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
2025-04-25 00:00:00
| authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2409.19286 | Qingming Zeng | Qingming Zeng (Harbin Institute of Technology, Shenzhen), Mo Li (The
Chinese University of Hongkong, Shenzhen), Ximing Fu (Harbin Institute of
Technology, Shenzhen), Chuanyi Liu (Harbin Institute of Technology, Shenzhen,
Peng Cheng Laboratory, Shenzhen), Hui Jiang (Tsinghua University, Baidu Inc) | Imitater: An Efficient Shared Mempool Protocol with Application to
Byzantine Fault Tolerance | 25 pages, 8 figures | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Byzantine Fault Tolerant (BFT) consensus, a cornerstone of blockchain
technology, has seen significant advancements. While existing BFT protocols
ensure security guarantees, they often suffer from efficiency challenges,
particularly under conditions of network instability or malicious exploitation
of system mechanisms.
We propose a novel Shared Mempool (SMP) protocol, named Imitater, which can
be seamlessly integrated into BFT protocols. By chaining microblocks and
applying coding techniques, Imitater efficiently achieves \emph{totality} and
\emph{availability}. Furthermore, a BFT protocol augmented with Imitater
ensures \emph{order preservation} of client transactions while mitigating the
risks of \emph{over-distribution} and \emph{unbalanced workload}.
In the experiment, we integrate Imitater into the HotStuff protocol,
resulting in Imitater-HS. The performance of Imitater-HS is validated in a
system with up to 256 nodes. Experimental results demonstrate the efficiency of
our approach: Imitater-HS achieves higher throughput and lower latency in the
presence of faulty nodes compared to Stratus-HS, the state-of-the-art protocol.
Notably, the throughput improvement increases with the number of faulty nodes.
| [
{
"version": "v1",
"created": "Sat, 28 Sep 2024 08:50:32 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 09:15:17 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zeng",
"Qingming",
"",
"Harbin Institute of Technology, Shenzhen"
],
[
"Li",
"Mo",
"",
"The\n Chinese University of Hongkong, Shenzhen"
],
[
"Fu",
"Ximing",
"",
"Harbin Institute of\n Technology, Shenzhen"
],
[
"Liu",
"Chuanyi",
"",
"Harbin Institute of Technology, Shenzhen,\n Peng Cheng Laboratory, Shenzhen"
],
[
"Jiang",
"Hui",
"",
"Tsinghua University, Baidu Inc"
]
] |
2409.19320 | Vikash Kumar Dubey Mr. | Vikash Kumar Dubey, Suman Chakraborty, Arunava Patra and Sagar
Chakraborty | Dynamical stability of evolutionarily stable strategy in asymmetric
games | The earlier version (arXiv:2409.19320v2) had an inadvertent error
that led to incorrect results. This revised version rectifies those results | null | null | null | q-bio.PE nlin.AO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Evolutionarily stable strategy (ESS) is the defining concept of evolutionary
game theory. It has a fairly unanimously accepted definition for the case of
symmetric games which are played in a homogeneous population where all
individuals are in same role. However, in asymmetric games, which are played in
a population with multiple subpopulations (each of which has individuals in one
particular role), situation is not as clear. Various generalizations of ESS
defined for such cases differ in how they correspond to fixed points of
replicator equation which models evolutionary dynamics of frequencies of
strategies in the population. Moreover, some of the definitions may even be
equivalent, and hence, redundant in the scheme of things. Along with reporting
some new results, this paper is partly indented as a contextual mini-review of
some of the most important definitions of ESS in asymmetric games. We present
the definitions coherently and scrutinize them closely while establishing
equivalences -- some of them hitherto unreported -- between them wherever
possible. Since it is desirable that a definition of ESS should correspond to
asymptotically stable fixed points of replicator dynamics, we bring forward the
connections between various definitions and their dynamical stabilities.
Furthermore, we find the use of principle of relative entropy to gain
information-theoretic insights into the concept of ESS in asymmetric games,
thereby establishing a three-fold connection between game theory, dynamical
system theory, and information theory in this context. We discuss our
conclusions also in the backdrop of asymmetric hypermatrix games where more
than two individuals interact simultaneously in the course of getting payoffs.
| [
{
"version": "v1",
"created": "Sat, 28 Sep 2024 11:11:26 GMT"
},
{
"version": "v2",
"created": "Sat, 30 Nov 2024 12:28:44 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 05:24:06 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Dubey",
"Vikash Kumar",
""
],
[
"Chakraborty",
"Suman",
""
],
[
"Patra",
"Arunava",
""
],
[
"Chakraborty",
"Sagar",
""
]
] |
2409.19777 | Oliver Hines | Christian L. Hines, Oliver J. Hines | Automatic debiasing of neural networks via moment-constrained learning | Code repository and license available at
https://github.com/crimbs/madnet | null | null | null | stat.ML cs.LG stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Causal and nonparametric estimands in economics and biostatistics can often
be viewed as the mean of a linear functional applied to an unknown outcome
regression function. Naively learning the regression function and taking a
sample mean of the target functional results in biased estimators, and a rich
debiasing literature has developed where one additionally learns the so-called
Riesz representer (RR) of the target estimand (targeted learning, double ML,
automatic debiasing etc.). Learning the RR via its derived functional form can
be challenging, e.g. due to extreme inverse probability weights or the need to
learn conditional density functions. Such challenges have motivated recent
advances in automatic debiasing (AD), where the RR is learned directly via
minimization of a bespoke loss. We propose moment-constrained learning as a new
RR learning approach that addresses some shortcomings in AD, constraining the
predicted moments and improving the robustness of RR estimates to optimization
hyperparamters. Though our approach is not tied to a particular class of
learner, we illustrate it using neural networks, and evaluate on the problems
of average treatment/derivative effect estimation using semi-synthetic data.
Our numerical experiments show improved performance versus state of the art
benchmarks.
| [
{
"version": "v1",
"created": "Sun, 29 Sep 2024 20:56:54 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 03:00:35 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Hines",
"Christian L.",
""
],
[
"Hines",
"Oliver J.",
""
]
] |
2409.19845 | Marco Aymone M. Aymone | Marco Aymone | Sign changes of the partial sums of a random multiplicative function
III: Average | 9 pages, v4: new examples and references added. Comments from the
referee | null | null | null | math.NT math.PR | http://creativecommons.org/licenses/by/4.0/ | Let $V(x)$ be the number of sign changes of the partial sums up to $x$, say
$M_f(x)$, of a Rademacher random multiplicative function $f$. We prove that the
averaged value of $V(x)$ is at least $\gg (\log x)(\log\log
x)^{-1/2-\epsilon}$. Our new method applies for the counting of sign changes of
the partial sums of a system of orthogonal random variables having variance $1$
under additional hypothesis on the moments of these partial sums. In
particular, we extend to larger classes of dependencies an old result of
Erd\H{o}s and Hunt on sign changes of partial sums of i.i.d. random variables.
In the arithmetic case, the main input in our method is the
``\textit{linearity}'' phase in $1\leq q\leq 1.9$ of the quantity $\log
\mathbb{E} |M_f(x)|^q$, provided by the Harper's \textit{better than squareroot
cancellation} phenomenon for small moments of $M_f(x)$.
| [
{
"version": "v1",
"created": "Mon, 30 Sep 2024 01:10:40 GMT"
},
{
"version": "v2",
"created": "Thu, 3 Oct 2024 16:19:50 GMT"
},
{
"version": "v3",
"created": "Wed, 23 Oct 2024 11:58:42 GMT"
},
{
"version": "v4",
"created": "Mon, 14 Apr 2025 16:38:55 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Aymone",
"Marco",
""
]
] |
2409.20203 | Rhombik Roy | Rhombik Roy and Ofir E. Alon | Assessing small accelerations using a bosonic Josephson junction | 26 pages, 11 figures | Phys. Rev. A 111, 043307 (2025) | 10.1103/PhysRevA.111.043307 | null | cond-mat.quant-gas quant-ph | http://creativecommons.org/licenses/by/4.0/ | Bosonic Josephson junctions provide a versatile platform for exploring
quantum tunneling and coherence phenomena in ultracold atomic systems. While
extensive research has examined the Josephson-junction dynamics in various
double-well configurations, most studies have been limited to inertial
reference frames. In the present work, we posed the question how placing a
Josephson junction in a non-inertial reference frame would impact the quantum
tunnelling. Our findings demonstrate that accelerating a Josephson junction
alters the tunneling dynamics. Conversely, tunneling behavior can be used to
assess the acceleration of the system. By analyzing the changes in physical
properties, we can assess the acceleration of the double-well. We begin with
the most simple non-inertial frame: moving with constant acceleration. The
tunneling time decreases exponentially as acceleration increases, making it
effective for measuring larger accelerations. However, for smaller
accelerations, accurate assessment requires accounting for many-body depletion,
which decreases linearly as acceleration rises. Next, we explore a more complex
scenario where the acceleration is time dependent. In this case, the
acceleration is mapped onto the tunneling time period and depletion, which
again serve as predictors of acceleration. We go further by conducting a
detailed analysis of the change in tunnelling dynamics when the system deviates
from constant or zero acceleration. The quantitative analysis show that the
depletion changes exponentially near constant acceleration, while around zero
acceleration, the change follows a polynomial pattern. All in all, we quantify
how the tunneling process, as well as the mean-field and many-body properties,
evolve in a non-inertial system of increasing complexity.
| [
{
"version": "v1",
"created": "Mon, 30 Sep 2024 11:33:05 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Roy",
"Rhombik",
""
],
[
"Alon",
"Ofir E.",
""
]
] |
2409.20527 | Haoyang Wang | Haoyang Wang (1), Haoran Guo (1), He Ba (1), Zhengxiong Li (2),
Lingfeng Tao (1) ((1) Oklahoma State University, (2) University of Colorado
Denver) | Bi-directional Momentum-based Haptic Feedback and Control System for
In-Hand Dexterous Telemanipulation | This work has been submitted to the IEEE for possible publication | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In-hand dexterous telemanipulation requires not only precise remote motion
control of the robot but also effective haptic feedback to the human operator
to ensure stable and intuitive interactions between them. Most existing haptic
devices for dexterous telemanipulation focus on force feedback and lack
effective torque rendering, which is essential for tasks involving object
rotation. While some torque feedback solutions in virtual reality
applications-such as those based on geared motors or mechanically coupled
actuators-have been explored, they often rely on bulky mechanical designs,
limiting their use in portable or in-hand applications. In this paper, we
propose a Bi-directional Momentum-based Haptic Feedback and Control (Bi-Hap)
system that utilizes a palm-sized momentum-actuated mechanism to enable
real-time haptic and torque feedback. The Bi-Hap system also integrates an
Inertial Measurement Unit (IMU) to extract the human's manipulation command to
establish a closed-loop learning-based telemanipulation framework. Furthermore,
an error-adaptive feedback strategy is introduced to enhance operator
perception and task performance in different error categories. Experimental
evaluations demonstrate that Bi-Hap achieved feedback capability with low
command following latency (Delay < 0.025 s) and highly accurate torque feedback
(RMSE < 0.010 Nm).
| [
{
"version": "v1",
"created": "Mon, 30 Sep 2024 17:26:23 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 21:34:44 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Wang",
"Haoyang",
""
],
[
"Guo",
"Haoran",
""
],
[
"Ba",
"He",
""
],
[
"Li",
"Zhengxiong",
""
],
[
"Tao",
"Lingfeng",
""
]
] |
2410.00318 | Hokin Deng | Haoran Sun, Qingying Gao, Haiyun Lyu, Dezhi Luo, Yijiang Li, Hokin
Deng | Probing Mechanical Reasoning in Large Vision Language Models | Published at the ICLR 2025 Workshop on Bidirectional Human-AI
Alignment (BiAlign) | null | null | null | cs.AI q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Mechanical reasoning is a hallmark of human intelligence, defined by its
ubiquitous yet irreplaceable role in human activities ranging from routine
tasks to civil engineering. Embedding machines with mechanical reasoning is
therefore an important step towards building human-level artificial
intelligence. Here, we leveraged 155 cognitive experiments to test the
understanding of system stability, gears and pulley systems, leverage
principle, inertia and motion, and fluid mechanics in 26 Vision Language Models
(VLMs). Results indicate that VLMs consistently perform worse than humans on
all domains, while demonstrate significant difficulty in reasoning about gear
systems and fluid mechanics. Notably, their performance on these tasks do not
improve as number of parameters increase, suggesting that current
attention-based architecture may fail to grasp certain underlying mechanisms
required for mechanical reasoning, particularly those pertaining to mental
simulations.
| [
{
"version": "v1",
"created": "Tue, 1 Oct 2024 01:33:10 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Feb 2025 05:47:39 GMT"
},
{
"version": "v3",
"created": "Sun, 13 Apr 2025 05:53:58 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Sun",
"Haoran",
""
],
[
"Gao",
"Qingying",
""
],
[
"Lyu",
"Haiyun",
""
],
[
"Luo",
"Dezhi",
""
],
[
"Li",
"Yijiang",
""
],
[
"Deng",
"Hokin",
""
]
] |
2410.00324 | Hokin Deng | Qingying Gao, Yijiang Li, Haiyun Lyu, Haoran Sun, Dezhi Luo, Hokin
Deng | Vision Language Models See What You Want but not What You See | Published at the ICLR 2025 Workshop on Bidirectional Human-AI
Alignment (BiAlign) | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Knowing others' intentions and taking others' perspectives are two core
components of human intelligence that are considered to be instantiations of
theory-of-mind. Infiltrating machines with these abilities is an important step
towards building human-level artificial intelligence. Here, to investigate
intentionality understanding and level-2 perspective-taking in Vision Language
Models (VLMs), we constructed the IntentBench and PerspectBench, which together
contains over 300 cognitive experiments grounded in real-world scenarios and
classic cognitive tasks. We found VLMs achieving high performance on
intentionality understanding but low performance on level-2 perspective-taking.
This suggests a potential dissociation between simulation-based and
theory-based theory-of-mind abilities in VLMs, highlighting the concern that
they are not capable of using model-based reasoning to infer others' mental
states. See $\href{https://growing-ai-like-a-child.github.io/}{Website}$
| [
{
"version": "v1",
"created": "Tue, 1 Oct 2024 01:52:01 GMT"
},
{
"version": "v2",
"created": "Fri, 13 Dec 2024 01:57:19 GMT"
},
{
"version": "v3",
"created": "Sun, 22 Dec 2024 07:13:52 GMT"
},
{
"version": "v4",
"created": "Thu, 13 Feb 2025 04:03:09 GMT"
},
{
"version": "v5",
"created": "Sun, 13 Apr 2025 05:41:27 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Gao",
"Qingying",
""
],
[
"Li",
"Yijiang",
""
],
[
"Lyu",
"Haiyun",
""
],
[
"Sun",
"Haoran",
""
],
[
"Luo",
"Dezhi",
""
],
[
"Deng",
"Hokin",
""
]
] |
2410.00332 | Hokin Deng | Dezhi Luo, Haiyun Lyu, Qingying Gao, Haoran Sun, Yijiang Li, Hokin
Deng | Vision Language Models Know Law of Conservation without Understanding
More-or-Less | Published at the ICLR 2025 Workshop on Bidirectional Human-AI
Alignment (BiAlign) | null | null | null | cs.AI q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Understanding law of conservation is a critical milestone in human cognitive
development considered to be supported by the apprehension of quantitative
concepts and the reversibility of operations. To assess whether this critical
component of human intelligence has emerged in Vision Language Models, we have
curated the ConserveBench, a battery of 365 cognitive experiments across four
dimensions of physical quantities: volume, solid quantity, length, and number.
The former two involve transformational tasks which require reversibility
understanding. The latter two involve non-transformational tasks which assess
quantity understanding. Surprisingly, we find that while Vision Language Models
are generally good at transformational tasks, they tend to fail at
non-transformational tasks. There is a dissociation between understanding the
reversibility of operations and understanding the concept of quantity, which
both are believed to be the cornerstones of understanding law of conservation
in humans. $\href{https://growing-ai-like-a-child.github.io/}{Website}$
| [
{
"version": "v1",
"created": "Tue, 1 Oct 2024 02:15:49 GMT"
},
{
"version": "v2",
"created": "Sat, 14 Dec 2024 02:38:07 GMT"
},
{
"version": "v3",
"created": "Sun, 22 Dec 2024 07:59:36 GMT"
},
{
"version": "v4",
"created": "Sun, 16 Feb 2025 03:35:10 GMT"
},
{
"version": "v5",
"created": "Sun, 13 Apr 2025 04:31:37 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Luo",
"Dezhi",
""
],
[
"Lyu",
"Haiyun",
""
],
[
"Gao",
"Qingying",
""
],
[
"Sun",
"Haoran",
""
],
[
"Li",
"Yijiang",
""
],
[
"Deng",
"Hokin",
""
]
] |
2410.00703 | Jun Zhou | Zhexuan Zeng, Jun Zhou, Yasen Wang, and Zuowei Ping | Koopman Spectral Analysis from Noisy Measurements based on Bayesian
Learning and Kalman Smoothing | null | null | null | null | eess.SY cs.SY math.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Koopman spectral analysis plays a crucial role in understanding and modeling
nonlinear dynamical systems as it reveals key system behaviors and long-term
dynamics. However, the presence of measurement noise poses a significant
challenge to accurately extracting spectral properties. In this work, we
propose a robust method for identifying the Koopman operator and extracting its
spectral characteristics in noisy environments. To address the impact of noise,
our approach tackles an identification problem that accounts for both
systematic errors from finite-dimensional approximations and measurement noise
in the data. By incorporating Bayesian learning and Kalman smoothing, the
method simultaneously identifies the Koopman operator and estimates system
states, effectively decoupling these two error sources. The method's efficiency
and robustness are demonstrated through extensive experiments, showcasing its
accuracy across varying noise levels.
| [
{
"version": "v1",
"created": "Tue, 1 Oct 2024 13:55:25 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 02:20:34 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zeng",
"Zhexuan",
""
],
[
"Zhou",
"Jun",
""
],
[
"Wang",
"Yasen",
""
],
[
"Ping",
"Zuowei",
""
]
] |
2410.00719 | James P. Kneller | James P. Kneller, Julien Froustey, Evan B. Grohs, Francois Foucart,
Gail C. McLaughlin, Sherwood Richers | Quantum Closures for Neutrino Moment Transport | 22 pages, 10 figures. Matches published version | Phys. Rev. D 111, 063046 (2025) | 10.1103/PhysRevD.111.063046 | N3AS-24-035 | hep-ph astro-ph.HE nucl-th | http://creativecommons.org/licenses/by/4.0/ | A computationally efficient method for calculating the transport of neutrino
flavor in simulations is to use angular moments of the neutrino one-body
reduced density matrix, i.e., `quantum moments'. As with any moment-based
radiation transport method, a closure is needed if the infinite tower of moment
evolution equations is truncated. We derive a general parameterization of a
quantum closure and the limits the parameters must satisfy in order for the
closure to be physical. We then derive from multi-angle calculations the
evolution of the closure parameters in two test cases which we then
progressively insert into a moment evolution code and show how the parameters
affect the moment results until the full multi-angle results are reproduced.
This parameterization paves the way to setting prescriptions for genuine
quantum closures adapted to neutrino transport in a range of situations.
| [
{
"version": "v1",
"created": "Tue, 1 Oct 2024 14:09:58 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Feb 2025 14:17:55 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Kneller",
"James P.",
""
],
[
"Froustey",
"Julien",
""
],
[
"Grohs",
"Evan B.",
""
],
[
"Foucart",
"Francois",
""
],
[
"McLaughlin",
"Gail C.",
""
],
[
"Richers",
"Sherwood",
""
]
] |
2410.00914 | Malika Golshan | Malika Golshan, Adrian E. Bayer | Massive $\nu$s through the CNN lens: interpreting the field-level
neutrino mass information in weak lensing | 10+5 pages, 5 figures, minor changes to reflect final version
accepted by JCAP | null | null | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern cosmological surveys probe the Universe deep into the nonlinear
regime, where massive neutrinos suppress cosmic structure. Traditional
cosmological analyses, which use the 2-point correlation function to extract
information, are no longer optimal in the nonlinear regime, and there is thus
much interest in extracting beyond-2-point information to improve constraints
on neutrino mass. Quantifying and interpreting the beyond-2-point information
is thus a pressing task. We study the field-level information in weak lensing
convergence maps using convolution neural networks. We find that the network
performance increases as higher source redshifts and smaller scales are
considered -- investigating up to a source redshift of 2.5 and $\ell_{\rm
max}\simeq10^4$ -- verifying that massive neutrinos leave a distinct effect on
weak lensing. However, the performance of the network significantly drops after
scaling out the 2-point information from the maps, implying that most of the
field-level information can be found in the 2-point correlation function alone.
