. We developed an adapted Feature Vector Difference AV methodology
to construct robust academic writing profiles for students, designed to capture
meaningful, individual characteristics of their writing. The method's
effectiveness was evaluated across multiple scenarios, including distinguishing
between student-authored and LLM-generated texts and testing resilience against
LLMs' attempts to mimic student writing styles. Results demonstrate the
enhanced AV classifier's ability to identify stylometric discrepancies and
measure human-AI collaboration at word and sentence levels while providing
educators with a transparent tool to support academic integrity investigations.
This work advances AV technology, offering actionable insights into the
dynamics of academic writing in an AI-driven era.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 00:36:36 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Oliveira",
"Eduardo Araujo",
""
],
[
"Mohoni",
"Madhavi",
""
],
[
"López-Pernas",
"Sonsoles",
""
],
[
"Saqr",
"Mohammed",
""
]
] |
2505.08830 | Wenhao Jiang | Wenhao Jiang and Yuchuan Luo and Guilin Deng and Silong Chen and Xu
Yang and Shihong Wu and Xinwen Gao and Lin Liu and Shaojing Fu | Federated Large Language Models: Feasibility, Robustness, Security and
Future Directions | 35 pages | null | null | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The integration of Large Language Models (LLMs) and Federated Learning (FL)
presents a promising solution for joint training on distributed data while
preserving privacy and addressing data silo issues. However, this emerging
field, known as Federated Large Language Models (FLLM), faces significant
challenges, including communication and computation overheads, heterogeneity,
privacy and security concerns. Current research has primarily focused on the
feasibility of FLLM, but future trends are expected to emphasize enhancing
system robustness and security. This paper provides a comprehensive review of
the latest advancements in FLLM, examining challenges from four critical
perspectives: feasibility, robustness, security, and future directions. We
present an exhaustive survey of existing studies on FLLM feasibility, introduce
methods to enhance robustness in the face of resource, data, and task
heterogeneity, and analyze novel risks associated with this integration,
including privacy threats and security challenges. We also review the latest
developments in defense mechanisms and explore promising future research
directions, such as few-shot learning, machine unlearning, and IP protection.
This survey highlights the pressing need for further research to enhance system
robustness and security while addressing the unique challenges posed by the
integration of FL and LLM.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 03:23:54 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Jiang",
"Wenhao",
""
],
[
"Luo",
"Yuchuan",
""
],
[
"Deng",
"Guilin",
""
],
[
"Chen",
"Silong",
""
],
[
"Yang",
"Xu",
""
],
[
"Wu",
"Shihong",
""
],
[
"Gao",
"Xinwen",
""
],
[
"Liu",
"Lin",
""
],
[
"Fu",
"Shaojing",
""
]
] |
2505.08831 | Arish Alreja | Arish Alreja, Michael J. Ward, Lisa S. Parker, R. Mark Richardson,
Louis-Philippe Morency, Taylor J. Abel, Avniel Singh Ghuman | Neural encoding of real world face perception | null | null | null | null | q-bio.NC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Social perception unfolds as we freely interact with people around us. We
investigated the neural basis of real world face perception using multi
electrode intracranial recordings in humans during spontaneous interactions
with friends, family, and others. Computational models reconstructed the faces
participants looked at during natural interactions, including facial
expressions and motion, from brain activity alone. The results highlighted a
critical role for the social vision pathway, a network of areas spanning
parietal, temporal, and occipital cortex. This network was more sharply tuned
to subtle expressions compared to intense expressions, which was confirmed with
controlled psychophysical experiments. These findings reveal that the human
social vision pathway encodes facial expressions and motion as deviations from
a neutral expression prototype during natural social interactions in real life.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 03:26:56 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Alreja",
"Arish",
""
],
[
"Ward",
"Michael J.",
""
],
[
"Parker",
"Lisa S.",
""
],
[
"Richardson",
"R. Mark",
""
],
[
"Morency",
"Louis-Philippe",
""
],
[
"Abel",
"Taylor J.",
""
],
[
"Ghuman",
"Avniel Singh",
""
]
] |
2505.08832 | Leif Svalgaard | Leif Svalgaard | A Possible Problem with the Sunspot Number | 8 Figures | null | null | null | astro-ph.IM astro-ph.SR physics.space-ph | http://creativecommons.org/licenses/by/4.0/ | Sunspots are areas of strong magnetic fields driven by a convective dynamo.
Rudolf Wolf devised his Sunspot Number (SN) series to describe its variation
with time. Most other solar phenomena vary in concert with SN, in particular
the microwave radiation. During the interval 2014-Jul to 2015-Dec, the
variation of the SN with the microwave flux (e.g. F10.7) was anomalous,
although microwave flux did not exhibit any anomaly with respect to the solar
magnetic field measured by the HMI instrument on the SDO spacecraft. This
points to a possible problem with the derivation of the SN using the 148,714
reports of sunspot observations received by the World Data Center Solar
Influences (SIDC/SILSO at ROB) since 2011 (when the last public raw data was
released). The unavailability of the raw data since then, contrary to avowed
open data policy of ROB, prevents independent assessment and possible
correction of the anomaly. We urge ROB to make the raw data available.
| [
{
"version": "v1",
"created": "Tue, 13 May 2025 04:17:29 GMT"
}
] | 2025-05-15T00:00:00 | [
[
"Svalgaard",
"Leif",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.