We quantify these findings in terms of the likelihood ratio and also use
Integrated Gradient saliency maps to interpret which parts of the map the
network is learning the most from. We find that, in the absence of noise, the
network extracts a similar amount of information from the most overdense and
underdense regions. However, upon adding noise, the information in underdense
regions is distorted as noise disproportionately washes out void-like
structures.
| [
{
"version": "v1",
"created": "Tue, 1 Oct 2024 17:59:26 GMT"
},
{
"version": "v2",
"created": "Wed, 2 Oct 2024 14:23:10 GMT"
},
{
"version": "v3",
"created": "Fri, 11 Apr 2025 18:10:20 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Golshan",
"Malika",
""
],
[
"Bayer",
"Adrian E.",
""
]
] |
2410.01432 | Minsu Kim | Minsu Kim, Sanghyeok Choi, Taeyoung Yun, Emmanuel Bengio, Leo Feng,
Jarrid Rector-Brooks, Sungsoo Ahn, Jinkyoo Park, Nikolay Malkin, Yoshua
Bengio | Adaptive teachers for amortized samplers | ICLR 2025, 27 pages, 12 figures | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Amortized inference is the task of training a parametric model, such as a
neural network, to approximate a distribution with a given unnormalized density
where exact sampling is intractable. When sampling is implemented as a
sequential decision-making process, reinforcement learning (RL) methods, such
as generative flow networks, can be used to train the sampling policy.
Off-policy RL training facilitates the discovery of diverse, high-reward
candidates, but existing methods still face challenges in efficient
exploration. We propose to use an adaptive training distribution (the \teacher)
to guide the training of the primary amortized sampler (the \student). The
\teacher, an auxiliary behavior model, is trained to sample high-loss regions
of the \student and can generalize across unexplored modes, thereby enhancing
mode coverage by providing an efficient training curriculum. We validate the
effectiveness of this approach in a synthetic environment designed to present
an exploration challenge, two diffusion-based sampling tasks, and four
biochemical discovery tasks demonstrating its ability to improve sample
efficiency and mode coverage. Source code is available at
https://github.com/alstn12088/adaptive-teacher.
| [
{
"version": "v1",
"created": "Wed, 2 Oct 2024 11:33:13 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 05:09:11 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Kim",
"Minsu",
""
],
[
"Choi",
"Sanghyeok",
""
],
[
"Yun",
"Taeyoung",
""
],
[
"Bengio",
"Emmanuel",
""
],
[
"Feng",
"Leo",
""
],
[
"Rector-Brooks",
"Jarrid",
""
],
[
"Ahn",
"Sungsoo",
""
],
[
"Park",
"Jinkyoo",
""
],
[
"Malkin",
"Nikolay",
""
],
[
"Bengio",
"Yoshua",
""
]
] |
2410.02321 | Zikun Zhang | Zikun Zhang, Zixiang Chen, Quanquan Gu | Convergence of Score-Based Discrete Diffusion Models: A Discrete-Time
Analysis | 26 pages, 1 figure | The Thirteenth International Conference on Learning
Representations, 2025 | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion models have achieved great success in generating high-dimensional
samples across various applications. While the theoretical guarantees for
continuous-state diffusion models have been extensively studied, the
convergence analysis of the discrete-state counterparts remains under-explored.
In this paper, we study the theoretical aspects of score-based discrete
diffusion models under the Continuous Time Markov Chain (CTMC) framework. We
introduce a discrete-time sampling algorithm in the general state space $[S]^d$
that utilizes score estimators at predefined time points. We derive convergence
bounds for the Kullback-Leibler (KL) divergence and total variation (TV)
distance between the generated sample distribution and the data distribution,
considering both scenarios with and without early stopping under reasonable
assumptions. Notably, our KL divergence bounds are nearly linear in the
dimension $d$, aligning with state-of-the-art results for diffusion models. Our
convergence analysis employs a Girsanov-based method and establishes key
properties of the discrete score function, which are essential for
characterizing the discrete-time sampling process.
| [
{
"version": "v1",
"created": "Thu, 3 Oct 2024 09:07:13 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 07:56:07 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhang",
"Zikun",
""
],
[
"Chen",
"Zixiang",
""
],
[
"Gu",
"Quanquan",
""
]
] |
2410.02696 | Jos\'e Paz-Mart\'in | Jos\'e Paz-Mart\'in, Andreas Sch\"uller, Alexandra Bourgouin, Araceli
Gago-Arias, Diego M. Gonzalez-Casta\~no, Nicol\'as G\'omez-Fern\'andez, Juan
Pardo-Montero, Faustino G\'omez | Evaluation of the two-voltage method for parallel-plate ionization
chambers irradiated with pulsed beams | null | null | 10.1002/mp.17814 | null | physics.med-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Air-vented ionization chambers exposed to clinical radiation beams may suffer
from recombination during the drift of the charge carriers towards the
electrodes. Thus, dosimetry protocols recommend the use of a correction factor,
usually denominated saturation factor ($k_{\rm sat}$), to correct the
ionization chamber readout for the incomplete collection of charge. The
two-voltage method is the recommended methodology for the calculation of the
saturation factor, however, it is based on the early Boag model, which only
takes into account the presence of positive and negative ions in the ionization
chamber and does not account for the electric field screening or the free
electron contribution to the signal.
The numerical simulation shows a better agreement with the experimental data
than the current analytical theories in terms of charge collection efficiency.
The classical two-voltage method, systematically overestimates the saturation
factor, with differences increasing with dose per pulse also present at low
dose per pulse. These results may have implications for the dosimetry with
ionization chambers in therapy modalities that use a dose per pulse higher than
conventional radiotherapy such as intraoperative radiotherapy but also in
conventional dose per pulse for ionization chambers that suffer from
significant charge recombination.
| [
{
"version": "v1",
"created": "Thu, 3 Oct 2024 17:20:38 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 14:21:44 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Paz-Martín",
"José",
""
],
[
"Schüller",
"Andreas",
""
],
[
"Bourgouin",
"Alexandra",
""
],
[
"Gago-Arias",
"Araceli",
""
],
[
"Gonzalez-Castaño",
"Diego M.",
""
],
[
"Gómez-Fernández",
"Nicolás",
""
],
[
"Pardo-Montero",
"Juan",
""
],
[
"Gómez",
"Faustino",
""
]
] |
2410.02761 | Zhipei Xu | Zhipei Xu, Xuanyu Zhang, Runyi Li, Zecheng Tang, Qing Huang, Jian
Zhang | FakeShield: Explainable Image Forgery Detection and Localization via
Multi-modal Large Language Models | Accepted by ICLR 2025 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid development of generative AI is a double-edged sword, which not
only facilitates content creation but also makes image manipulation easier and
more difficult to detect. Although current image forgery detection and
localization (IFDL) methods are generally effective, they tend to face two
challenges: \textbf{1)} black-box nature with unknown detection principle,
\textbf{2)} limited generalization across diverse tampering methods (e.g.,
Photoshop, DeepFake, AIGC-Editing). To address these issues, we propose the
explainable IFDL task and design FakeShield, a multi-modal framework capable of
evaluating image authenticity, generating tampered region masks, and providing
a judgment basis based on pixel-level and image-level tampering clues.
Additionally, we leverage GPT-4o to enhance existing IFDL datasets, creating
the Multi-Modal Tamper Description dataSet (MMTD-Set) for training FakeShield's
tampering analysis capabilities. Meanwhile, we incorporate a Domain Tag-guided
Explainable Forgery Detection Module (DTE-FDM) and a Multi-modal Forgery
Localization Module (MFLM) to address various types of tamper detection
interpretation and achieve forgery localization guided by detailed textual
descriptions. Extensive experiments demonstrate that FakeShield effectively
detects and localizes various tampering techniques, offering an explainable and
superior solution compared to previous IFDL methods. The code is available at
https://github.com/zhipeixu/FakeShield.
| [
{
"version": "v1",
"created": "Thu, 3 Oct 2024 17:59:34 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Oct 2024 14:15:56 GMT"
},
{
"version": "v3",
"created": "Tue, 5 Nov 2024 13:14:23 GMT"
},
{
"version": "v4",
"created": "Sat, 12 Apr 2025 08:08:04 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Xu",
"Zhipei",
""
],
[
"Zhang",
"Xuanyu",
""
],
[
"Li",
"Runyi",
""
],
[
"Tang",
"Zecheng",
""
],
[
"Huang",
"Qing",
""
],
[
"Zhang",
"Jian",
""
]
] |
2410.03620 | Mauricio Rodr\'iguez-Mayorga | Mauricio Rodr\'iguez-Mayorga, Pierre-Fran\c{c}ois Loos, Fabien
Bruneval, and Lucas Visscher | Time-Reversal Symmetry in RDMFT and pCCD with Complex-Valued Orbitals | null | J. Chem. Phys. 162, 054716 (2025) | 10.1063/5.0242504 | null | physics.chem-ph quant-ph | http://creativecommons.org/licenses/by/4.0/ | Reduced density matrix functional theory (RDMFT) and coupled cluster theory
restricted to paired double excitations (pCCD) are emerging as efficient
methodologies for accounting for the so-called non-dynamic electronic
correlation effects. Up to now, molecular calculations have been performed with
real-valued orbitals. However, before extending the applicability of these
methodologies to extended systems, where Bloch states are employed, the
subtleties of working with complex-valued orbitals and the consequences of
imposing time-reversal symmetry must be carefully addressed. In this work, we
describe the theoretical and practical implications of adopting time-reversal
symmetry in RDMFT and pCCD when allowing for complex-valued orbital
coefficients. The theoretical considerations primarily affect the optimization
algorithms, while the practical implications raise fundamental questions about
the stability of solutions. Specifically, we find that complex solutions lower
the energy when non-dynamic electronic correlation effects are pronounced. We
present numerical examples to illustrate and discuss these instabilities and
possible problems introduced by N-representability violations.
| [
{
"version": "v1",
"created": "Fri, 4 Oct 2024 17:22:03 GMT"
},
{
"version": "v2",
"created": "Mon, 9 Dec 2024 17:33:47 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Rodríguez-Mayorga",
"Mauricio",
""
],
[
"Loos",
"Pierre-François",
""
],
[
"Bruneval",
"Fabien",
""
],
[
"Visscher",
"Lucas",
""
]
] |
2410.03974 | Jaemoo Choi | Milena Gazdieva, Jaemoo Choi, Alexander Kolesov, Jaewoong Choi, Petr
Mokrov, Alexander Korotin | Robust Barycenter Estimation using Semi-Unbalanced Neural Optimal
Transport | 30 pages, 7 figures | null | null | null | stat.ML cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Aggregating data from multiple sources can be formalized as an Optimal
Transport (OT) barycenter problem, which seeks to compute the average of
probability distributions with respect to OT discrepancies. However, in
real-world scenarios, the presence of outliers and noise in the data measures
can significantly hinder the performance of traditional statistical methods for
estimating OT barycenters. To address this issue, we propose a novel scalable
approach for estimating the robust continuous barycenter, leveraging the dual
formulation of the (semi-)unbalanced OT problem. To the best of our knowledge,
this paper is the first attempt to develop an algorithm for robust barycenters
under the continuous distribution setup. Our method is framed as a min-max
optimization problem and is adaptable to general cost functions. We rigorously
establish the theoretical underpinnings of the proposed method and demonstrate
its robustness to outliers and class imbalance through a number of illustrative
experiments. Our source code is publicly available at
https://github.com/milenagazdieva/U-NOTBarycenters.
| [
{
"version": "v1",
"created": "Fri, 4 Oct 2024 23:27:33 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 04:16:25 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Gazdieva",
"Milena",
""
],
[
"Choi",
"Jaemoo",
""
],
[
"Kolesov",
"Alexander",
""
],
[
"Choi",
"Jaewoong",
""
],
[
"Mokrov",
"Petr",
""
],
[
"Korotin",
"Alexander",
""
]
] |
2410.04026 | Ahmad Al-Dweik | Ahmad Y. Al-Dweik, Abdallah Sayyed-Ahmad | Efficient tensor-based approach to solving linear systems involving
Kronecker sum of matrices | null | null | null | null | math.GM | http://creativecommons.org/licenses/by/4.0/ | A novel tensor-based formula for solving the linear systems involving
Kronecker sum is proposed. Such systems are directly related to the matrix and
tensor forms of Sylvester equation. The new tensor-based formula demonstrates
the well-known fact that a Sylvester tensor equation has a unique solution if
the sum of spectra of the matrices does not contain zero. We have showcased the
effectiveness of the method by efficiently solving the 2D and 3D discretized
Poisson equations, as well as the 2D steady-state convection-diffusion
equation, on a rectangular domain with Dirichlet boundary conditions. The
results suggest that this approach is well-suited for high-dimensional
problems.
| [
{
"version": "v1",
"created": "Sat, 5 Oct 2024 04:06:08 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 04:06:39 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Al-Dweik",
"Ahmad Y.",
""
],
[
"Sayyed-Ahmad",
"Abdallah",
""
]
] |
2410.04116 | Alessio Zaccone | Alessio Zaccone | Thickness-dependent conductivity of nanometric semiconductor thin films | null | Phys. Rev. Materials 9, 046001 (2025) | 10.1103/PhysRevMaterials.9.046001 | null | cond-mat.mes-hall cond-mat.mtrl-sci cond-mat.soft cond-mat.stat-mech physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The miniaturization of electronic devices has led to the prominence, in
technological applications, of semiconductor thin films that are only a few
nanometers thick. In spite of intense research, the thickness-dependent
resistivity or conductivity of semiconductor thin films is not understood at a
fundamental physical level. We develop a theory based on quantum confinement
which yields the dependence of the concentration of intrinsic carriers on the
film thickness. The theory predicts that the resistivity $\rho$, in the 1-10 nm
thickness range, increases exponentially as $\rho \sim \exp(const/L^{1/2})$
upon decreasing the film thickness $L$. This law is able to reproduce the
remarkable increase in resistivity observed experimentally in Si thin films,
whereas the effect of surface scattering (Fuchs-Sondheimer theory) alone cannot
explain the data when the film thickness is lower than 10 nm.
| [
{
"version": "v1",
"created": "Sat, 5 Oct 2024 10:49:03 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 16:33:57 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zaccone",
"Alessio",
""
]
] |
2410.04601 | Seungjun Yi | Seungjun Yi, Jaeyoung Lim, Juyong Yoon | ProtoMed-LLM: An Automatic Evaluation Framework for Large Language
Models in Medical Protocol Formulation | Oral Presentation at the 2025 Conference on Health IT and Analytics
(CHITA 2025) | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Automated generation of scientific protocols executable by robots can
significantly accelerate scientific research processes. Large Language Models
(LLMs) excel at Scientific Protocol Formulation Tasks (SPFT), but the
evaluation of their capabilities rely on human evaluation. Here, we propose a
flexible, automatic framework to evaluate LLMs' capability on SPFT:
ProtoMed-LLM. This framework prompts the target model and GPT-4 to extract
pseudocode from biology protocols using only predefined lab actions and
evaluates the output of the target model using LLAM-EVAL, the pseudocode
generated by GPT-4 serving as a baseline and Llama-3 acting as the evaluator.
Our adaptable prompt-based evaluation method, LLAM-EVAL, offers significant
flexibility in terms of evaluation model, material, criteria, and is free of
cost. We evaluate GPT variations, Llama, Mixtral, Gemma, Cohere, and Gemini.
Overall, we find that GPT and Cohere are powerful scientific protocol
formulators. We also introduce BIOPROT 2.0, a dataset with biology protocols
and corresponding pseudocodes, which can aid LLMs in formulation and evaluation
of SPFT. Our work is extensible to assess LLMs on SPFT across various domains
and other fields that require protocol generation for specific goals.
| [
{
"version": "v1",
"created": "Sun, 6 Oct 2024 19:28:55 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 20:35:36 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Yi",
"Seungjun",
""
],
[
"Lim",
"Jaeyoung",
""
],
[
"Yoon",
"Juyong",
""
]
] |
2410.05168 | Yuelyu Ji | Yuelyu Ji, Zhuochun Li, Rui Meng, and Daqing He | ReasoningRank: Teaching Student Models to Rank through Reasoning-Based
Knowledge Distillation | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Reranking documents based on their relevance to a given query is a critical
task in information retrieval. Traditional reranking methods often lack
transparency and rely on proprietary models, hindering reproducibility and
interpretability. We propose Reason-to-Rank (R2R), a novel open-source
reranking approach that enhances transparency by generating two types of
reasoning: direct relevance reasoning, which explains how a document addresses
the query, and comparison reasoning, which justifies the relevance of one
document over another. We leverage large language models (LLMs) as teacher
models to generate these explanations and distill this knowledge into smaller,
openly available student models. Our student models are trained to generate
meaningful reasoning and rerank documents, achieving competitive performance
across multiple datasets, including MSMARCO and BRIGHT. Experiments demonstrate
that R2R not only improves reranking accuracy but also provides valuable
insights into the decision-making process. By offering a structured and
interpretable solution with openly accessible resources, R2R aims to bridge the
gap between effectiveness and transparency in information retrieval, fostering
reproducibility and further research in the field.
| [
{
"version": "v1",
"created": "Mon, 7 Oct 2024 16:25:39 GMT"
},
{
"version": "v2",
"created": "Sun, 17 Nov 2024 17:26:23 GMT"
},
{
"version": "v3",
"created": "Sat, 30 Nov 2024 16:10:21 GMT"
},
{
"version": "v4",
"created": "Mon, 14 Apr 2025 14:58:01 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Ji",
"Yuelyu",
""
],
[
"Li",
"Zhuochun",
""
],
[
"Meng",
"Rui",
""
],
[
"He",
"Daqing",
""
]
] |
2410.05260 | Kaifeng Zhao | Kaifeng Zhao, Gen Li, Siyu Tang | DartControl: A Diffusion-Based Autoregressive Motion Model for Real-Time
Text-Driven Motion Control | Updated ICLR camera ready version | null | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-conditioned human motion generation, which allows for user interaction
through natural language, has become increasingly popular. Existing methods
typically generate short, isolated motions based on a single input sentence.
However, human motions are continuous and can extend over long periods,
carrying rich semantics. Creating long, complex motions that precisely respond
to streams of text descriptions, particularly in an online and real-time
setting, remains a significant challenge. Furthermore, incorporating spatial
constraints into text-conditioned motion generation presents additional
challenges, as it requires aligning the motion semantics specified by text
descriptions with geometric information, such as goal locations and 3D scene
geometry. To address these limitations, we propose DartControl, in short DART,
a Diffusion-based Autoregressive motion primitive model for Real-time
Text-driven motion control. Our model effectively learns a compact motion
primitive space jointly conditioned on motion history and text inputs using
latent diffusion models. By autoregressively generating motion primitives based
on the preceding history and current text input, DART enables real-time,
sequential motion generation driven by natural language descriptions.
Additionally, the learned motion primitive space allows for precise spatial
motion control, which we formulate either as a latent noise optimization
problem or as a Markov decision process addressed through reinforcement
learning. We present effective algorithms for both approaches, demonstrating
our model's versatility and superior performance in various motion synthesis
tasks. Experiments show our method outperforms existing baselines in motion
realism, efficiency, and controllability. Video results are available on the
project page: https://zkf1997.github.io/DART/.
| [
{
"version": "v1",
"created": "Mon, 7 Oct 2024 17:58:22 GMT"
},
{
"version": "v2",
"created": "Sun, 2 Mar 2025 10:58:06 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 10:43:50 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhao",
"Kaifeng",
""
],
[
"Li",
"Gen",
""
],
[
"Tang",
"Siyu",
""
]
] |
2410.05431 | Martin Andrae | Martin Andrae, Tomas Landelius, Joel Oskarsson, Fredrik Lindsten | Continuous Ensemble Weather Forecasting with Diffusion models | 25 pages, 17 figures. Code is available at
https://github.com/martinandrae/Continuous-Ensemble-Forecasting | null | null | null | cs.LG physics.ao-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Weather forecasting has seen a shift in methods from numerical simulations to
data-driven systems. While initial research in the area focused on
deterministic forecasting, recent works have used diffusion models to produce
skillful ensemble forecasts. These models are trained on a single forecasting
step and rolled out autoregressively. However, they are computationally
expensive and accumulate errors for high temporal resolution due to the many
rollout steps. We address these limitations with Continuous Ensemble
Forecasting, a novel and flexible method for sampling ensemble forecasts in
diffusion models. The method can generate temporally consistent ensemble
trajectories completely in parallel, with no autoregressive steps. Continuous
Ensemble Forecasting can also be combined with autoregressive rollouts to yield
forecasts at an arbitrary fine temporal resolution without sacrificing
accuracy. We demonstrate that the method achieves competitive results for
global weather forecasting with good probabilistic properties.
| [
{
"version": "v1",
"created": "Mon, 7 Oct 2024 18:51:23 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 17:30:20 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Andrae",
"Martin",
""
],
[
"Landelius",
"Tomas",
""
],
[
"Oskarsson",
"Joel",
""
],
[
"Lindsten",
"Fredrik",
""
]
] |
2410.06580 | Ramesh Johari | Ramesh Johari, Hannah Li, Anushka Murthy, Gabriel Y. Weintraub | When Does Interference Matter? Decision-Making in Platform Experiments | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | This paper investigates decision-making in A/B experiments for online
platforms and marketplaces. In such settings, due to constraints on inventory,
A/B experiments typically lead to biased estimators because of interference;
this phenomenon has been well studied in recent literature. By contrast, there
has been relatively little discussion of the impact of interference on
decision-making. In this paper, we analyze a benchmark Markovian model of an
inventory-constrained platform, where arriving customers book listings that are
limited in supply; our analysis builds on a self-contained analysis of general
A/B experiments for Markov chains. We focus on the commonly used frequentist
hypothesis testing approach for making launch decisions based on data from
customer-randomized experiments, and we study the impact of interference on (1)
false positive probability and (2) statistical power.
We obtain three main findings. First, we show that for monotone treatments --
i.e., those where the treatment changes booking probabilities in the same
direction relative to control in all states -- the false positive probability
of the na\"ive difference-in-means estimator with classical variance estimation
is correctly controlled. This result stems from a novel analysis of A/A
experiments with arbitrary dependence structures. Second, we demonstrate that
for monotone treatments, when the state space is large, the statistical power
of this na\"ive approach is higher than that of any similar pipeline using a
debiased estimator. Taken together, these two findings suggest that platforms
may be better off not debiasing when treatments are monotone. Finally, we
numerically investigate false positive probability and statistical power when
treatments are non-monotone, and we show that the performance of the na\"ive
approach can be arbitrarily worse than a debiased approach in such cases.
| [
{
"version": "v1",
"created": "Wed, 9 Oct 2024 06:25:25 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Oct 2024 04:29:16 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 16:31:00 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Johari",
"Ramesh",
""
],
[
"Li",
"Hannah",
""
],
[
"Murthy",
"Anushka",
""
],
[
"Weintraub",
"Gabriel Y.",
""
]
] |
2410.06604 | Lianbao Jia | Lian-Bao Jia | A try for dark energy in quantum field theory: The vacuum energy of
neutrino field | 4 pages, 3 figures, discussions improved | null | null | null | hep-ph astro-ph.CO gr-qc | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The quartic-divergent vacuum energy poses an ultraviolet (UV) challenge (the
cosmological constant problem) in probing the nature of dark energy. Here we
try to evaluate the contribution of the vacuum energy to dark energy with a
method of the UV-free scheme. The result indicates that it is not a problem in
the UV region but a question of the contributions of heavy fields being
suppressed. Then, we propose an effective description via scale decoupling. The
parameter spaces suggest that the active vacuum energy of neutrino fields can
naturally meet the observation of dark energy density, and a neutrino with a
typical mass $\sim$ 10 meV is expected. The normal ordering neutrinos are
preferred by naturalness, and the neutrino mass window set by dark energy is
6.3 meV $\lesssim m_1 \lesssim$ 16.3 meV, 10.7 meV $\lesssim m_2 \lesssim$ 18.4
meV, 50.5 meV $\lesssim m_3 \lesssim$ 52.7 meV.
| [
{
"version": "v1",
"created": "Wed, 9 Oct 2024 06:55:58 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Oct 2024 04:08:56 GMT"
},
{
"version": "v3",
"created": "Wed, 27 Nov 2024 09:05:26 GMT"
},
{
"version": "v4",
"created": "Mon, 14 Apr 2025 07:34:05 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Jia",
"Lian-Bao",
""
]
] |
2410.06900 | Hajime Kawahara | Hajime Kawahara, Yui Kawashima, Shotaro Tada, Hiroyuki Tako Ishikawa,
Ko Hosokawa, Yui Kasagi, Takayuki Kotani, Kento Masuda, Stevanus Nuguroho,
Motohide Tamura, Hibiki Yama, Daniel Kitzmann, Nicolas Minesi, Brett M.
Morris | Differentiable Modeling of Planet and Substellar Atmosphere:
High-Resolution Emission, Transmission, and Reflection Spectroscopy with
ExoJAX2 | 36 pages, 15 figures, accepted by ApJ. ExoJAX2 (version 2.0) has been
released, available at https://github.com/HajimeKawahara/exojax | null | null | null | astro-ph.EP astro-ph.IM astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | Modeling based on differentiable programming holds great promise for
astronomy, enabling advanced techniques such as gradient-based posterior
sampling and optimization. This paradigm motivated us to develop ExoJAX
(Kawahara et al. 2022), the first auto-differentiable spectrum model of
exoplanets and brown dwarfs. ExoJAX directly calculates cross-sections as
functions of temperature and pressure to minimize interpolation errors in
high-dispersion spectra, although initial work focused on narrowband emission
spectroscopy. Here, we introduce a fast, memory-efficient opacity algorithm and
differentiable radiative transfer for emission, transmission, and reflection
spectroscopy. In the era of data-rich JWST observations, retrieval analyses are
often forced to bin high-resolution spectra due to computational bottlenecks.
The new algorithm efficiently handles native-resolution data, preserving the
full information content and dynamic range. The advances proposed in this paper
enable broader applications, demonstrated by retrievals of GL229 B's
high-dispersion emission, WASP-39 b's JWST mid-resolution transmission at
original resolution (R $\sim$ 2,700), and Jupiter's reflection spectrum. We
derive a C/O ratio for GL229 B consistent with its host star, constrain WASP-39
b's radial velocity from molecular line structures, and infer Jupiter's
metallicity in line with previous estimates.
| [
{
"version": "v1",
"created": "Wed, 9 Oct 2024 14:02:42 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 13:36:20 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Apr 2025 08:23:03 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Kawahara",
"Hajime",
""
],
[
"Kawashima",
"Yui",
""
],
[
"Tada",
"Shotaro",
""
],
[
"Ishikawa",
"Hiroyuki Tako",
""
],
[
"Hosokawa",
"Ko",
""
],
[
"Kasagi",
"Yui",
""
],
[
"Kotani",
"Takayuki",
""
],
[
"Masuda",
"Kento",
""
],
[
"Nuguroho",
"Stevanus",
""
],
[
"Tamura",
"Motohide",
""
],
[
"Yama",
"Hibiki",
""
],
[
"Kitzmann",
"Daniel",
""
],
[
"Minesi",
"Nicolas",
""
],
[
"Morris",
"Brett M.",
""
]
] |
2410.07177 | Hanrong Ye | Hanrong Ye, Haotian Zhang, Erik Daxberger, Lin Chen, Zongyu Lin,
Yanghao Li, Bowen Zhang, Haoxuan You, Dan Xu, Zhe Gan, Jiasen Lu, Yinfei Yang | MM-Ego: Towards Building Egocentric Multimodal LLMs for Video QA | Accepted by ICLR 2025 | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This research aims to comprehensively explore building a multimodal
foundation model for egocentric video understanding. To achieve this goal, we
work on three fronts. First, as there is a lack of QA data for egocentric video
understanding, we automatically generate 7M high-quality QA samples for
egocentric videos ranging from 30 seconds to one hour long in Ego4D based on
human-annotated data. This is one of the largest egocentric QA datasets.
Second, we contribute a challenging egocentric QA benchmark with 629 videos and
7,026 questions to evaluate the models' ability in recognizing and memorizing
visual details across videos of varying lengths. We introduce a new de-biasing
evaluation method to help mitigate the unavoidable language bias present in the
models being evaluated. Third, we propose a specialized multimodal architecture
featuring a novel "Memory Pointer Prompting" mechanism. This design includes a
\textit{global glimpse} step to gain an overarching understanding of the entire
video and identify key visual information, followed by a fallback step that
utilizes the key visual information to generate responses. This enables the
model to more effectively comprehend extended video content. With the data,
benchmark, and model, we build MM-Ego, an egocentric multimodal LLM that shows
powerful performance on egocentric video understanding.
| [
{
"version": "v1",
"created": "Wed, 9 Oct 2024 17:59:59 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 12:27:56 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Ye",
"Hanrong",
""
],
[
"Zhang",
"Haotian",
""
],
[
"Daxberger",
"Erik",
""
],
[
"Chen",
"Lin",
""
],
[
"Lin",
"Zongyu",
""
],
[
"Li",
"Yanghao",
""
],
[
"Zhang",
"Bowen",
""
],
[
"You",
"Haoxuan",
""
],
[
"Xu",
"Dan",
""
],
[
"Gan",
"Zhe",
""
],
[
"Lu",
"Jiasen",
""
],
[
"Yang",
"Yinfei",
""
]
] |
2410.07241 | Daniel Durnford | E. Alfonso-Pita, E. Behnke, M. Bressler, B. Broerman, K. Clark, R.
Coppejans, J. Corbett, M. Crisler, C. E. Dahl, K. Dering, A. de St. Croix, D.
Durnford, P. Giampa, J. Hall, O. Harris, H. Hawley-Herrera, N. Lamb, M.
Laurin, I. Levine, W. H. Lippincott, R. Neilson, M.-C. Piro, D. Pyda, Z.
Sheng, G. Sweeney, E. V\'azquez-J\'auregui, S. Westerdale, T. J. Whitis, A.
Wright, J. Zhang, R. Zhang, A. Zu\~niga-Reyes | Low-threshold response of a scintillating xenon bubble chamber to
nuclear and electronic recoils | 11 pages, 7 figures | Phys. Rev. D 111, 032002 (2025) | 10.1103/PhysRevD.111.032002 | null | physics.ins-det hep-ex nucl-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A device filled with pure xenon first demonstrated the ability to operate
simultaneously as a bubble chamber and scintillation detector in 2017. Initial
results from data taken at thermodynamic thresholds down to ~4 keV showed
sensitivity to ~20 keV nuclear recoils with no observable bubble nucleation by
$\gamma$-ray interactions. This paper presents results from further operation
of the same device at thermodynamic thresholds as low as 0.50 keV, hardware
limited. The bubble chamber has now been shown to have sensitivity to ~1 keV
nuclear recoils while remaining insensitive to bubble nucleation by
$\gamma$-rays. A robust calibration of the chamber's nuclear recoil nucleation
response, as a function of nuclear recoil energy and thermodynamic state, is
presented. Stringent upper limits are established for the probability of bubble
nucleation by $\gamma$-ray-induced Auger cascades, with a limit of
$<1.1\times10^{-6}$ set at 0.50 keV, the lowest thermodynamic threshold
explored.
| [
{
"version": "v1",
"created": "Mon, 7 Oct 2024 21:17:22 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Feb 2025 00:21:49 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Alfonso-Pita",
"E.",
""
],
[
"Behnke",
"E.",
""
],
[
"Bressler",
"M.",
""
],
[
"Broerman",
"B.",
""
],
[
"Clark",
"K.",
""
],
[
"Coppejans",
"R.",
""
],
[
"Corbett",
"J.",
""
],
[
"Crisler",
"M.",
""
],
[
"Dahl",
"C. E.",
""
],
[
"Dering",
"K.",
""
],
[
"Croix",
"A. de St.",
""
],
[
"Durnford",
"D.",
""
],
[
"Giampa",
"P.",
""
],
[
"Hall",
"J.",
""
],
[
"Harris",
"O.",
""
],
[
"Hawley-Herrera",
"H.",
""
],
[
"Lamb",
"N.",
""
],
[
"Laurin",
"M.",
""
],
[
"Levine",
"I.",
""
],
[
"Lippincott",
"W. H.",
""
],
[
"Neilson",
"R.",
""
],
[
"Piro",
"M. -C.",
""
],
[
"Pyda",
"D.",
""
],
[
"Sheng",
"Z.",
""
],
[
"Sweeney",
"G.",
""
],
[
"Vázquez-Jáuregui",
"E.",
""
],
[
"Westerdale",
"S.",
""
],
[
"Whitis",
"T. J.",
""
],
[
"Wright",
"A.",
""
],
[
"Zhang",
"J.",
""
],
[
"Zhang",
"R.",
""
],
[
"Zuñiga-Reyes",
"A.",
""
]
] |
2410.07580 | Meseret Asrat Demis\=e | Meseret Asrat | Kalb-Ramond field, black holes and black strings in (2 + 1)D | v4: typos corrected; no change on the results | null | null | null | hep-th gr-qc | http://creativecommons.org/licenses/by/4.0/ | New rotating dilaton black hole and black string solutions in three spacetime
dimensions are obtained. The background spacetime interpolates between Anti-de
Sitter and a (an asymptotically) flat spacetime. The new black strings are
characterized by their masses, angular momenta and axion charges. The uncharged
black string solutions have only a single horizon. They contain a curvature
singularity enclosed inside their horizons. Thus, they are also black hole
solutions of Einstein-scalar gravity. On the other hand, the charged black
string solutions have two horizons. They may or may not contain a curvature
singularity depending on the ratio of their inner and outer horizons radii.
When they contain a singularity, the singularity is either at or enclosed
inside their inner horizons. We also obtain, in addition to the uncharged black
strings, other new black hole solutions of Einstein-scalar gravity by applying
duality transformation on the new charged black strings. The new black holes
are described by their masses and angular momenta. We show that their angular
momenta do not depend on their horizons radii. We also find that the masses of
a class of the black holes can be made negative by adjusting, using the gauge
ambiguity, the asymptotic value of the (independent component of the)
Kalb-Ramond field in the dual black strings. We also discuss black hole and
black string solutions with a curvature singularity at or beyond their outer or
event horizons. Black hole solutions with a ring curvature singularity in
between their inner and outer horizons are also presented.
| [
{
"version": "v1",
"created": "Thu, 10 Oct 2024 03:29:11 GMT"
},
{
"version": "v2",
"created": "Tue, 7 Jan 2025 07:41:42 GMT"
},
{
"version": "v3",
"created": "Mon, 17 Mar 2025 08:28:03 GMT"
},
{
"version": "v4",
"created": "Mon, 14 Apr 2025 02:46:38 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Asrat",
"Meseret",
""
]
] |
2410.07942 | Cl\'ement de Seguins Pazzis | Cl\'ement de Seguins Pazzis | Spaces of triangularizable matrices | 39 pages (updated with a new section of the appendix) | null | null | null | math.RA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let F be a field. We investigate the greatest possible dimension t_n(F) for a
vector space of n-by-n matrices with entries in F and in which every element is
triangularizable over the ground field F. It is obvious that t_n(F) is greater
than or equal to n(n+1)/2, and we prove that equality holds if and only if F is
not quadratically closed or n=1, excluding finite fields with characteristic 2.
If F is infinite and not quadratically closed, we give an explicit description
of the solutions with the critical dimension t_n(F), reducing the problem to
the one of deciding for which integers k between 2 and n all k-by-k symmetric
matrices over F are triangularizable.
| [
{
"version": "v1",
"created": "Thu, 10 Oct 2024 14:08:30 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 08:27:06 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Pazzis",
"Clément de Seguins",
""
]
] |
2410.08437 | Daniel Bramblett | Rushang Karia, Daniel Bramblett, Daksh Dobhal, Siddharth Srivastava | Autonomous Evaluation of LLMs for Truth Maintenance and Reasoning Tasks | null | null | null | null | cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents AutoEval, a novel benchmark for scaling Large Language
Model (LLM) assessment in formal tasks with clear notions of correctness, such
as truth maintenance in translation and logical reasoning. AutoEval is the
first benchmarking paradigm that offers several key advantages necessary for
scaling objective evaluation of LLMs without human labeling: (a) ability to
evaluate LLMs of increasing sophistication by auto-generating tasks at
different levels of difficulty; (b) auto-generation of ground truth that
eliminates dependence on expensive and time-consuming human annotation; (c) the
use of automatically generated, randomized datasets that mitigate the ability
of successive LLMs to overfit to static datasets used in many contemporary
benchmarks. Empirical analysis shows that an LLM's performance on AutoEval is
highly indicative of its performance on a diverse array of other benchmarks
focusing on translation and reasoning tasks, making it a valuable autonomous
evaluation paradigm in settings where hand-curated datasets can be hard to
obtain and/or update.
| [
{
"version": "v1",
"created": "Fri, 11 Oct 2024 00:56:37 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 21:03:16 GMT"
},
{
"version": "v3",
"created": "Fri, 11 Apr 2025 20:44:16 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Karia",
"Rushang",
""
],
[
"Bramblett",
"Daniel",
""
],
[
"Dobhal",
"Daksh",
""
],
[
"Srivastava",
"Siddharth",
""
]
] |
2410.09014 | Julian Rey | Juan \'Alvarez Ruiz, Juli\'an Rey | Gravitational waves in ultra-slow-roll and their anisotropy at two loops | 34 pages, 4 figures. Figures corrected and discussion improved.
Version accepted for publication in JCAP | JCAP 04 (2025) 026 | 10.1088/1475-7516/2025/04/026 | DESY-24-149 | astro-ph.CO gr-qc hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We compute the non-Gaussian corrections to the energy density and
anisotropies of gravitational waves induced during the radiation era after an
ultra-slow-roll phase of inflation by using a diagrammatic approach, and
present the corresponding Feynman rules. Our two-loop calculation includes both
the intrinsic non-Gaussianity of the inflaton perturbation $\delta\phi$ and the
non-Gaussianity arising from the nonlinear relation between the latter and the
curvature perturbation $\mathcal{R}$, which we find to be subdominant with
respect to the former. We apply our formalism to an analytical model in which
the ultra-slow-roll phase is followed by a constant-roll stage with a
nonvanishing second slow-roll parameter $\eta$, and address the renormalization
of the one-loop scalar power spectrum in this scenario.
| [
{
"version": "v1",
"created": "Fri, 11 Oct 2024 17:30:02 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 08:24:46 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Ruiz",
"Juan Álvarez",
""
],
[
"Rey",
"Julián",
""
]
] |
2410.09303 | Buu Phan | Buu Phan, Brandon Amos, Itai Gat, Marton Havasi, Matthew Muckley,
Karen Ullrich | Exact Byte-Level Probabilities from Tokenized Language Models for
FIM-Tasks and Model Ensembles | null | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Tokenization is associated with many poorly understood shortcomings in
language models (LMs), yet remains an important component for long sequence
scaling purposes. This work studies how tokenization impacts model performance
by analyzing and comparing the stochastic behavior of tokenized models with
their byte-level, or token-free, counterparts. We discover that, even when the
two models are statistically equivalent, their predictive distributions over
the next byte can be substantially different, a phenomenon we term as
``tokenization bias''. To fully characterize this phenomenon, we introduce the
Byte-Token Representation Lemma, a framework that establishes a mapping between
the learned token distribution and its equivalent byte-level distribution. From
this result, we develop a next-byte sampling algorithm that eliminates
tokenization bias without requiring further training or optimization. In other
words, this enables zero-shot conversion of tokenized LMs into statistically
equivalent token-free ones. We demonstrate its broad applicability with two use
cases: fill-in-the-middle (FIM) tasks and model ensembles. In FIM tasks where
input prompts may terminate mid-token, leading to out-of-distribution
tokenization, our method mitigates performance degradation and achieves 18%
improvement in FIM coding benchmarks, while consistently outperforming the
standard token healing fix. For model ensembles where each model employs a
distinct vocabulary, our approach enables seamless integration, resulting in
improved performance up to 3.7% over individual models across various standard
baselines in reasoning, knowledge, and coding. Code is available at:
https://github.com/facebookresearch/Exact-Byte-Level-Probabilities-from-Tokenized-LMs
| [
{
"version": "v1",
"created": "Fri, 11 Oct 2024 23:30:42 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 20:53:36 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Phan",
"Buu",
""
],
[
"Amos",
"Brandon",
""
],
[
"Gat",
"Itai",
""
],
[
"Havasi",
"Marton",
""
],
[
"Muckley",
"Matthew",
""
],
[
"Ullrich",
"Karen",
""
]
] |
2410.09505 | Haoran Wang | Haoran Wang, Yaoru Sun, Zeshen Tang, Haibo Shi, Chenyuan Jiao | HG2P: Hippocampus-inspired High-reward Graph and Model-Free Q-Gradient
Penalty for Path Planning and Motion Control | null | null | null | null | cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Goal-conditioned hierarchical reinforcement learning (HRL) decomposes complex
reaching tasks into a sequence of simple subgoal-conditioned tasks, showing
significant promise for addressing long-horizon planning in large-scale
environments. This paper bridges the goal-conditioned HRL based on graph-based
planning to brain mechanisms, proposing a hippocampus-striatum-like
dual-controller hypothesis. Inspired by the brain mechanisms of organisms
(i.e., the high-reward preferences observed in hippocampal replay) and
instance-based theory, we propose a high-return sampling strategy for
constructing memory graphs, improving sample efficiency. Additionally, we
derive a model-free lower-level Q-function gradient penalty to resolve the
model dependency issues present in prior work, improving the generalization of
Lipschitz constraints in applications. Finally, we integrate these two
extensions, High-reward Graph and model-free Gradient Penalty (HG2P), into the
state-of-the-art framework ACLG, proposing a novel goal-conditioned HRL
framework, HG2P+ACLG. Experimentally, the results demonstrate that our method
outperforms state-of-the-art goal-conditioned HRL algorithms on a variety of
long-horizon navigation tasks and robotic manipulation tasks.
| [
{
"version": "v1",
"created": "Sat, 12 Oct 2024 11:46:31 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 18:55:06 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Wang",
"Haoran",
""
],
[
"Sun",
"Yaoru",
""
],
[
"Tang",
"Zeshen",
""
],
[
"Shi",
"Haibo",
""
],
[
"Jiao",
"Chenyuan",
""
]
] |
2410.09880 | Saeed Hassanpour | Shuai Jiang, Christina Robinson, Joseph Anderson, William Hisey, Lynn
Butterly, Arief Suriawinata, Saeed Hassanpour | Improving Colorectal Cancer Screening and Risk Assessment through
Predictive Modeling on Medical Images and Records | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Colonoscopy screening effectively identifies and removes polyps before they
progress to colorectal cancer (CRC), but current follow-up guidelines rely
primarily on histopathological features, overlooking other important CRC risk
factors. Variability in polyp characterization among pathologists also hinders
consistent surveillance decisions. Advances in digital pathology and deep
learning enable the integration of pathology slides and medical records for
more accurate CRC risk prediction. Using data from the New Hampshire
Colonoscopy Registry, including longitudinal follow-up, we adapted a
transformer-based model for histopathology image analysis to predict 5-year CRC
risk. We further explored multi-modal fusion strategies to combine clinical
records with deep learning-derived image features. Training the model to
predict intermediate clinical variables improved 5-year CRC risk prediction
(AUC = 0.630) compared to direct prediction (AUC = 0.615, p = 0.013).
Incorporating both imaging and non-imaging data, without requiring manual slide
review, further improved performance (AUC = 0.674) compared to traditional
features from colonoscopy and microscopy reports (AUC = 0.655, p = 0.001).
These results highlight the value of integrating diverse data modalities with
computational methods to enhance CRC risk stratification.
| [
{
"version": "v1",
"created": "Sun, 13 Oct 2024 15:39:53 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 19:21:04 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Jiang",
"Shuai",
""
],
[
"Robinson",
"Christina",
""
],
[
"Anderson",
"Joseph",
""
],
[
"Hisey",
"William",
""
],
[
"Butterly",
"Lynn",
""
],
[
"Suriawinata",
"Arief",
""
],
[
"Hassanpour",
"Saeed",
""
]
] |
2410.09913 | Kha Le Nhat | Kha Nhat Le, Hoang-Tuan Nguyen, Hung Tien Tran, Thanh Duc Ngo | Stratified Domain Adaptation: A Progressive Self-Training Approach for
Scene Text Recognition | [WACV 2025] 15 pages, 12 figures, 5 tables, include supplementary
materials, source code: https://github.com/KhaLee2307/StrDA | 2025 IEEE/CVF Winter Conference on Applications of Computer Vision
(WACV) (pp. 8990-9000) | 10.1109/WACV61041.2025.00871 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Unsupervised domain adaptation (UDA) has become increasingly prevalent in
scene text recognition (STR), especially where training and testing data reside
in different domains. The efficacy of existing UDA approaches tends to degrade
when there is a large gap between the source and target domains. To deal with
this problem, gradually shifting or progressively learning to shift from domain
to domain is the key issue. In this paper, we introduce the Stratified Domain
Adaptation (StrDA) approach, which examines the gradual escalation of the
domain gap for the learning process. The objective is to partition the training
data into subsets so that the progressively self-trained model can adapt to
gradual changes. We stratify the training data by evaluating the proximity of
each data sample to both the source and target domains. We propose a novel
method for employing domain discriminators to estimate the out-of-distribution
and domain discriminative levels of data samples. Extensive experiments on
benchmark scene-text datasets show that our approach significantly improves the
performance of baseline (source-trained) STR models.
| [
{
"version": "v1",
"created": "Sun, 13 Oct 2024 16:40:48 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Oct 2024 15:08:08 GMT"
},
{
"version": "v3",
"created": "Tue, 29 Oct 2024 10:45:05 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Le",
"Kha Nhat",
""
],
[
"Nguyen",
"Hoang-Tuan",
""
],
[
"Tran",
"Hung Tien",
""
],
[
"Ngo",
"Thanh Duc",
""
]
] |
2410.10212 | Jiajie Yu | Jiajie Yu, Yuhong Wang, Wei Ma | Large Language Model-Enhanced Reinforcement Learning for Generic Bus
Holding Control Strategies | 51 pages, 19 figures | null | null | null | cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bus holding control is a widely-adopted strategy for maintaining stability
and improving the operational efficiency of bus systems. Traditional
model-based methods often face challenges with the low accuracy of bus state
prediction and passenger demand estimation. In contrast, Reinforcement Learning
(RL), as a data-driven approach, has demonstrated great potential in
formulating bus holding strategies. RL determines the optimal control
strategies in order to maximize the cumulative reward, which reflects the
overall control goals. However, translating sparse and delayed control goals in
real-world tasks into dense and real-time rewards for RL is challenging,
normally requiring extensive manual trial-and-error. In view of this, this
study introduces an automatic reward generation paradigm by leveraging the
in-context learning and reasoning capabilities of Large Language Models (LLMs).
This new paradigm, termed the LLM-enhanced RL, comprises several LLM-based
modules: reward initializer, reward modifier, performance analyzer, and reward
refiner. These modules cooperate to initialize and iteratively improve the
reward function according to the feedback from training and test results for
the specified RL-based task. Ineffective reward functions generated by the LLM
are filtered out to ensure the stable evolution of the RL agents' performance
over iterations. To evaluate the feasibility of the proposed LLM-enhanced RL
paradigm, it is applied to extensive bus holding control scenarios that vary in
the number of bus lines, stops, and passenger demand. The results demonstrate
the superiority, generalization capability, and robustness of the proposed
paradigm compared to vanilla RL strategies, the LLM-based controller,
physics-based feedback controllers, and optimization-based controllers. This
study sheds light on the great potential of utilizing LLMs in various smart
mobility applications.
| [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 07:10:16 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 14:17:56 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Yu",
"Jiajie",
""
],
[
"Wang",
"Yuhong",
""
],
[
"Ma",
"Wei",
""
]
] |
2410.10486 | Laura Caravenna Prof | Mohamed Bentaibi, Laura Caravenna, Jean-Paul A. Gauthier and Francesco
Rossi | Consensus in Multiagent Systems under communication failure | null | null | null | null | math.OC cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | We consider multi-agent systems with cooperative interactions and study the
convergence to consensus in the case of time-dependent connections, with
possible communication failure.
We prove a new condition ensuring consensus: we define a graph in which
directed arrows correspond to connection functions that converge (in the weak
sense) to some function with a positive integral on all intervals of the form
$[t,+\infty)$. If the graph has a node reachable from all other indices,
i.e.~``globally reachable'', then the system converges to consensus. We show
that this requirement generalizes some known sufficient conditions for
convergence, such as Moreau's or the Persistent Excitation one. We also give a
second new condition, transversal to the known ones: total connectedness of the
undirected graph formed by the non-vanishing of limiting functions.
| [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 13:28:36 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 11:58:01 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Bentaibi",
"Mohamed",
""
],
[
"Caravenna",
"Laura",
""
],
[
"Gauthier",
"Jean-Paul A.",
""
],
[
"Rossi",
"Francesco",
""
]
] |
2410.10537 | Tom\'a\v{s} Jirsa | Jan Vrba, Jakub Steinbach, Tom\'a\v{s} Jirsa, Laura Verde, Roberta De
Fazio, Yuwen Zeng, Kei Ichiji, Luk\'a\v{s} H\'ajek, Zuzana Sedl\'akov\'a,
Zuzana Urb\'aniov\'a, Martin Chovanec, Jan Mare\v{s}, Noriyasu Homma | Reproducible Machine Learning-based Voice Pathology Detection:
Introducing the Pitch Difference Feature | Code repository:
https://github.com/aailab-uct/Automated-Robust-and-Reproducible-Voice-Pathology-Detection,
Supplementary materials: https://doi.org/10.5281/zenodo.14793017 | null | 10.1016/j.jvoice.2025.03.028 | null | cs.SD cs.AI cs.LG eess.AS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Purpose: We introduce a novel methodology for voice pathology detection using
the publicly available Saarbr\"ucken Voice Database (SVD) and a robust feature
set combining commonly used acoustic handcrafted features with two novel ones:
pitch difference (relative variation in fundamental frequency) and NaN feature
(failed fundamental frequency estimation).
Methods: We evaluate six machine learning (ML) algorithms -- support vector
machine, k-nearest neighbors, naive Bayes, decision tree, random forest, and
AdaBoost -- using grid search for feasible hyperparameters and 20480 different
feature subsets. Top 1000 classification models -- feature subset combinations
for each ML algorithm are validated with repeated stratified cross-validation.
To address class imbalance, we apply K-Means SMOTE to augment the training
data.
Results: Our approach achieves 85.61%, 84.69% and 85.22% unweighted average
recall (UAR) for females, males and combined results respectively. We
intentionally omit accuracy as it is a highly biased metric for imbalanced
data.
Conclusion: Our study demonstrates that by following the proposed methodology
and feature engineering, there is a potential in detection of various voice
pathologies using ML models applied to the simplest vocal task, a sustained
utterance of the vowel /a:/. To enable easier use of our methodology and to
support our claims, we provide a publicly available GitHub repository with DOI
10.5281/zenodo.13771573. Finally, we provide a REFORMS checklist to enhance
readability, reproducibility and justification of our approach
| [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 14:17:52 GMT"
},
{
"version": "v2",
"created": "Mon, 3 Feb 2025 12:57:09 GMT"
},
{
"version": "v3",
"created": "Fri, 14 Mar 2025 13:47:31 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Vrba",
"Jan",
""
],
[
"Steinbach",
"Jakub",
""
],
[
"Jirsa",
"Tomáš",
""
],
[
"Verde",
"Laura",
""
],
[
"De Fazio",
"Roberta",
""
],
[
"Zeng",
"Yuwen",
""
],
[
"Ichiji",
"Kei",
""
],
[
"Hájek",
"Lukáš",
""
],
[
"Sedláková",
"Zuzana",
""
],
[
"Urbániová",
"Zuzana",
""
],
[
"Chovanec",
"Martin",
""
],
[
"Mareš",
"Jan",
""
],
[
"Homma",
"Noriyasu",
""
]
] |
2410.10935 | Orlando Luongo | Youri Carloni, Orlando Luongo | Phase-space analysis of dark energy models in non-minimally coupled
theories of gravity | 30 pages, 9 figures, 7 tables | Class.Quant.Grav. 42 (2025) 7, 075014 | 10.1088/1361-6382/adc06e | null | gr-qc astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We analyze scalar field dark energy models minimally and non-minimally
coupled to gravity, postulating that a Yukawa-like interacting term is \emph{in
form} equivalent for general relativity, teleparallel and
symmetric-teleparallel theories. Our analysis is pursued within two scalar
field representations, where a quintessence and phantom pictures are associated
with quasiquintessence and quasiphantom exotic fields. In the latter, we
suggest how the phion-pressure can be built up without exhibiting a direct
kinetic term. Accordingly, the stability analysis reveals that this
quasiquintessence field provides a viable description of the universe
indicating, when minimally coupled, how to unify dark energy and dark matter by
showing an attractor point where $w_{\phi}=0$. Conversely, in the non-minimally
coupling, the alternative field only leaves an attractor where dark energy
dominates, mimicking \emph{de facto} a cosmological constant behavior. A direct
study is conducted comparing the standard case with the alternative one,
overall concluding that the behavior of quintessence is well established across
all the gravity scenarios. However, considering the phantom field non-minimal
coupled to gravity, the results are inconclusive for power-law potentials in
Einstein theory, and for the inverse square power (ISP) potential in both
teleparallel and symmetric-teleparallel theories. Finally, we study the growth
of matter perturbations and establish that only the fifth power and quadratic
potentials, when used to describe quasiphantom field minimally coupled to
gravity, exhibit behavior similar to the $\Lambda$CDM model.
| [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 17:58:29 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 17:04:06 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Carloni",
"Youri",
""
],
[
"Luongo",
"Orlando",
""
]
] |
2410.11333 | Yuzhong Cheng | Yuzhong Cheng and Hiroki Masuda | Statistical inference for ergodic diffusion with Markovian switching | 25 pages | null | null | null | math.ST stat.TH | http://creativecommons.org/licenses/by/4.0/ | This study explores a Gaussian quasi-likelihood approach for estimating
parameters of diffusion processes with Markovian regime switching. Assuming the
ergodicity under high-frequency sampling, we will show the asymptotic normality
of the unknown parameters contained in the drift and diffusion coefficients and
present a consistent explicit estimator for the generator of the Markov chain.
Simulation experiments are conducted to illustrate the theoretical results
obtained.
| [
{
"version": "v1",
"created": "Tue, 15 Oct 2024 07:01:40 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 04:56:23 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Cheng",
"Yuzhong",
""
],
[
"Masuda",
"Hiroki",
""
]
] |
2410.11525 | Katherine Newhall | Caitlyn Obrero and Mastawal Tirfe and Carmen Lee and Sourabh Saptarshi
and Christopher Rock and Karen E. Daniels and Katherine A. Newhall | Electrical Transport in Tunably-Disordered Metamaterials | 9 pages, 8 figures | null | null | null | cond-mat.dis-nn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Naturally occurring materials are often disordered, with their bulk
properties being challenging to predict from the structure, due to the lack of
underlying crystalline axes. In this paper, we develop a digital pipeline from
algorithmically-created configurations with tunable disorder to 3D printed
materials, as a tool to aid in the study of such materials, using electrical
resistance as a test case. The designed material begins with a random point
cloud that is iteratively evolved using Lloyd's algorithm to approach
uniformity, with the points being connected via a Delaunay triangulation to
form a disordered network metamaterial. Utilizing laser powder bed fusion
additive manufacturing with stainless steel 17-4 PH and titanium alloy
Ti-6Al-4V, we are able to experimentally measure the bulk electrical
resistivity of the disordered network. The effective resistance of the
structure calculated from the combinatorial weighted graph Laplacian is in good
agreement with experimental data. However, the effective resistance is
sensitive to anisotropy and global network topology, preventing a single
network statistic or disorder characterization from predicting global
resistivity.
| [
{
"version": "v1",
"created": "Tue, 15 Oct 2024 11:54:39 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 18:16:24 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Obrero",
"Caitlyn",
""
],
[
"Tirfe",
"Mastawal",
""
],
[
"Lee",
"Carmen",
""
],
[
"Saptarshi",
"Sourabh",
""
],
[
"Rock",
"Christopher",
""
],
[
"Daniels",
"Karen E.",
""
],
[
"Newhall",
"Katherine A.",
""
]
] |
2410.11709 | Nina Wiedemann | Nina Wiedemann, Th\'eo Uscidda, Martin Raubal | GeOT: A spatially explicit framework for evaluating spatio-temporal
predictions | null | null | null | null | cs.LG cs.CE | http://creativecommons.org/licenses/by/4.0/ | When predicting observations across space and time, the spatial layout of
errors impacts a model's real-world utility. For instance, in bike sharing
demand prediction, error patterns translate to relocation costs. However,
commonly used error metrics in GeoAI evaluate predictions point-wise,
neglecting effects such as spatial heterogeneity, autocorrelation, and the
Modifiable Areal Unit Problem. We put forward Optimal Transport (OT) as a
spatial evaluation metric and loss function. The proposed framework, called
GeOT, assesses the performance of prediction models by quantifying the
transport costs associated with their prediction errors. Through experiments on
real and synthetic data, we demonstrate that 1) the spatial distribution of
prediction errors relates to real-world costs in many applications, 2) OT
captures these spatial costs more accurately than existing metrics, and 3) OT
enhances comparability across spatial and temporal scales. Finally, we advocate
for leveraging OT as a loss function in neural networks to improve the spatial
accuracy of predictions. Experiments with bike sharing, charging station, and
traffic datasets show that spatial costs are significantly reduced with only
marginal changes to non-spatial error metrics. Thus, this approach not only
offers a spatially explicit tool for model evaluation and selection, but also
integrates spatial considerations into model training. All code is available at
https://github.com/mie-lab/geospatialOT.
| [
{
"version": "v1",
"created": "Tue, 15 Oct 2024 15:46:03 GMT"
},
{
"version": "v2",
"created": "Wed, 23 Oct 2024 15:35:57 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 10:54:19 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Wiedemann",
"Nina",
""
],
[
"Uscidda",
"Théo",
""
],
[
"Raubal",
"Martin",
""
]
] |
2410.11951 | Valentina La Torre | Valentina La Torre and Fabio Pacucci | HOLESOM: Constraining the Properties of Slowly-Accreting Massive Black
Holes with Self-Organizing Maps | 19 pages, 12 figures, 2 tables, accepted for publication on The
Astrophysical Journal | null | null | null | astro-ph.HE astro-ph.GA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Accreting massive black holes (MBHs, with M$_\bullet > 10^3$ M$_{\odot}$) are
known for their panchromatic emission, spanning from radio to gamma rays. While
MBHs accreting at significant fractions of their Eddington rate are readily
detectable, those accreting at much lower rates in radiatively inefficient
modes often go unnoticed, blending in with other astrophysical sources. This
challenge is particularly relevant for gas-starved MBHs in external galaxies
and those possibly wandering in the Milky Way. We present HOLESOM, a machine
learning-powered tool based on the Self-Organizing Maps (SOMs) algorithm,
specifically designed to identify slowly-accreting MBHs using sparse
photometric data. Trained on a comprehensive set of $\sim$ 20, 000 spectral
energy distributions (SEDs), HOLESOM can (i) determine if the photometry of a
source is consistent with slowly-accreting MBHs and (ii) estimate its black
hole mass and Eddington ratio, including uncertainties. We validate HOLESOM
through extensive tests on synthetic data and real-world cases, including
Sagittarius A* (Sgr A*), demonstrating its effectiveness in identifying
slowly-accreting MBHs. Additionally, we derive analytical relations between
radio and X-ray luminosities to further constrain physical parameters. The
primary strength of HOLESOM lies in its ability to accurately identify MBH
candidates, which can then be targeted for follow-up photometric and
spectroscopic observations. Fast and scalable, HOLESOM offers a robust
framework for automatically scanning large multi-wavelength datasets, making it
a valuable tool for unveiling hidden MBH populations in the local Universe.
| [
{
"version": "v1",
"created": "Tue, 15 Oct 2024 18:00:02 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Oct 2024 17:43:46 GMT"
},
{
"version": "v3",
"created": "Fri, 29 Nov 2024 23:40:25 GMT"
},
{
"version": "v4",
"created": "Mon, 14 Apr 2025 17:09:44 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"La Torre",
"Valentina",
""
],
[
"Pacucci",
"Fabio",
""
]
] |
2410.12385 | Yaqi Zhao | Yaqi Zhao, Jinchuan Hou, Kan He, Nicol\`o Lo Piparo, Xiangyi Meng | A nonconvex entanglement monotone determining the characteristic length
of entanglement distribution in continuous-variable quantum networks | null | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum networks (QNs) promise to enhance the performance of various quantum
technologies in the near future by distributing entangled states over long
distances. The first step towards this is to develop novel entanglement
measures that are both informative and computationally tractable at large
scales. While numerous such entanglement measures exist for discrete-variable
(DV) systems, a comprehensive exploration for experimentally preferred
continuous-variable (CV) systems is lacking. Here, we introduce a class of CV
entanglement measures, among which we identify a nonconvex entanglement
monotone -- the ratio negativity, which possesses a simple, scalable form that
determines the exponential decay of optimal entanglement swapping on a chain of
pure Gaussian states. This characterization opens avenues for leveraging
statistical physics tools to analyze swapping-protocol-based CV QNs.
| [
{
"version": "v1",
"created": "Wed, 16 Oct 2024 09:11:20 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 06:25:12 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhao",
"Yaqi",
""
],
[
"Hou",
"Jinchuan",
""
],
[
"He",
"Kan",
""
],
[
"Piparo",
"Nicolò Lo",
""
],
[
"Meng",
"Xiangyi",
""
]
] |
2410.12877 | Alessandro Stolfo | Alessandro Stolfo, Vidhisha Balachandran, Safoora Yousefi, Eric
Horvitz, Besmira Nushi | Improving Instruction-Following in Language Models through Activation
Steering | ICLR 2025 | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | The ability to follow instructions is crucial for numerous real-world
applications of language models. In pursuit of deeper insights and more
powerful capabilities, we derive instruction-specific vector representations
from language models and use them to steer models accordingly. These vectors
are computed as the difference in activations between inputs with and without
instructions, enabling a modular approach to activation steering. We
demonstrate how this method can enhance model adherence to constraints such as
output format, length, and word inclusion, providing inference-time control
over instruction following. Our experiments across four models demonstrate how
we can use the activation vectors to guide models to follow constraints even
without explicit instructions and to enhance performance when instructions are
present. Additionally, we explore the compositionality of activation steering,
successfully applying multiple instructions simultaneously. Finally, we
demonstrate that steering vectors computed on instruction-tuned models can
transfer to improve base models. Our findings demonstrate that activation
steering offers a practical and scalable approach for fine-grained control in
language generation. Our code and data are available at
https://github.com/microsoft/llm-steer-instruct.
| [
{
"version": "v1",
"created": "Tue, 15 Oct 2024 08:38:20 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 09:04:45 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Stolfo",
"Alessandro",
""
],
[
"Balachandran",
"Vidhisha",
""
],
[
"Yousefi",
"Safoora",
""
],
[
"Horvitz",
"Eric",
""
],
[
"Nushi",
"Besmira",
""
]
] |
2410.13053 | Toghrul Karimov | Toghrul Karimov | Verification of Linear Dynamical Systems via O-Minimality of the Real
Numbers | ICALP 2025 | null | null | null | cs.LO | http://creativecommons.org/licenses/by/4.0/ | A discrete-time linear dynamical system (LDS) is given by an update matrix $M
\in \mathbb{R}^{d\times d}$, and has the trajectories $\langle s, Ms, M^2s,
\ldots \rangle$ for $s \in \mathbb{R}^d$. Reachability-type decision problems
of linear dynamical systems, most notably the Skolem Problem, lie at the
forefront of decidability: typically, sound and complete algorithms are known
only in low dimensions, and these rely on sophisticated tools from number
theory and Diophantine approximation. Recently, however, o-minimality has
emerged as a counterpoint to these number-theoretic tools that allows us to
decide certain modifications of the classical problems of LDS without any
dimension restrictions. In this paper, we first introduce the Decomposition
Method, a framework that captures all applications of o-minimality to decision
problems of LDS that are currently known to us. We then use the Decomposition
Method to show decidability of the Robust Safety Problem (restricted to bounded
initial sets) in arbitrary dimension: given a matrix $M$, a bounded
semialgebraic set $S$ of initial points, and a semialgebraic set $T$ of unsafe
points, it is decidable whether there exists $\varepsilon > 0$ such that all
orbits that begin in the $\varepsilon$-ball around $S$ avoid $T$.
| [
{
"version": "v1",
"created": "Wed, 16 Oct 2024 21:30:03 GMT"
},
{
"version": "v2",
"created": "Mon, 13 Jan 2025 18:18:08 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 17:03:36 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Karimov",
"Toghrul",
""
]
] |
2410.13538 | Suneel Dutt | Abhinaba Upadhyay, Arvind Kumar, Harleen Dahiya, Suneel Dutt | Magnetic moments of $\frac{1}{2}^-$ baryon resonances in hot and dense
strange hadronic matter | 40 pages, 8 tables, 13 figures, Accepted for Publication in PTEP | null | null | null | hep-ph nucl-th | http://creativecommons.org/licenses/by/4.0/ | This work primarily focusses on determining the magnetic moments of
$\frac{1}{2}^-$ baryon resonances in the presence of hot and dense hadronic
matter. In the chiral $SU(3)$ quark mean field model approach, we have
essentially accounted for the effects on in-medium scalar meson fields to
investigate the impact of high densities on the in-medium baryon masses and
their constituent quarks. In light of chiral constituent quark model $\chi$CQM,
we have calculated the magnetic moments of $\frac{1}{2}^-$ baryon resonances
and scrutinized the effects due to its internal constituents: the valence
quarks, sea quarks and the orbital moment of sea quarks. Furthermore, we have
investigated the effective baryonic magnetic moments for the finite magnitudes
of isospin asymmetry and strangeness fraction.
| [
{
"version": "v1",
"created": "Thu, 17 Oct 2024 13:29:07 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 11:49:54 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Upadhyay",
"Abhinaba",
""
],
[
"Kumar",
"Arvind",
""
],
[
"Dahiya",
"Harleen",
""
],
[
"Dutt",
"Suneel",
""
]
] |
2410.14558 | Leonardo Antonio Navarro Labastida | Leonardo A. Navarro-Labastida | Thermal quantum information capacity in a topological insulator | null | null | null | null | quant-ph cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | Thermal effects in a one-dimensional Su-Schrieffer-Hegger (SSH) topological
insulator are studied. Particularly, we focus on quantum information processing
(QIP) capacity for thermal ensembles. To evaluate QIP an optimized quantum
Fisher information (OQFI) is introduced as a quantifier of entanglement and
topological phases are calculated by a definition in real space for the
electric polarization of mixture states. For the thermal ensemble, there is a
relationship between the Fisher metric and the electric polarization in such a
way that in the topological region, there is more entanglement, therefore,
these generate more robustness and protection in the quantum information due to
thermal effects. Also, long-range hopping effects are studied and it is found
that in this case, the OQFI captures these topological phase transitions in the
limit of low temperature by this formalism in real space.
| [
{
"version": "v1",
"created": "Fri, 18 Oct 2024 16:03:46 GMT"
},
{
"version": "v2",
"created": "Fri, 6 Dec 2024 21:01:18 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 17:58:19 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Navarro-Labastida",
"Leonardo A.",
""
]
] |
2410.14763 | Hamed Fayyaz | Hamed Fayyaz, Raphael Poulain, Rahmatollah Beheshti | Enabling Scalable Evaluation of Bias Patterns in Medical LLMs | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have shown impressive potential in helping with
numerous medical challenges. Deploying LLMs in high-stakes applications such as
medicine, however, brings in many concerns. One major area of concern relates
to biased behaviors of LLMs in medical applications, leading to unfair
treatment of individuals. To pave the way for the responsible and impactful
deployment of Med LLMs, rigorous evaluation is a key prerequisite. Due to the
huge complexity and variability of different medical scenarios, existing work
in this domain has primarily relied on using manually crafted datasets for bias
evaluation. In this study, we present a new method to scale up such bias
evaluations by automatically generating test cases based on rigorous medical
evidence. We specifically target the challenges of a) domain-specificity of
bias characterization, b) hallucinating while generating the test cases, and c)
various dependencies between the health outcomes and sensitive attributes. To
that end, we offer new methods to address these challenges integrated with our
generative pipeline, using medical knowledge graphs, medical ontologies, and
customized general LLM evaluation frameworks in our method. Through a series of
extensive experiments, we show that the test cases generated by our proposed
method can effectively reveal bias patterns in Med LLMs at larger and more
flexible scales than human-crafted datasets. We publish a large bias evaluation
dataset using our pipeline, which is dedicated to a few medical case studies. A
live demo of our application for vignette generation is available at
https://vignette.streamlit.app. Our code is also available at
https://github.com/healthylaife/autofair.
| [
{
"version": "v1",
"created": "Fri, 18 Oct 2024 14:17:03 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 03:42:38 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Fayyaz",
"Hamed",
""
],
[
"Poulain",
"Raphael",
""
],
[
"Beheshti",
"Rahmatollah",
""
]
] |
2410.14949 | Saptarshi Roy | Vansh Bansal, Saptarshi Roy, Purnamrita Sarkar, Alessandro Rinaldo | On the Wasserstein Convergence and Straightness of Rectified Flow | 38 pages | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Diffusion models have emerged as a powerful tool for image generation and
denoising. Typically, generative models learn a trajectory between the starting
noise distribution and the target data distribution. Recently Liu et al.
(2023b) proposed Rectified Flow (RF), a generative model that aims to learn
straight flow trajectories from noise to data using a sequence of convex
optimization problems with close ties to optimal transport. If the trajectory
is curved, one must use many Euler discretization steps or novel strategies,
such as exponential integrators, to achieve a satisfactory generation quality.
In contrast, RF has been shown to theoretically straighten the trajectory
through successive rectifications, reducing the number of function evaluations
(NFEs) while sampling. It has also been shown empirically that RF may improve
the straightness in two rectifications if one can solve the underlying
optimization problem within a sufficiently small error. In this paper, we make
two contributions. First, we provide a theoretical analysis of the Wasserstein
distance between the sampling distribution of RF and the target distribution.
Our error rate is characterized by the number of discretization steps and a
novel formulation of straightness stronger than that in the original work.
Secondly, we present general conditions guaranteeing uniqueness and
straightness of 1-RF, which is in line with previous empirical findings. As a
byproduct of our analysis, we show that, in one dimension, RF started at the
standard Gaussian distribution yields the Monge map. Additionally, we also
present empirical results on both simulated and real datasets to validate our
theoretical findings. The code is available at
https://github.com/bansal-vansh/rectified-flow.
| [
{
"version": "v1",
"created": "Sat, 19 Oct 2024 02:36:11 GMT"
},
{
"version": "v2",
"created": "Mon, 18 Nov 2024 15:35:52 GMT"
},
{
"version": "v3",
"created": "Fri, 6 Dec 2024 11:00:14 GMT"
},
{
"version": "v4",
"created": "Sat, 21 Dec 2024 07:12:20 GMT"
},
{
"version": "v5",
"created": "Sun, 23 Feb 2025 02:24:40 GMT"
},
{
"version": "v6",
"created": "Sun, 13 Apr 2025 06:13:45 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Bansal",
"Vansh",
""
],
[
"Roy",
"Saptarshi",
""
],
[
"Sarkar",
"Purnamrita",
""
],
[
"Rinaldo",
"Alessandro",
""
]
] |
2410.15038 | Siyuan Yan | Siyuan Yan, Zhen Yu, Clare Primiero, Cristina Vico-Alonso, Zhonghua
Wang, Litao Yang, Philipp Tschandl, Ming Hu, Lie Ju, Gin Tan, Vincent Tang,
Aik Beng Ng, David Powell, Paul Bonnington, Simon See, Elisabetta Magnaterra,
Peter Ferguson, Jennifer Nguyen, Pascale Guitera, Jose Banuls, Monika Janda,
Victoria Mar, Harald Kittler, H. Peter Soyer, Zongyuan Ge | A Multimodal Vision Foundation Model for Clinical Dermatology | 74 pages; Preprint; The code can be found at
https://github.com/SiyuanYan1/PanDerm | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diagnosing and treating skin diseases require advanced visual skills across
domains and the ability to synthesize information from multiple imaging
modalities. While current deep learning models excel at specific tasks like
skin cancer diagnosis from dermoscopic images, they struggle to meet the
complex, multimodal requirements of clinical practice. Here, we introduce
PanDerm, a multimodal dermatology foundation model pretrained through
self-supervised learning on over 2 million real-world skin disease images from
11 clinical institutions across 4 imaging modalities. We evaluated PanDerm on
28 diverse benchmarks, including skin cancer screening, risk stratification,
differential diagnosis of common and rare skin conditions, lesion segmentation,
longitudinal monitoring, and metastasis prediction and prognosis. PanDerm
achieved state-of-the-art performance across all evaluated tasks, often
outperforming existing models when using only 10% of labeled data. We conducted
three reader studies to assess PanDerm's potential clinical utility. PanDerm
outperformed clinicians by 10.2% in early-stage melanoma detection through
longitudinal analysis, improved clinicians' skin cancer diagnostic accuracy by
11% on dermoscopy images, and enhanced non-dermatologist healthcare providers'
differential diagnosis by 16.5% across 128 skin conditions on clinical
photographs. These results demonstrate PanDerm's potential to improve patient
care across diverse clinical scenarios and serve as a model for developing
multimodal foundation models in other medical specialties, potentially
accelerating the integration of AI support in healthcare. The code can be found
at https://github.com/SiyuanYan1/PanDerm.
| [
{
"version": "v1",
"created": "Sat, 19 Oct 2024 08:48:01 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Mar 2025 00:32:24 GMT"
},
{
"version": "v3",
"created": "Sun, 13 Apr 2025 05:58:18 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Yan",
"Siyuan",
""
],
[
"Yu",
"Zhen",
""
],
[
"Primiero",
"Clare",
""
],
[
"Vico-Alonso",
"Cristina",
""
],
[
"Wang",
"Zhonghua",
""
],
[
"Yang",
"Litao",
""
],
[
"Tschandl",
"Philipp",
""
],
[
"Hu",
"Ming",
""
],
[
"Ju",
"Lie",
""
],
[
"Tan",
"Gin",
""
],
[
"Tang",
"Vincent",
""
],
[
"Ng",
"Aik Beng",
""
],
[
"Powell",
"David",
""
],
[
"Bonnington",
"Paul",
""
],
[
"See",
"Simon",
""
],
[
"Magnaterra",
"Elisabetta",
""
],
[
"Ferguson",
"Peter",
""
],
[
"Nguyen",
"Jennifer",
""
],
[
"Guitera",
"Pascale",
""
],
[
"Banuls",
"Jose",
""
],
[
"Janda",
"Monika",
""
],
[
"Mar",
"Victoria",
""
],
[
"Kittler",
"Harald",
""
],
[
"Soyer",
"H. Peter",
""
],
[
"Ge",
"Zongyuan",
""
]
] |
2410.16350 | Daniel Durnford | M.M. Arora, L. Balogh, C. Beaufort, A. Brossard, M. Chapellier, J.
Clarke, E.C. Corcoran, J.-M. Coquillat, A. Dastgheibi-Fard, Y. Deng, D.
Durnford, C. Garrah, G. Gerbier, I. Giomataris, G. Giroux, P. Gorel, M. Gros,
P. Gros, O. Guillaudin, E.W. Hoppe, I. Katsioulas, F. Kelly, P. Knights, P.
Lautridou, A. Makowski, I. Manthos, R.D. Martin, J. Matthews, H.M. McCallum,
H. Meadows, L. Millins, J.-F. Muraz, T. Neep, K. Nikolopoulos, N. Panchal,
M.-C. Piro, N. Rowe, D. Santos, G. Savvidis, I. Savvidis, D. Spathara, F.
Vazquez de Sola Fernandez, R. Ward | The ionization yield in a methane-filled spherical proportional counter | 11 pages, 8 figures | Ionization yield in a methane-filled spherical proportional
counter, Phys. Rev. D 111, 072007 (2025) | 10.1103/PhysRevD.111.072007 | null | physics.ins-det hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spherical proportional counters (SPCs) are gaseous particle detectors
sensitive to single ionization electrons in their target media, with large
detector volumes and low background rates. The $\mbox{NEWS-G}$ collaboration
employs this technology to search for low-mass dark matter, having previously
performed searches with detectors at the Laboratoire Souterrain de Modane
(LSM), including a recent campaign with a 135 cm diameter SPC filled with
methane. While in situ calibrations of the detector response were carried out
at the LSM, measurements of the mean ionization yield and fluctuations of
methane gas in SPCs were performed using a 30 cm diameter detector. The results
of multiple measurements taken at different operating voltages are presented. A
UV laser system was used to measure the mean gas gain of the SPC, along with
$\mathrm{^{37}Ar}$ and aluminum-fluorescence calibration sources. These
measurements will inform the energy response model of future operating
detectors.
| [
{
"version": "v1",
"created": "Mon, 21 Oct 2024 17:59:46 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 23:05:20 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Arora",
"M. M.",
""
],
[
"Balogh",
"L.",
""
],
[
"Beaufort",
"C.",
""
],
[
"Brossard",
"A.",
""
],
[
"Chapellier",
"M.",
""
],
[
"Clarke",
"J.",
""
],
[
"Corcoran",
"E. C.",
""
],
[
"Coquillat",
"J. -M.",
""
],
[
"Dastgheibi-Fard",
"A.",
""
],
[
"Deng",
"Y.",
""
],
[
"Durnford",
"D.",
""
],
[
"Garrah",
"C.",
""
],
[
"Gerbier",
"G.",
""
],
[
"Giomataris",
"I.",
""
],
[
"Giroux",
"G.",
""
],
[
"Gorel",
"P.",
""
],
[
"Gros",
"M.",
""
],
[
"Gros",
"P.",
""
],
[
"Guillaudin",
"O.",
""
],
[
"Hoppe",
"E. W.",
""
],
[
"Katsioulas",
"I.",
""
],
[
"Kelly",
"F.",
""
],
[
"Knights",
"P.",
""
],
[
"Lautridou",
"P.",
""
],
[
"Makowski",
"A.",
""
],
[
"Manthos",
"I.",
""
],
[
"Martin",
"R. D.",
""
],
[
"Matthews",
"J.",
""
],
[
"McCallum",
"H. M.",
""
],
[
"Meadows",
"H.",
""
],
[
"Millins",
"L.",
""
],
[
"Muraz",
"J. -F.",
""
],
[
"Neep",
"T.",
""
],
[
"Nikolopoulos",
"K.",
""
],
[
"Panchal",
"N.",
""
],
[
"Piro",
"M. -C.",
""
],
[
"Rowe",
"N.",
""
],
[
"Santos",
"D.",
""
],
[
"Savvidis",
"G.",
""
],
[
"Savvidis",
"I.",
""
],
[
"Spathara",
"D.",
""
],
[
"Fernandez",
"F. Vazquez de Sola",
""
],
[
"Ward",
"R.",
""
]
] |
2410.16675 | Oluwafemi Odu | Oluwafemi Odu, Daniel M\'endez Beltran, Emiliano Berrones Guti\'errez,
Alvine B. Belle, Gerhard Yu, Melika Sherafat | SmartGSN: an online tool to semi-automatically manage assurance cases | null | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Developing industry-wide standards and ensuring producers of mission-critical
systems comply with them is crucial to fostering consumer acceptance. Producers
of such systems can rely on assurance cases to demonstrate to regulatory
authorities how they have complied with such standards to help prevent system
failure, which could result in fatalities and environmental damage. In this
paper, we introduce SmartGSN, an innovative online tool that relies on Large
Language Models to (semi-)automate the management of assurance cases complying
with GSN - a very popular notation to graphically represent assurance cases.
The evaluation of SmartGSN demonstrates its strong capability to detect
assurance case patterns within the assurance cases manually created for five
systems spanning several application domains. SmartGSN is accessible online at
[https://smartgsn.vercel.app], and a demonstration video can be viewed at
[https://youtu.be/qLrTHf-SZbM]
| [
{
"version": "v1",
"created": "Tue, 22 Oct 2024 04:16:19 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 15:33:44 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Odu",
"Oluwafemi",
""
],
[
"Beltran",
"Daniel Méndez",
""
],
[
"Gutiérrez",
"Emiliano Berrones",
""
],
[
"Belle",
"Alvine B.",
""
],
[
"Yu",
"Gerhard",
""
],
[
"Sherafat",
"Melika",
""
]
] |
2410.17248 | V\'it R\r{u}\v{z}i\v{c}ka | V\'it R\r{u}\v{z}i\v{c}ka, Andrew Markham | HyperspectralViTs: General Hyperspectral Models for On-board Remote
Sensing | 14 pages, Accepted manuscript version, published version may differ
in minor details and formatting | null | 10.1109/JSTARS.2025.3557527 | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | On-board processing of hyperspectral data with machine learning models would
enable unprecedented amount of autonomy for a wide range of tasks, for example
methane detection or mineral identification. This can enable early warning
system and could allow new capabilities such as automated scheduling across
constellations of satellites. Classical methods suffer from high false positive
rates and previous deep learning models exhibit prohibitive computational
requirements. We propose fast and accurate machine learning architectures which
support end-to-end training with data of high spectral dimension without
relying on hand-crafted products or spectral band compression preprocessing. We
evaluate our models on two tasks related to hyperspectral data processing. With
our proposed general architectures, we improve the F1 score of the previous
methane detection state-of-the-art models by 27% on a newly created synthetic
dataset and by 13% on the previously released large benchmark dataset. We also
demonstrate that training models on the synthetic dataset improves performance
of models finetuned on the dataset of real events by 6.9% in F1 score in
contrast with training from scratch. On a newly created dataset for mineral
identification, our models provide 3.5% improvement in the F1 score in contrast
to the default versions of the models. With our proposed models we improve the
inference speed by 85% in contrast to previous classical and deep learning
approaches by removing the dependency on classically computed features. With
our architecture, one capture from the EMIT sensor can be processed within 30
seconds on realistic proxy of the ION-SCV 004 satellite.
| [
{
"version": "v1",
"created": "Tue, 22 Oct 2024 17:59:55 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Oct 2024 15:06:36 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Apr 2025 12:37:11 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Růžička",
"Vít",
""
],
[
"Markham",
"Andrew",
""
]
] |
2410.17254 | Gunther Leobacher | Gunther Leobacher, Tapio Rajala, Alexander Steinicke, J\"org
Thuswaldner | Measure and dimension theory of permeable sets and its applications to
fractals | null | null | null | null | math.GN math.DS math.GT math.MG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study {\it permeable} sets. These are sets \(\Theta \subset \mathbb{R}^d\)
which have the property that each two points \(x,y\in \mathbb{R}^d\) can be
connected by a short path \(\gamma\) which has small (or even empty, apart from
the end points of \(\gamma\)) intersection with \(\Theta\). We investigate
relations between permeability and Lebesgue measure and establish theorems on
the relation of permeability with several notions of dimension. It turns out
that for most notions of dimension each subset of \(\mathbb{R}^d\) of dimension
less than \(d-1\) is permeable. We use our permeability result on the Nagata
dimension to characterize permeability properties of self-similar sets with
certain finiteness properties.
| [
{
"version": "v1",
"created": "Thu, 3 Oct 2024 18:57:50 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 15:18:01 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Leobacher",
"Gunther",
""
],
[
"Rajala",
"Tapio",
""
],
[
"Steinicke",
"Alexander",
""
],
[
"Thuswaldner",
"Jörg",
""
]
] |
2410.17509 | Jinghan Jia | Jinghan Jia, Jiancheng Liu, Yihua Zhang, Parikshit Ram, Nathalie
Baracaldo, Sijia Liu | WAGLE: Strategic Weight Attribution for Effective and Modular Unlearning
in Large Language Models | NeurIPS'24 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The need for effective unlearning mechanisms in large language models (LLMs)
is increasingly urgent, driven by the necessity to adhere to data regulations
and foster ethical generative AI practices. Despite growing interest of LLM
unlearning, much of the existing research has focused on varied unlearning
method designs to boost effectiveness and efficiency. However, the inherent
relationship between model weights and LLM unlearning has not been extensively
examined. In this paper, we systematically explore how model weights interact
with unlearning processes in LLMs and we design the weight attribution-guided
LLM unlearning method, WAGLE, which unveils the interconnections between
'influence' of weights and 'influence' of data to forget and retain in LLM
generation. By strategically guiding the LLM unlearning across different types
of unlearning methods and tasks, WAGLE can erase the undesired content, while
maintaining the performance of the original tasks. We refer to the weight
attribution-guided LLM unlearning method as WAGLE, which unveils the
interconnections between 'influence' of weights and 'influence' of data to
forget and retain in LLM generation. Our extensive experiments show that WAGLE
boosts unlearning performance across a range of LLM unlearning methods such as
gradient difference and (negative) preference optimization, applications such
as fictitious unlearning, malicious use prevention, and copyrighted information
removal, and models including Zephyr-7b-beta and Llama2-7b. To the best of our
knowledge, our work offers the first principled method for attributing and
pinpointing the influential weights in enhancing LLM unlearning. It stands in
contrast to previous methods that lack weight attribution and simpler weight
attribution techniques.
| [
{
"version": "v1",
"created": "Wed, 23 Oct 2024 02:22:07 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 21:23:46 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Jia",
"Jinghan",
""
],
[
"Liu",
"Jiancheng",
""
],
[
"Zhang",
"Yihua",
""
],
[
"Ram",
"Parikshit",
""
],
[
"Baracaldo",
"Nathalie",
""
],
[
"Liu",
"Sijia",
""
]
] |
2410.17672 | Haoyue Zhang | Hao-Yue Zhang, Yi-Xuan Yao, Bin-Yao Huang, Jing-Yi-Ran Jin and Qing Ai | Non-Hermitian Hamiltonian Approach for Two-Dimensional Coherent Spectra
of Driven Systems | null | null | 10.1021/acs.jctc.4c01737 | null | quant-ph physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two-dimensional coherent spectroscopy (2DCS) offers significant advantages in
terms of high temporal and frequency resolutions and signal-to-noise ratio.
Until now, the response-function (RF) formalism has been the prevalent
theoretical description. In this study, we compare the non-Hermitian
Hamiltonian (NHH) method with the RF formalism in a three-level system with a
constant control field. We obtain the signals from both approaches and compare
their population dynamics and 2DCS. We propose the quasi-Green functions for
the NHH method, which allows all dominant Liouville paths to be inferred. We
further simulate the 2DCS of Rh(CO)$_2$C$_5$H$_7$O$_2$ (RDC) dissolved in
hexane with the NHH method, which is in good agreement with the previous
experiments. Although the NHH method overestimates relaxations, it provides all
important paths by analytical solutions, which are different from the four
paths used in the RF formalism. Our results demonstrate that the NHH method is
more suitable than the RF formalism for investigating the systems including
relaxation and control fields via the 2DCS.
| [
{
"version": "v1",
"created": "Wed, 23 Oct 2024 08:46:46 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 07:49:48 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhang",
"Hao-Yue",
""
],
[
"Yao",
"Yi-Xuan",
""
],
[
"Huang",
"Bin-Yao",
""
],
[
"Jin",
"Jing-Yi-Ran",
""
],
[
"Ai",
"Qing",
""
]
] |
2410.17817 | Naomi Andrew | Naomi Andrew, Paige Hillen, Robert Alonzo Lyman, Catherine Eva Pfaff | A hyperbolic free-by-cyclic group determined by its finite quotients | 4 pages. v2: final version, to appear in Glasgow Mathematical Journal | null | 10.1017/S0017089525000096 | null | math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that the group $\langle a,b,c,t : a^t=b,b^t=c,c^t=ca^{-1} \rangle$ is
profinitely rigid amongst free-by-cyclic groups, providing the first example of
a hyperbolic free-by-cyclic group with this property.
| [
{
"version": "v1",
"created": "Wed, 23 Oct 2024 12:21:05 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 18:25:09 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Andrew",
"Naomi",
""
],
[
"Hillen",
"Paige",
""
],
[
"Lyman",
"Robert Alonzo",
""
],
[
"Pfaff",
"Catherine Eva",
""
]
] |
2410.18194 | Brando Miranda | Elyas Obbad, Iddah Mlauzi, Brando Miranda, Rylan Schaeffer, Kamal
Obbad, Suhana Bedi, Sanmi Koyejo | ZIP-FIT: Embedding-Free Data Selection via Compression-Based Alignment | null | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Data selection is crucial for optimizing language model (LM) performance on
specific tasks, yet most existing methods fail to effectively consider the
target task distribution.
Current approaches either ignore task-specific requirements entirely or rely
on approximations that fail to capture the nuanced patterns needed for tasks
like Autoformalization or code generation.
Methods that do consider the target distribution often rely on simplistic,
sometimes noisy, representations, like hashed n-gram features, which can lead
to collisions and introduce noise.
We introduce ZIP-FIT, a data selection framework that uses gzip compression
to directly measure alignment between potential training data and the target
task distribution.
In extensive evaluations on Autoformalization and Python code generation,
ZIP-FIT significantly outperforms leading baselines like DSIR and D4.
Models trained on ZIP-FIT-selected data achieve their lowest cross-entropy
loss up to 85.1\% faster than baselines, demonstrating that better task
alignment leads to more efficient learning.
In addition, ZIP-FIT performs selection up to 65.8\% faster than DSIR and two
orders of magnitude faster than D4.
Notably, ZIP-FIT shows that smaller, well-aligned datasets often outperform
larger but less targeted ones, demonstrating that a small amount of higher
quality data is superior to a large amount of lower quality data.
Our results imply that task-aware data selection is crucial for efficient
domain adaptation, and that compression offers a principled way to measure task
alignment.
By showing that targeted data selection can dramatically improve
task-specific performance, our work provides new insights into the relationship
between data quality, task alignment, and model learning efficiency.
| [
{
"version": "v1",
"created": "Wed, 23 Oct 2024 18:01:06 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 19:39:18 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Obbad",
"Elyas",
""
],
[
"Mlauzi",
"Iddah",
""
],
[
"Miranda",
"Brando",
""
],
[
"Schaeffer",
"Rylan",
""
],
[
"Obbad",
"Kamal",
""
],
[
"Bedi",
"Suhana",
""
],
[
"Koyejo",
"Sanmi",
""
]
] |
2410.18229 | Paul Howard | Paul Howard and Eleftherios Tachtsis | Three forms of the Erd\H{o}s-Dushnik-Miller Theorem | null | null | null | null | math.LO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We continue the study of the Erd\H{o}s-Dushnik-Miller theorem (A graph with
an uncountable set of vertices has either an infinite independent set or an
uncountable clique) in set theory without the axiom of choice. We show that
there are three inequivalent versions of this theorem and we give some results
about the positions of these versions in the deductive hierarchy of weak choice
principles.
| [
{
"version": "v1",
"created": "Wed, 23 Oct 2024 19:21:47 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Mar 2025 19:31:27 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Apr 2025 18:58:43 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Howard",
"Paul",
""
],
[
"Tachtsis",
"Eleftherios",
""
]
] |
2410.18491 | Hengxiang Zhang | Hengxiang Zhang, Hongfu Gao, Qiang Hu, Guanhua Chen, Lili Yang, Bingyi
Jing, Hongxin Wei, Bing Wang, Haifeng Bai, Lei Yang | ChineseSafe: A Chinese Benchmark for Evaluating Safety in Large Language
Models | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid development of Large language models (LLMs), understanding the
capabilities of LLMs in identifying unsafe content has become increasingly
important. While previous works have introduced several benchmarks to evaluate
the safety risk of LLMs, the community still has a limited understanding of
current LLMs' capability to recognize illegal and unsafe content in Chinese
contexts. In this work, we present a Chinese safety benchmark (ChineseSafe) to
facilitate research on the content safety of large language models. To align
with the regulations for Chinese Internet content moderation, our ChineseSafe
contains 205,034 examples across 4 classes and 10 sub-classes of safety issues.
For Chinese contexts, we add several special types of illegal content:
political sensitivity, pornography, and variant/homophonic words. Moreover, we
employ two methods to evaluate the legal risks of popular LLMs, including
open-sourced models and APIs. The results reveal that many LLMs exhibit
vulnerability to certain types of safety issues, leading to legal risks in
China. Our work provides a guideline for developers and researchers to
facilitate the safety of LLMs. Our results are also available at
https://huggingface.co/spaces/SUSTech/ChineseSafe-Benchmark. Additionally, we
release a test set comprising 200,000 examples, which is publicly accessible at
https://huggingface.co/datasets/SUSTech/ChineseSafe.
| [
{
"version": "v1",
"created": "Thu, 24 Oct 2024 07:25:29 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 06:16:49 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhang",
"Hengxiang",
""
],
[
"Gao",
"Hongfu",
""
],
[
"Hu",
"Qiang",
""
],
[
"Chen",
"Guanhua",
""
],
[
"Yang",
"Lili",
""
],
[
"Jing",
"Bingyi",
""
],
[
"Wei",
"Hongxin",
""
],
[
"Wang",
"Bing",
""
],
[
"Bai",
"Haifeng",
""
],
[
"Yang",
"Lei",
""
]
] |
2410.19449 | Maxence Noble | Maxence Noble, Louis Grenioux, Marylou Gabri\'e, Alain Oliviero Durmus | Learned Reference-based Diffusion Sampling for multi-modal distributions | Accepted at ICLR 2025 | null | null | null | stat.ML cs.LG stat.CO | http://creativecommons.org/licenses/by/4.0/ | Over the past few years, several approaches utilizing score-based diffusion
have been proposed to sample from probability distributions, that is without
having access to exact samples and relying solely on evaluations of
unnormalized densities. The resulting samplers approximate the time-reversal of
a noising diffusion process, bridging the target distribution to an
easy-to-sample base distribution. In practice, the performance of these methods
heavily depends on key hyperparameters that require ground truth samples to be
accurately tuned. Our work aims to highlight and address this fundamental
issue, focusing in particular on multi-modal distributions, which pose
significant challenges for existing sampling methods. Building on existing
approaches, we introduce Learned Reference-based Diffusion Sampler (LRDS), a
methodology specifically designed to leverage prior knowledge on the location
of the target modes in order to bypass the obstacle of hyperparameter tuning.
LRDS proceeds in two steps by (i) learning a reference diffusion model on
samples located in high-density space regions and tailored for multimodality,
and (ii) using this reference model to foster the training of a diffusion-based
sampler. We experimentally demonstrate that LRDS best exploits prior knowledge
on the target distribution compared to competing algorithms on a variety of
challenging distributions.
| [
{
"version": "v1",
"created": "Fri, 25 Oct 2024 10:23:34 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Feb 2025 15:48:19 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Apr 2025 08:04:13 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Noble",
"Maxence",
""
],
[
"Grenioux",
"Louis",
""
],
[
"Gabrié",
"Marylou",
""
],
[
"Durmus",
"Alain Oliviero",
""
]
] |
2410.19631 | Julien Roy | Ihor Neporozhnii, Julien Roy, Emmanuel Bengio, Jason Hartford | Efficient Biological Data Acquisition through Inference Set Design | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In drug discovery, highly automated high-throughput laboratories are used to
screen a large number of compounds in search of effective drugs. These
experiments are expensive, so one might hope to reduce their cost by only
experimenting on a subset of the compounds, and predicting the outcomes of the
remaining experiments. In this work, we model this scenario as a sequential
subset selection problem: we aim to select the smallest set of candidates in
order to achieve some desired level of accuracy for the system as a whole. Our
key observation is that, if there is heterogeneity in the difficulty of the
prediction problem across the input space, selectively obtaining the labels for
the hardest examples in the acquisition pool will leave only the relatively
easy examples to remain in the inference set, leading to better overall system
performance. We call this mechanism inference set design, and propose the use
of a confidence-based active learning solution to prune out these challenging
examples. Our algorithm includes an explicit stopping criterion that interrupts
the acquisition loop when it is sufficiently confident that the system has
reached the target performance. Our empirical studies on image and molecular
datasets, as well as a real-world large-scale biological assay, show that
active learning for inference set design leads to significant reduction in
experimental cost while retaining high system performance.
| [
{
"version": "v1",
"created": "Fri, 25 Oct 2024 15:34:03 GMT"
},
{
"version": "v2",
"created": "Mon, 25 Nov 2024 17:51:33 GMT"
},
{
"version": "v3",
"created": "Sat, 1 Mar 2025 23:46:21 GMT"
},
{
"version": "v4",
"created": "Fri, 11 Apr 2025 20:59:24 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Neporozhnii",
"Ihor",
""
],
[
"Roy",
"Julien",
""
],
[
"Bengio",
"Emmanuel",
""
],
[
"Hartford",
"Jason",
""
]
] |
2410.21491 | Hongyang Li | Hongyang Li, Caesar Wu, Mohammed Chadli, Said Mammar, Pascal Bouvry | Trustworthiness of Stochastic Gradient Descent in Distributed Learning | null | null | null | null | cs.LG cs.AI cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Distributed learning (DL) uses multiple nodes to accelerate training,
enabling efficient optimization of large-scale models. Stochastic Gradient
Descent (SGD), a key optimization algorithm, plays a central role in this
process. However, communication bottlenecks often limit scalability and
efficiency, leading to increasing adoption of compressed SGD techniques to
alleviate these challenges. Despite addressing communication overheads,
compressed SGD introduces trustworthiness concerns, as gradient exchanges among
nodes are vulnerable to attacks like gradient inversion (GradInv) and
membership inference attacks (MIA). The trustworthiness of compressed SGD
remains unexplored, leaving important questions about its reliability
unanswered.
In this paper, we provide a trustworthiness evaluation of compressed versus
uncompressed SGD. Specifically, we conducted empirical studies using GradInv
attacks, revealing that compressed SGD demonstrates significantly higher
resistance to privacy leakage compared to uncompressed SGD. In addition, our
findings suggest that MIA may not be a reliable metric for assessing privacy
risks in distributed learning.
| [
{
"version": "v1",
"created": "Mon, 28 Oct 2024 20:02:05 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 11:19:21 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Li",
"Hongyang",
""
],
[
"Wu",
"Caesar",
""
],
[
"Chadli",
"Mohammed",
""
],
[
"Mammar",
"Said",
""
],
[
"Bouvry",
"Pascal",
""
]
] |
2410.21834 | Sa Wang | Sa Wang, Shuang Li, Yao Li, Ben-Wei Zhang, Enke Wang | Probing the mass effect of heavy quark jets in high-energy nuclear
collisions | 12 pages, 12 figures. The version to be published in CPC | Chin.Phys.C 49 (2025) 6, 064101 | 10.1088/1674-1137/adb385 | null | hep-ph nucl-th | http://creativecommons.org/licenses/by/4.0/ | The production of heavy quark (HQ) jets provides a new arena to address the
mass effect of jet quenching in heavy-ion physics. This paper presents a
theoretical study of HQ jet yield suppression in Pb+Pb collisions at the LHC
and focuses on the energy loss of HQ jets produced by different mechanisms. The
p+p baseline is carried out by the SHERPA generator, and the jet-medium
interactions are described by the SHELL transport model, which considers the
elastic and inelastic partonic energy loss in the quark-gluon plasma (QGP). In
p+p collisions, our numerical results indicate that the HQ jets from gluon
splitting ($g \rightarrow Q$-jet) give the dominant contribution at high $p_T$,
and it shows more dispersive structures than the HQ-initiated one ($Q
\rightarrow Q$-jet). In nucleus-nucleus collisions, our calculations are
consistent with the inclusive and b-jet $R_{AA}$ recently measured by the ATLAS
collaboration, which suggests a remarkable manifestation of the mass effect of
jet energy loss. As a result of the dispersive substructure, the $g \rightarrow
Q$-jet will lose more energy than the $Q \rightarrow Q$-jet in the QGP. Due to
the significant contribution of $g \rightarrow c$-jet, the $R_{AA}$ of c-jet
will be comparable or even smaller than that of inclusive jet. To
experimentally distinguish the $g \rightarrow Q$-jet and $Q \rightarrow Q$-jet,
we propose the event selection strategies based on their topological features
and test the performances. By isolating the $c \rightarrow c$-jet and $b
\rightarrow b$-jet, the jets initiated by heavy quarks, we predict that the
order of their $R_{AA}$ are in line with the mass hierarchy of energy loss.
Future measurements on the $R_{AA}$ of $Q \rightarrow Q$-jet and $g \rightarrow
Q$-jet will provide a unique chance to test the flavor/mass dependence of
energy loss at the jet level.
| [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 07:57:38 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Feb 2025 11:54:09 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Wang",
"Sa",
""
],
[
"Li",
"Shuang",
""
],
[
"Li",
"Yao",
""
],
[
"Zhang",
"Ben-Wei",
""
],
[
"Wang",
"Enke",
""
]
] |
2410.22455 | Alessandra Rizzo | Alessandra Rizzo | Classification of 1+0 two-dimensional Hamiltonian operators | null | null | null | null | math-ph math.MP nlin.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we study Hamiltonian operators which are sum of a first order
operator and of a Poisson tensor, in two spatial independent variables. In
particular, a complete classification of these operators is presented in two
and three components, analyzing both the cases of degenerate and non degenerate
leading coefficients.
| [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 18:42:13 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 09:41:47 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Rizzo",
"Alessandra",
""
]
] |
2410.23113 | Jiangming Yao | E. F. Zhou, C. R. Ding, J. M. Yao, B. Bally, H. Hergert, C. F. Jiao,
and T. R. Rodr\'iguez | Ab initio nuclear shape coexistence and emergence of island of inversion
around $N=20$ | 8 pages with 5 figures and 1 table, to appear in Phys. Lett. B | Physics Letters B 865, 139464 (2025) | 10.1016/j.physletb.2025.139464 | null | nucl-th nucl-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We extend a nuclear ab initio framework based on chiral two- and
three-nucleon interactions to investigate shape coexistence and the degradation
of the $N=20$ magic number in both even-even and odd-even magnesium isotopes.
The quantum-number projected generator coordinate method, combined with the
in-medium similarity renormalization group (IMSRG), is employed to compute
their low-lying states. This approach reasonably reproduces the coexistence of
weakly and strongly deformed states at comparable energies, and allows us to
track the emergence of the $N=20$ island of inversion through the continuous
IMSRG evolution of the chiral Hamiltonian. Our results indicate that the ground
state of $^{33}$Mg with spin-parity $3/2^-$ is predominantly a strongly
deformed configuration with $K^\pi = 3/2^-$, while the lowest $7/2^-$ state is
predicted to be a shape isomer, consisting of a mixture of weakly deformed
configurations with different $K$ values. The results highlight the essential
roles of both dynamical and static collective correlations in reproducing the
ordering of nuclear states with distinct shapes.
| [
{
"version": "v1",
"created": "Wed, 30 Oct 2024 15:24:47 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Apr 2025 14:50:10 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhou",
"E. F.",
""
],
[
"Ding",
"C. R.",
""
],
[
"Yao",
"J. M.",
""
],
[
"Bally",
"B.",
""
],
[
"Hergert",
"H.",
""
],
[
"Jiao",
"C. F.",
""
],
[
"Rodríguez",
"T. R.",
""
]
] |
2410.23519 | Daniel Julio Garcia | D. J. Garcia, J. G. Sereni, A. A. Aligia | Specific heat of Gd$^{3+}$- and Eu$^{2+}$-based magnetic compounds | 20 pages, 6 figures | null | null | null | cond-mat.str-el cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We have studied theoretically the specific heat of a large number of
non-frustrated magnetic structures described by the Heisenberg model for
systems with total angular momentum $J=7/2$, corresponding to the 4f$^7$
configuration of Gd$^{+3}$ and Eu$^{+2}$. For a given critical temperature
(determined by the magnitude of the exchange interactions), we find that, to a
high degree of accuracy, the specific heat is governed by two primary
parameters: the effective number of neighbors, which dictates the extent of
quantum fluctuations, and the axial anisotropy. Using these two parameters we
fit the specific heat of four Gd compounds and two Eu compounds, achieving a
remarkable agreement. Our work opens the possibility of describing the specific
heat of 4f$^7$ systems within a general framework.
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 00:03:58 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 18:36:21 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Garcia",
"D. J.",
""
],
[
"Sereni",
"J. G.",
""
],
[
"Aligia",
"A. A.",
""
]
] |
2410.24037 | Sunjae Yoon | Sunjae Yoon, Gwanhyeong Koo, Younghwan Lee, Chang D. Yoo | TPC: Test-time Procrustes Calibration for Diffusion-based Human Image
Animation | 24 pages, 16 figures, NeurIPS 2024 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human image animation aims to generate a human motion video from the inputs
of a reference human image and a target motion video. Current diffusion-based
image animation systems exhibit high precision in transferring human identity
into targeted motion, yet they still exhibit irregular quality in their
outputs. Their optimal precision is achieved only when the physical
compositions (i.e., scale and rotation) of the human shapes in the reference
image and target pose frame are aligned. In the absence of such alignment,
there is a noticeable decline in fidelity and consistency. Especially, in
real-world environments, this compositional misalignment commonly occurs,
posing significant challenges to the practical usage of current systems. To
this end, we propose Test-time Procrustes Calibration (TPC), which enhances the
robustness of diffusion-based image animation systems by maintaining optimal
performance even when faced with compositional misalignment, effectively
addressing real-world scenarios. The TPC provides a calibrated reference image
for the diffusion model, enhancing its capability to understand the
correspondence between human shapes in the reference and target images. Our
method is simple and can be applied to any diffusion-based image animation
system in a model-agnostic manner, improving the effectiveness at test time
without additional training.
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 15:34:49 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 09:02:09 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Yoon",
"Sunjae",
""
],
[
"Koo",
"Gwanhyeong",
""
],
[
"Lee",
"Younghwan",
""
],
[
"Yoo",
"Chang D.",
""
]
] |
2411.00140 | Sanaz Mahmoodi Takaghaj | Sanaz Mahmoodi Takaghaj | ViT-LCA: A Neuromorphic Approach for Vision Transformers | Accepted to Artificial Intelligence Circuits And Systems(AICAS), 2015 | null | null | null | cs.NE cs.ET | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The recent success of Vision Transformers has generated significant interest
in attention mechanisms and transformer architectures. Although existing
methods have proposed spiking self-attention mechanisms compatible with spiking
neural networks, they often face challenges in effective deployment on current
neuromorphic platforms. This paper introduces a novel model that combines
vision transformers with the Locally Competitive Algorithm (LCA) to facilitate
efficient neuromorphic deployment. Our experiments show that ViT-LCA achieves
higher accuracy on ImageNet-1K dataset while consuming significantly less
energy than other spiking vision transformer counterparts. Furthermore,
ViT-LCA's neuromorphic-friendly design allows for more direct mapping onto
current neuromorphic architectures.
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 18:41:30 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 01:39:54 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Takaghaj",
"Sanaz Mahmoodi",
""
]
] |
2411.00246 | Lorenzo Basile | Lorenzo Basile, Valentino Maiorca, Luca Bortolussi, Emanuele Rodol\`a,
Francesco Locatello | ResiDual Transformer Alignment with Spectral Decomposition | Published in Transactions on Machine Learning Research (TMLR) | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | When examined through the lens of their residual streams, a puzzling property
emerges in transformer networks: residual contributions (e.g., attention heads)
sometimes specialize in specific tasks or input attributes. In this paper, we
analyze this phenomenon in vision transformers, focusing on the spectral
geometry of residuals, and explore its implications for modality alignment in
vision-language models. First, we link it to the intrinsically low-dimensional
structure of visual head representations, zooming into their principal
components and showing that they encode specialized roles across a wide variety
of input data distributions. Then, we analyze the effect of head specialization
in multimodal models, focusing on how improved alignment between text and
specialized heads impacts zero-shot classification performance. This
specialization-performance link consistently holds across diverse pre-training
data, network sizes, and objectives, demonstrating a powerful new mechanism for
boosting zero-shot classification through targeted alignment. Ultimately, we
translate these insights into actionable terms by introducing ResiDual, a
technique for spectral alignment of the residual stream. Much like panning for
gold, it lets the noise from irrelevant unit principal components (i.e.,
attributes) wash away to amplify task-relevant ones. Remarkably, this dual
perspective on modality alignment yields fine-tuning level performance on
different data distributions while modelling an extremely interpretable and
parameter-efficient transformation, as we extensively show on 70 pre-trained
network-dataset combinations (7 models, 10 datasets).
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 22:51:45 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 13:51:08 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Basile",
"Lorenzo",
""
],
[
"Maiorca",
"Valentino",
""
],
[
"Bortolussi",
"Luca",
""
],
[
"Rodolà",
"Emanuele",
""
],
[
"Locatello",
"Francesco",
""
]
] |
2411.00264 | Sina Rismanchian | Sina Rismanchian, Yasaman Razeghi, Sameer Singh, Shayan Doroudi | TurtleBench: A Visual Programming Benchmark in Turtle Geometry | null | null | null | null | cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Humans have the ability to reason about geometric patterns in images and
scenes from a young age. However, developing large multimodal models (LMMs)
capable of similar reasoning remains a challenge, highlighting the need for
robust evaluation methods to assess these capabilities. We introduce \Turtle, a
benchmark designed to evaluate LMMs' capacity to interpret geometric patterns
-- given visual examples, textual instructions, or both -- and generate precise
code outputs. Inspired by turtle geometry, a notion used to teach children
foundational coding and geometric concepts, TurtleBench features tasks with
patterned shapes that have underlying algorithmic logic. Our evaluation reveals
that leading LMMs struggle significantly with these tasks, with GPT-4o
achieving only 19\% accuracy on the simplest tasks and few-shot prompting only
marginally improves their performance ($<2\%$). \Turtle highlights the gap
between human and AI performance in intuitive and visual geometrical
understanding, setting the stage for future research in this area. \Turtle
stands as one of the few benchmarks to evaluate the integration of visual
understanding and code generation capabilities in LMMs, setting the stage for
future research. Code and Dataset for this paper is provided here:
\href{https://github.com/sinaris76/TurtleBench}{https://github.com/sinaris76/TurtleBench}
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 23:52:06 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 21:25:02 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Rismanchian",
"Sina",
""
],
[
"Razeghi",
"Yasaman",
""
],
[
"Singh",
"Sameer",
""
],
[
"Doroudi",
"Shayan",
""
]
] |
2411.00601 | Evangelia Tzimpimpaki | Evangelia Tzimpimpaki, Thrasyvoulos Spyropoulos | Diversity in Network-Friendly Recommendations | 9 pages, 4 figures | null | null | null | cs.PF | http://creativecommons.org/licenses/by/4.0/ | In recent years, the Internet has been dominated by content-rich platforms,
employing recommendation systems to provide users with more appealing content
(e.g., videos in YouTube, movies in Netflix). While traditional content
recommendations are oblivious to network conditions, the paradigm of
Network-Friendly Recommendations (NFR) has recently emerged, favoring content
that improves network performance (e.g. cached near the user), while still
being appealing to the user. However, NFR algorithms sometimes achieve their
goal by shrinking the pool of content recommended to users. The undesirable
side-effect is reduced content diversity, a phenomenon known as
``content/filter bubble''. This reduced diversity is problematic for both
users, who are prevented from exploring a broader range of content, and content
creators (e.g. YouTubers) whose content may be recommended less frequently,
leading to perceived unfairness. In this paper, we first investigate - using
real data and state-of-the-art NFR schemes - the extent of this phenomenon. We
then formulate a ``Diverse-NFR'' optimization problem (i.e., network-friendly
recommendations with - sufficient - content diversity), and through a series of
transformation steps, we manage to reduce it to a linear program that can be
solved fast and optimally. Our findings show that Diverse-NFR can achieve high
network gains (comparable to non-diverse NFR) while maintaining diversity
constraints. To our best knowledge, this is the first work that incorporates
diversity issues into network-friendly recommendation algorithms.
| [
{
"version": "v1",
"created": "Fri, 1 Nov 2024 14:03:21 GMT"
},
{
"version": "v2",
"created": "Mon, 4 Nov 2024 09:40:27 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 11:20:56 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Tzimpimpaki",
"Evangelia",
""
],
[
"Spyropoulos",
"Thrasyvoulos",
""
]
] |
2411.01009 | Rui-Zhen Huang | Wen-Tao Xu and Rui-Zhen Huang | Finite Correlation Length Scaling of Disorder Parameter at Quantum
Criticality | 7+4 pages, 4+3 figures; tittle changed | Phys. Rev. Lett. 134, 146503 (2025) | 10.1103/PhysRevLett.134.146503 | null | cond-mat.str-el cond-mat.stat-mech quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The disorder parameter, defined as the expectation value of the symmetry
transformation acting on a subsystem, can be used to characterize symmetric
phases as an analogy to detecting spontaneous symmetry breaking (SSB) phases
using local order parameters. In a dual picture, disorder parameters actually
detect SSB of higher-form symmetries. In this work, we show that the non-local
disorder parameters can be conveniently and efficiently evaluated using
infinite projected entangled pair states (iPEPS). Moreover, we propose a finite
correlation length scaling theory of the disorder parameter within the quantum
critical region and validate the scaling theory with variationally optimized
iPEPS. We find from the finite correlation length scaling that the disorder
parameter satisfies perimeter law at a critical point, i.e., it decays
exponentially with the boundary size of the subsystem, indicating spontaneous
higher-form symmetry breaking at the critical point of the dual model.
| [
{
"version": "v1",
"created": "Fri, 1 Nov 2024 20:19:57 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 06:45:38 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Xu",
"Wen-Tao",
""
],
[
"Huang",
"Rui-Zhen",
""
]
] |
2411.01551 | Quanyu Tang | Yizhe Ji, Quanyu Tang, Wei Wang, Hao Zhang | New arithmetic invariants for cospectral graphs | 15 pages, 1 figure. v2 added some applications in the Introduction | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An invariant for cospectral graphs is a property shared by all cospectral
graphs. In this paper, we establish three novel arithmetic invariants for
cospectral graphs, revealing deep connections between spectral properties and
combinatorial structures. More precisely, one of our main results shows that
for any two cospectral graphs $G$ and $H$ with adjacency matrices $A(G)$ and
$A(H)$, respectively, the following congruence holds for all integers $m\geq
0$:\[e^{\rm T}A(G)^me\equiv e^{\rm T}A(H)^me \pmod{4},\] where $e$ is the
all-one vector. Moreover, we present a number of fascinating applications.
Specifically: i) Resolving a conjecture proposed by the third author, we
demonstrate that under certain conditions, every graph cospectral with a graph
$G$ is determined by its generalized spectrum. ii) We demonstrate that whenever
the complements of two trees are cospectral, then one tree has a perfect
matching if and only if the other does. An analogous result holds for the
existence of triangles in general graphs. iii) An unexpected connection to the
polynomial reconstruction problem is also provided, showing that the parity of
the constant term of the characteristic polynomial is reconstructible.
| [
{
"version": "v1",
"created": "Sun, 3 Nov 2024 12:54:43 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 06:22:58 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Ji",
"Yizhe",
""
],
[
"Tang",
"Quanyu",
""
],
[
"Wang",
"Wei",
""
],
[
"Zhang",
"Hao",
""
]
] |
2411.01894 | R\'emy Portelas | Emilien Bir\'e, Anthony Kobanda, Ludovic Denoyer, R\'emy Portelas | Efficient Active Imitation Learning with Random Network Distillation | Accepted at ICLR 2025 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Developing agents for complex and underspecified tasks, where no clear
objective exists, remains challenging but offers many opportunities. This is
especially true in video games, where simulated players (bots) need to play
realistically, and there is no clear reward to evaluate them. While imitation
learning has shown promise in such domains, these methods often fail when
agents encounter out-of-distribution scenarios during deployment. Expanding the
training dataset is a common solution, but it becomes impractical or costly
when relying on human demonstrations. This article addresses active imitation
learning, aiming to trigger expert intervention only when necessary, reducing
the need for constant expert input along training. We introduce Random Network
Distillation DAgger (RND-DAgger), a new active imitation learning method that
limits expert querying by using a learned state-based out-of-distribution
measure to trigger interventions. This approach avoids frequent expert-agent
action comparisons, thus making the expert intervene only when it is useful. We
evaluate RND-DAgger against traditional imitation learning and other active
approaches in 3D video games (racing and third-person navigation) and in a
robotic locomotion task and show that RND-DAgger surpasses previous methods by
reducing expert queries. https://sites.google.com/view/rnd-dagger
| [
{
"version": "v1",
"created": "Mon, 4 Nov 2024 08:50:52 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 10:25:47 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Biré",
"Emilien",
""
],
[
"Kobanda",
"Anthony",
""
],
[
"Denoyer",
"Ludovic",
""
],
[
"Portelas",
"Rémy",
""
]
] |
2411.01935 | Steffen Lennart B\"otzel | Steffen B\"otzel, Frank Lechermann, Takasada Shibauchi, Ilya M. Eremin | Theory of potential impurity scattering in pressurized superconducting
La$_3$Ni$_2$O$_7$ | 9 pages, 2 figures + supplemental material (4 pages, 4 figures) | null | 10.1038/s42005-025-02056-7 | null | cond-mat.supr-con cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | Recently discovered high-T$_c$ superconductivity in pressurized bilayer
nickelate La$_3$Ni$_2$O$_7$ (La-327) is believed to be driven by the
non-phononic repulsive interaction. Depending on the strength of the interlayer
repulsion, the symmetry of the superconducting order parameter is expected to
be either $d$-wave or sign-changing bonding-antibonding $s_{\pm}$-wave.
Unfortunately, due to the need of high pressure to reach superconducting phase,
conventional spectroscopic probes to validate the symmetry of the order
parameter are hard to use. Here, we study the effect of the point-like
non-magnetic impurities on the superconducting state of La-327 and show that
$s_{\pm}$-wave and $d$-wave symmetries show a very different behavior as a
function of impurity concentration, which can be studied experimentally by
irradiating the La-327 samples by electrons prior applying the pressure. While
$d-$wave superconducting state will be conventionally suppressed, the
$s_{\pm}$-wave state shows more subtle behavior, depending on the asymmetry
between bonding and antibonding subspaces. For the electronic structure,
predicted to realize in La-327, the $s_{\pm}-$wave state will be robust against
complete suppression and the transition temperature, $T_c$ demonstrates a
transition from convex to concave behavior, indicating a crossover from
$s_{\pm}$-wave to $s_{++}$-wave symmetry as a function of impurity
concentration. We further analyze the sensitivity of the obtained results with
respect to the potential electronic structure modification.
| [
{
"version": "v1",
"created": "Mon, 4 Nov 2024 09:58:03 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Bötzel",
"Steffen",
""
],
[
"Lechermann",
"Frank",
""
],
[
"Shibauchi",
"Takasada",
""
],
[
"Eremin",
"Ilya M.",
""
]
] |
2411.02253 | Johannes Teutsch | Oleksii Molodchyk, Johannes Teutsch, Timm Faulwasser | Towards safe Bayesian optimization with Wiener kernel regression | null | null | null | null | stat.ML cs.LG cs.SY eess.SY math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bayesian Optimization (BO) is a data-driven strategy for
minimizing/maximizing black-box functions based on probabilistic surrogate
models. In the presence of safety constraints, the performance of BO crucially
relies on tight probabilistic error bounds related to the uncertainty
surrounding the surrogate model. For the case of Gaussian Process surrogates
and Gaussian measurement noise, we present a novel error bound based on the
recently proposed Wiener kernel regression. We prove that under rather mild
assumptions, the proposed error bound is tighter than bounds previously
documented in the literature, leading to enlarged safety regions. We draw upon
a numerical example to demonstrate the efficacy of the proposed error bound in
safe BO.
| [
{
"version": "v1",
"created": "Mon, 4 Nov 2024 16:43:16 GMT"
},
{
"version": "v2",
"created": "Tue, 26 Nov 2024 15:57:54 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 11:36:12 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Molodchyk",
"Oleksii",
""
],
[
"Teutsch",
"Johannes",
""
],
[
"Faulwasser",
"Timm",
""
]
] |
2411.02489 | Lucas Pupim | Lucas V. Pupim and Mathias S. Scheurer | Adatom engineering magnetic order in superconductors: Applications to
altermagnetic superconductivity | 10 pages, 6 figures | Phys. Rev. Lett. 134, 146001 (2025) | 10.1103/PhysRevLett.134.146001 | null | cond-mat.supr-con cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | We study theoretically how superlattices based on adatoms on surfaces of
unconventional superconductors can be used to engineer novel pairing states
that break time-reversal symmetry and exhibit non-trivial magnetic point
symmetries. We illustrate this using a square-lattice Hubbard model with
$d$-wave superconductivity and a subleading $s$-wave state as an example. An
adatom superlattice with square-lattice symmetries is shown to stabilize an
"orbital-altermagnetic superconductor'', a state that exhibits loop current
patterns and associated orbital magnetic moments, which preserve superlattice
translations but are odd under four-fold rotations. This state is further
characterized by a non-zero Berry curvature quadrupole moment and, upon
including spin-orbit coupling, by an altermagnetic spin splitting of the bands
and non-trivial spin textures in the superlattice unit cell, with zero net spin
moment.
| [
{
"version": "v1",
"created": "Mon, 4 Nov 2024 19:00:00 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Pupim",
"Lucas V.",
""
],
[
"Scheurer",
"Mathias S.",
""
]
] |
2411.03577 | Hisashi Morioka | Kazunori Ando, Hiroshi Isozaki, Hisashi Morioka | A remark on the absence of eigenvalues in continuous spectra for
discrete Schr\"{o}dinger operators on periodic lattices | null | null | null | null | math.SP math-ph math.AP math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove a Rellich-Vekua type theorem for Schr\"{o}dinger operators with
exponentially decreasing potentials on a class of lattices including square,
triangular, hexagonal lattices and their ladders. We also discuss the unique
continuation theorem and the non-existence of eigenvalues embedded in the
continuous spectrum.
| [
{
"version": "v1",
"created": "Wed, 6 Nov 2024 00:35:00 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 11:13:01 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Ando",
"Kazunori",
""
],
[
"Isozaki",
"Hiroshi",
""
],
[
"Morioka",
"Hisashi",
""
]
] |
2411.03689 | Xiaoming Wang | Jack Coleman, Daozhi Han, Xiaoming Wang | An efficient scheme for approximating long-time dynamics of a class of
non-linear models | null | null | null | null | math.NA cs.NA math.DS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We propose a novel, highly efficient, mean-reverting-SAV-BDF2-based,
long-time unconditionally stable numerical scheme for a class of
finite-dimensional nonlinear models important in geophysical fluid dynamics.
The scheme is highly efficient in that only a fixed symmetric positive definite
linear problem (with varying right-hand sides) is solved at each time step. The
solutions remain uniformly bounded for all time. We show that the scheme
accurately captures the long-time dynamics of the underlying geophysical model,
with the global attractors and invariant measures of the scheme converging to
those of the original model as the step size approaches zero.
In our numerical experiments, we adopt an indirect approach, using statistics
from long-time simulations to approximate the invariant measures. Our results
suggest that the convergence rate of the long-term statistics, as a function of
terminal time, is approximately first-order under the Jensen-Shannon metric and
half-order under the total variation metric. This implies that extremely long
simulations are required to achieve high-precision approximations of the
invariant measure (or climate). Nevertheless, the second-order scheme
significantly outperforms its first-order counterpart, requiring far less time
to reach a small neighborhood of statistical equilibrium for a given step size.
| [
{
"version": "v1",
"created": "Wed, 6 Nov 2024 06:14:46 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 12:00:56 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Apr 2025 06:55:17 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Coleman",
"Jack",
""
],
[
"Han",
"Daozhi",
""
],
[
"Wang",
"Xiaoming",
""
]
] |
2411.03915 | Stefan Karpitschka | Youchuang Chao, Hansol Jeon, Stefan Karpitschka | Non-monotonic motion of sliding droplets on strained soft solids | 7 pages, 4 figures, 3 ancillary files | null | null | null | cond-mat.soft | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Soft materials are ubiquitous in technological applications that require
deformability, for instance, in flexible, water-repellent coatings. However,
the wetting properties of pre-strained soft materials are only beginning to be
explored. Here we study the sliding dynamics of droplets on pre-strained soft
silicone gels, both in tension and in compression. Intriguingly, in compression
we find a non-monotonic strain dependence of the sliding speed: mild
compressions decelerate the droplets, but stronger compressions lead again to
faster droplet motion. Upon further compression, creases nucleate under the
droplets until finally, the entire surface undergoes the creasing instability,
causing a ``run-and-stop" motion. We quantitatively elucidate the speed
modification for moderate pre-strains by incremental viscoelasticity, while the
acceleration for larger pre-strains turns out to be linked to the solid
pressure, presumably through a lubrication effect of expelled oligomers.
| [
{
"version": "v1",
"created": "Wed, 6 Nov 2024 13:47:32 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 09:04:55 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Chao",
"Youchuang",
""
],
[
"Jeon",
"Hansol",
""
],
[
"Karpitschka",
"Stefan",
""
]
] |
2411.03961 | Yang Zhang | Yang Zhang and Xuan Ye | Regularized stress tensor of vector fields in de Sitter space | 42 pages, 10 figures | Universe 11, 72 (2025) | 10.3390/universe11020072 | null | gr-qc math-ph math.MP quant-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We study the Stueckelberg field in de Sitter space, which is a massive vector
field with the gauge fixing (GF) term $\frac{1}{2\zeta} (A^\mu\,_{;\, \mu})^2$.
We obtain the vacuum stress tensor, which consists of the transverse,
longitudinal, temporal, and GF parts, and each contains various UV divergences.
By the minimal subtraction rule, we regularize each part of the stress tensor
to its pertinent adiabatic order. The transverse stress tensor is regularized
to the 0th adiabatic order, the longitudinal, temporal, and GF stress tensors
are regularized to the 2nd adiabatic order. The resulting total regularized
vacuum stress tensor is convergent and maximally-symmetric, has a positive
energy density, and respects the covariant conservation, and thus can be
identified as the cosmological constant that drives the de Sitter inflation.
Under the Lorenz condition $A^\mu\,_{;\, \mu}=0$, the regularized Stueckelberg
stress tensor reduces to the regularized Proca stress tensor that contains only
the transverse and longitudinal modes. In the massless limit, the regularized
Stueckelberg stress tensor becomes zero, and is the same as that of the Maxwell
field with the GF term, and no trace anomaly exists. If the order of adiabatic
regularization were lower than our prescription, some divergences would remain.
If the order were higher, say, under the conventional 4th-order regularization,
more terms than necessary would be subtracted off, leading to an unphysical
negative energy density and the trace anomaly simultaneously.
| [
{
"version": "v1",
"created": "Wed, 6 Nov 2024 14:47:48 GMT"
},
{
"version": "v2",
"created": "Thu, 7 Nov 2024 02:18:10 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 09:16:20 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhang",
"Yang",
""
],
[
"Ye",
"Xuan",
""
]
] |
2411.05465 | Beno\^it Mahault | Yu Duan and Jaime Agudo-Canalejo and Ramin Golestanian and Beno\^it
Mahault | Phase Coexistence in Nonreciprocal Quorum-Sensing Active Matter | 22 pages, 13 figures. Comments are welcome | Phys. Rev. Research 7, 013234 (2025) | 10.1103/PhysRevResearch.7.013234 | null | cond-mat.soft cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | Motility and nonreciprocity are two primary mechanisms for self-organization
in active matter. In a recent study [Phys. Rev. Lett. 131, 148301 (2023)], we
explored their joint influence in a minimal model of two-species quorum-sensing
active particles interacting via mutual motility regulation. Our results
notably revealed a highly dynamic phase of chaotic chasing bands that is absent
when either nonreciprocity or self-propulsion is missing. Here, we examine
further the phase behavior of nonreciprocal quorum-sensing active particles,
distinguishing between the regimes of weak and strong nonreciprocity. In the
weakly nonreciprocal regime, this system exhibits multi-component
motility-induced phase separation. We establish an analytical criterion for the
associated phase coexistence, enabling a quantitative prediction of the phase
diagram. For strong nonreciprocity, where the dynamics is chase-and-run-like,
we numerically determine the phase behavior and show that it strongly depends
on the scale of observation. In small systems, our numerical simulations reveal
a phenomenology consistent with phenomenological models, comprising traveling
phase-separated domains and spiral-like defect patterns. However, we show that
these structures are generically unstable in large systems, where they are
superseded by bulk phase coexistence between domains that are either
homogeneous or populated by mesoscopic chasing bands. Crucially, this implies
that collective motion totally vanishes at large scales, while the breakdown of
our analytical criterion for this phase coexistence with multi-scale structures
prevents us from predicting the corresponding phase diagram.
| [
{
"version": "v1",
"created": "Fri, 8 Nov 2024 10:34:48 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Duan",
"Yu",
""
],
[
"Agudo-Canalejo",
"Jaime",
""
],
[
"Golestanian",
"Ramin",
""
],
[
"Mahault",
"Benoît",
""
]
] |
2411.05732 | Benjamin Hardin | Yandika Sirgabsou, Benjamin Hardin, Fran\c{c}ois Leblanc, Efi Raili,
Pericle Salvini, David Jackson, Marina Jirotka, and Lars Kunze | A risk model and analysis method for the psychological safety of human
and autonomous vehicles interaction | null | null | null | null | cs.HC cs.CY | http://creativecommons.org/licenses/by-sa/4.0/ | This paper addresses the critical issue of psychological safety in the design
and operation of autonomous vehicles, which are increasingly integrated with
artificial intelligence technologies. While traditional safety standards focus
primarily on physical safety, this paper emphasizes the psychological
implications that arise from human interactions with autonomous vehicles,
highlighting the importance of trust and perceived risk as significant factors
influencing user acceptance. Through a review of existing safety techniques,
the paper defines psychological safety in the context of autonomous vehicles,
proposes a risk model to identify and assess psychological risks, and adopts a
system-theoretic analysis method. The paper illustrates the potential
psychological hazards using a scenario involving a family's experience with an
autonomous vehicle, aiming to systematically evaluate situations that could
lead to psychological harm. By establishing a framework that incorporates
psychological safety alongside physical safety, the paper contributes to the
broader discourse on the safe deployment of autonomous vehicle and aims to
guide future developments in user-cantered design and regulatory practices.
| [
{
"version": "v1",
"created": "Fri, 8 Nov 2024 17:44:40 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 03:34:11 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Sirgabsou",
"Yandika",
""
],
[
"Hardin",
"Benjamin",
""
],
[
"Leblanc",
"François",
""
],
[
"Raili",
"Efi",
""
],
[
"Salvini",
"Pericle",
""
],
[
"Jackson",
"David",
""
],
[
"Jirotka",
"Marina",
""
],
[
"Kunze",
"Lars",
""
]
] |
2411.06145 | Pengyu Xue | Pengyu Xue, Linhao Wu, Zhen Yang, Chengyi Wang, Xiang Li, Yuxiang
Zhang, Jia Li, Ruikai Jin, Yifei Pei, Zhaoyan Shen, Xiran Lyu and Jacky Wai
Keung | ClassEval-T: Evaluating Large Language Models in Class-Level Code
Translation | null | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, Large Language Models (LLMs) have dramatically advanced the
performance of automated code translation, making their computational accuracy
score reach up to over 80% on many previous benchmarks. However, most code
samples in these benchmarks are short, standalone, statement/method-level, and
algorithmic, which is not aligned with practical coding tasks. Therefore, it is
still unknown the actual capability of LLMs in translating code samples written
for daily development. To achieve this, we construct a class-level code
translation benchmark, ClassEval-T, and make the first attempt to extensively
assess recent LLMs' performance on class-level code translation. ClassEval-T is
extended from ClassEval, a well-known class-level Python code generation
benchmark consisting of multiple practical coding topics, such as database
operation and game design, and diverse contextual dependencies (e.g., fields,
methods, and libraries). It cost us 360 person-hours to accomplish the manual
migration to Java and C++ with complete code samples and associated test
suites. Subsequently, we design three translation strategies (i.e., holistic,
min-dependency, and standalone) for class-level code translations and evaluate
eight recent LLMs of commercial, general, and code kinds in diverse families
and sizes on ClassEval-T. Experimental results demonstrate a remarkable
performance drop compared with the most widely studied method-level code
translation benchmark, and obvious discrepancies among LLMs appear, showing the
effectiveness of ClassEval-T in measuring recent LLMs. Afterwards, we further
discuss the usage scenarios for diverse translation strategies and LLMs'
ability to dependency awareness when translating class samples. Finally, 1,243
failure cases made by the best-performing LLM under test are analyzed and
categorized in this paper for practical guidance and future enlightenment.
| [
{
"version": "v1",
"created": "Sat, 9 Nov 2024 11:13:14 GMT"
},
{
"version": "v2",
"created": "Tue, 19 Nov 2024 07:19:27 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Mar 2025 12:36:51 GMT"
},
{
"version": "v4",
"created": "Mon, 14 Apr 2025 08:45:07 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Xue",
"Pengyu",
""
],
[
"Wu",
"Linhao",
""
],
[
"Yang",
"Zhen",
""
],
[
"Wang",
"Chengyi",
""
],
[
"Li",
"Xiang",
""
],
[
"Zhang",
"Yuxiang",
""
],
[
"Li",
"Jia",
""
],
[
"Jin",
"Ruikai",
""
],
[
"Pei",
"Yifei",
""
],
[
"Shen",
"Zhaoyan",
""
],
[
"Lyu",
"Xiran",
""
],
[
"Keung",
"Jacky Wai",
""
]
] |
2411.06770 | Zhijie Chen | Zhijie Chen, Qiaobo Li, Arindam Banerjee | Sketched Adaptive Federated Deep Learning: A Sharp Convergence Analysis | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Combining gradient compression methods (e.g., CountSketch, quantization) and
adaptive optimizers (e.g., Adam, AMSGrad) is a desirable goal in federated
learning (FL), with potential benefits on both fewer communication rounds and
less per-round communication. In spite of the preliminary empirical success of
sketched adaptive methods, existing convergence analyses show the communication
cost to have a linear dependence on the ambient dimension, i.e., number of
parameters, which is prohibitively high for modern deep learning models. In
this work, we introduce specific sketched adaptive federated learning (SAFL)
algorithms and, as our main contribution, provide theoretical convergence
analyses in different FL settings with guarantees on communication cost
depending only logarithmically (instead of linearly) on the ambient dimension.
Unlike existing analyses, we show that the entry-wise sketching noise existent
in the preconditioners and the first moments of SAFL can be implicitly
addressed by leveraging the recently-popularized anisotropic curvatures in deep
learning losses, e.g., fast decaying loss Hessian eigen-values. In the i.i.d.
client setting of FL, we show that SAFL achieves asymptotic $O(1/\sqrt{T})$
convergence, and converges faster in the initial epochs. In the non-i.i.d.
client setting, where non-adaptive methods lack convergence guarantees, we show
that SACFL (SAFL with clipping) algorithms can provably converge in spite of
the additional heavy-tailed noise. Our theoretical claims are supported by
empirical studies on vision and language tasks, and in both fine-tuning and
training-from-scratch regimes. Surprisingly, as a by-product of our analysis,
the proposed SAFL methods are competitive with the state-of-the-art
communication-efficient federated learning algorithms based on error feedback.
| [
{
"version": "v1",
"created": "Mon, 11 Nov 2024 07:51:22 GMT"
},
{
"version": "v2",
"created": "Tue, 12 Nov 2024 02:24:58 GMT"
},
{
"version": "v3",
"created": "Sun, 13 Apr 2025 03:44:21 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Chen",
"Zhijie",
""
],
[
"Li",
"Qiaobo",
""
],
[
"Banerjee",
"Arindam",
""
]
] |
2411.06854 | Qi Song | Qi Song, Jie Lou, Yan Chen | Condensation of Composite Bosonic Trions in Interacting Bose-Fermi
Mixtures | 5 pages,5 figures | null | null | null | cond-mat.quant-gas | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We reveal a quantum coherent state characterized by composite bosonic trions,
wherein paired fermions further bind with bosons, in one-dimensional Bose-Fermi
mixtures. This phase emerges in two distinct models, featuring onsite
boson-fermion attraction that induces negative binding energy for the composite
trions. The first is the Bose-Fermi pair-hopping model, in which spinless
fermions undergo pair hopping to form preformed pairs. The second is the
extended Bose-Fermi Hubbard model, which describes dipolar bosons and fermions
with density-density interactions. Notably, the formation of composite trions
is independent of the specific pairing interactions of fermions. Extensive
density matrix renormalization group calculations demonstrate the
quasi-condensation of the composite trions, evidenced by algebraically decaying
correlations of trions, gapped single particle excitations, and suppressed
fermion pair correlations. Our findings provide novel insights into a general
three-body pairing mechanism among particles with distinct quantum statistics.
| [
{
"version": "v1",
"created": "Mon, 11 Nov 2024 10:37:02 GMT"
},
{
"version": "v2",
"created": "Sun, 8 Dec 2024 09:02:46 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Apr 2025 06:04:11 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Song",
"Qi",
""
],
[
"Lou",
"Jie",
""
],
[
"Chen",
"Yan",
""
]
] |
2411.06886 | Emilio Lauret | Emilio A. Lauret and Juan Sebasti\'an Rodr\'iguez | Spectrally distinguishing symmetric spaces II | null | null | null | null | math.DG math.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The action of the subgroup $\operatorname{G}_2$ of $\operatorname{SO}(7)$
(resp.\ $\operatorname{Spin}(7)$ of $\operatorname{SO}(8)$) on the Grassmannian
space
$M=\frac{\operatorname{SO}(7)}{\operatorname{SO}(5)\times\operatorname{SO}(2)}$
(resp.\
$M=\frac{\operatorname{SO}(8)}{\operatorname{SO}(5)\times\operatorname{SO}(3)}$)
is still transitive. We prove that the spectrum (i.e.\ the collection of
eigenvalues of its Laplace-Beltrami operator) of a symmetric metric $g_0$ on
$M$ coincides with the spectrum of a $\operatorname{G}_2$-invariant (resp.\
$\operatorname{Spin}(7)$-invariant) metric $g$ on $M$ only if $g_0$ and $g$ are
isometric. As a consequence, each non-flat compact irreducible symmetric space
of non-group type is spectrally unique among the family of all currently known
homogeneous metrics on its underlying differentiable manifold.
| [
{
"version": "v1",
"created": "Mon, 11 Nov 2024 11:38:48 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 19:00:14 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Lauret",
"Emilio A.",
""
],
[
"Rodríguez",
"Juan Sebastián",
""
]
] |
2411.07027 | Masatoshi Yamada | Linlin Huang, Mamiya Kawaguchi, Yadikaer Maitiniyazi, Shinya
Matsuzaki, Akio Tomiya and Masatoshi Yamada | Functional renormalization group study of a four-fermion model with $CP$
violation | 17 pages, 3 figures, version published in Phys. Rev. D | Phys.Rev.D 111 (2025) 7, 074015 | 10.1103/PhysRevD.111.074015 | null | hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | We perform a functional renormalization group analysis of a four-fermion
model with $CP$ and $P$ violation to explore the nonperturbative infrared
dynamics of quantum chromodynamics (QCD) within the Wilsonian renormalization
group framework, particularly in the context of spontaneous $CP$-violation
models. Our analysis of the fixed-point structure reveals that, in the
large-$N_c$ limit, the $CP$-violating $\bar{\theta}$ parameter is dynamically
induced and approaches $\pi \cdot (N_f/2)$ (where $N_f$ is the number of
flavors) as the system enters the chirally broken phase. This behavior arises
due to criticality and the large anomalous dimensions of the $U(1)_A$-violating
four-fermion couplings. Furthermore, this trend appears to persist beyond the
leading large-$N_c$ approximation, provided that the infrared dynamics of QCD
remains dominated by the scalar condensate of the quark bilinear, as expected.
Notably, our findings highlight that $CP$-violating four-fermion interactions,
which are perturbatively irrelevant, can become relevant in the chirally broken
phase through nonperturbative effects, with potential implications for
spontaneous $CP$-violation scenarios.
| [
{
"version": "v1",
"created": "Mon, 11 Nov 2024 14:33:49 GMT"
},
{
"version": "v2",
"created": "Fri, 4 Apr 2025 04:27:17 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Huang",
"Linlin",
""
],
[
"Kawaguchi",
"Mamiya",
""
],
[
"Maitiniyazi",
"Yadikaer",
""
],
[
"Matsuzaki",
"Shinya",
""
],
[
"Tomiya",
"Akio",
""
],
[
"Yamada",
"Masatoshi",
""
]
] |
2411.07107 | Brian DuSell | Alexandra Butoi and Ghazal Khalighinejad and Anej Svete and Josef
Valvoda and Ryan Cotterell and Brian DuSell | Training Neural Networks as Recognizers of Formal Languages | 44 pages, 3 figures. ICLR 2025. Official camera-ready version;
applies minor corrections to previous version | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Characterizing the computational power of neural network architectures in
terms of formal language theory remains a crucial line of research, as it
describes lower and upper bounds on the reasoning capabilities of modern AI.
However, when empirically testing these bounds, existing work often leaves a
discrepancy between experiments and the formal claims they are meant to
support. The problem is that formal language theory pertains specifically to
recognizers: machines that receive a string as input and classify whether it
belongs to a language. On the other hand, it is common instead to evaluate
language models on proxy tasks, e.g., language modeling or sequence-to-sequence
transduction, that are similar in only an informal sense to the underlying
theory. We correct this mismatch by training and evaluating neural networks
directly as binary classifiers of strings, using a general method that can be
applied to a wide variety of languages. As part of this, we extend an algorithm
recently proposed by Sn{\ae}bjarnarson et al. (2025) for efficient
length-controlled sampling of strings from regular languages. We provide
results on a variety of languages across the Chomsky hierarchy for three neural
architectures: a simple RNN, an LSTM, and a causally-masked transformer. We
find that the RNN and LSTM often outperform the transformer, and that auxiliary
training objectives such as language modeling can help, although no single
objective uniformly improves performance across languages and architectures.
Our contributions will facilitate theoretically sound empirical testing of
language recognition claims in future work. We have released our datasets as a
benchmark called FLaRe (Formal Language Recognition), along with our code.
| [
{
"version": "v1",
"created": "Mon, 11 Nov 2024 16:33:25 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 14:51:27 GMT"
},
{
"version": "v3",
"created": "Fri, 11 Apr 2025 21:42:48 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Butoi",
"Alexandra",
""
],
[
"Khalighinejad",
"Ghazal",
""
],
[
"Svete",
"Anej",
""
],
[
"Valvoda",
"Josef",
""
],
[
"Cotterell",
"Ryan",
""
],
[
"DuSell",
"Brian",
""
]
] |
2411.07273 | Jeffrey Considine | Jeffrey Considine | Compressed Game Solving | Submitted to Computers and Games 2024 | null | null | null | cs.DS | http://creativecommons.org/licenses/by/4.0/ | We recast move generators for solving board games as operations on compressed
sets of strings. We aim for compressed representations with space sublinear in
the number of game positions for interesting sets of positions, move generation
in time roughly linear in the compressed size and membership tests in constant
time. To the extent that we achieve these tradeoffs empirically, we can
strongly solve board games in time sublinear in the state space. We demonstrate
this concept with the game Breakthrough where we empirically realize compressed
representations taking roughly $n^{0.5}$ to $n^{0.7}$ space to store relevant
sets of $n$ positions.
| [
{
"version": "v1",
"created": "Sun, 10 Nov 2024 18:46:59 GMT"
},
{
"version": "v2",
"created": "Fri, 15 Nov 2024 04:33:10 GMT"
},
{
"version": "v3",
"created": "Sun, 13 Apr 2025 01:36:27 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Considine",
"Jeffrey",
""
]
] |
2411.07696 | Jes\'us Alejandro Miranda Hern\'andez | Gabriel L\'opez Castro, Alejandro Miranda, Pablo Roig | Isospin breaking corrections in $2\pi$ production in tau decays and
$e^+e^-$ annihilation: consequences for the muon $g-2$ and CVC tests | 32 pages, 10 figures, 14 tables. Minor changes. Version accepted for
publication in PRD | Phys. Rev. D 111(2025)073004 | 10.1103/PhysRevD.111.073004 | null | hep-ph hep-ex hep-lat | http://creativecommons.org/licenses/by/4.0/ | We revisit the isospin-breaking corrections relating the $e^+e^-$ hadronic
cross-section and the tau decay spectral function, focusing on the di-pion
channel, that gives the dominant contribution to the hadronic vacuum
polarization piece of the muon $g-2$. We test different types of
electromagnetic and weak form factors and show that both, the Gounaris-Sakurai
and a dispersive-based approach, describe accurately $\tau$ lepton and $e^+e^-$
data (less when KLOE measurements are included in the fits) and comply
reasonably well with analyticity constraints. From these results we obtain the
isospin-breaking contribution to the conserved vector current (CVC) prediction
of the ${\rm BR}(\tau \to \pi\pi\nu_{\tau})$ and to the $2\pi$ hadronic vacuum
polarization (HVP) contribution to the muon $g-2$, in agreement with previous
determinations and with similar precision. Our results abound in the
convenience of using tau data-based results in the updated data-driven
prediction of the muon $g-2$ in the Standard Model.
| [
{
"version": "v1",
"created": "Tue, 12 Nov 2024 10:23:38 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Jan 2025 14:53:01 GMT"
},
{
"version": "v3",
"created": "Fri, 28 Mar 2025 10:16:47 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Castro",
"Gabriel López",
""
],
[
"Miranda",
"Alejandro",
""
],
[
"Roig",
"Pablo",
""
]
] |
2411.08253 | Nishanth Kumar | Nishanth Kumar, William Shen, Fabio Ramos, Dieter Fox, Tom\'as
Lozano-P\'erez, Leslie Pack Kaelbling, Caelan Reed Garrett | Open-World Task and Motion Planning via Vision-Language Model Inferred
Constraints | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Foundation models trained on internet-scale data, such as Vision-Language
Models (VLMs), excel at performing a wide variety of common sense tasks like
visual question answering. Despite their impressive capabilities, these models
cannot currently be directly applied to challenging robot manipulation problems
that require complex and precise continuous reasoning over long horizons. Task
and Motion Planning (TAMP) systems can control high-dimensional continuous
systems over long horizons via a hybrid search over traditional primitive robot
skills. However, these systems require detailed models of how the robot can
impact its environment, preventing them from directly interpreting and
addressing novel human objectives, for example, an arbitrary natural language
goal. We propose deploying VLMs within TAMP systems by having them generate
discrete and continuous language-parameterized constraints that enable TAMP to
reason about open-world concepts. Specifically, we propose algorithms for VLM
partial planning that constrain a TAMP system's discrete temporal search and
VLM continuous constraints interpretation to augment the traditional
manipulation constraints that TAMP systems seek to satisfy. Experiments
demonstrate that our approach -- OWL-TAMP -- outperforms several related
baselines, including those that solely use TAMP or VLMs for planning, across
several long-horizon manipulation tasks specified directly through natural
language. We additionally demonstrate that our approach is compatible with a
variety of TAMP systems and can be deployed to solve challenging manipulation
tasks on real-world hardware.
| [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 00:02:32 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Apr 2025 20:58:37 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Kumar",
"Nishanth",
""
],
[
"Shen",
"William",
""
],
[
"Ramos",
"Fabio",
""
],
[
"Fox",
"Dieter",
""
],
[
"Lozano-Pérez",
"Tomás",
""
],
[
"Kaelbling",
"Leslie Pack",
""
],
[
"Garrett",
"Caelan Reed",
""
]
] |
2411.08395 | Yuelin Zhang | Yuelin Zhang, Long Lei, Wanquan Yan, Tianyi Zhang, Raymond Shing-Yan
Tang, Shing Shin Cheng | MambaXCTrack: Mamba-based Tracker with SSM Cross-correlation and Motion
Prompt for Ultrasound Needle Tracking | Accepted by RAL | null | 10.1109/LRA.2025.3558377 | null | cs.CV cs.RO | http://creativecommons.org/licenses/by/4.0/ | Ultrasound (US)-guided needle insertion is widely employed in percutaneous
interventions. However, providing feedback on the needle tip position via US
imaging presents challenges due to noise, artifacts, and the thin imaging plane
of US, which degrades needle features and leads to intermittent tip visibility.
In this paper, a Mamba-based US needle tracker MambaXCTrack utilizing
structured state space models cross-correlation (SSMX-Corr) and implicit motion
prompt is proposed, which is the first application of Mamba in US needle
tracking. The SSMX-Corr enhances cross-correlation by long-range modeling and
global searching of distant semantic features between template and search maps,
benefiting the tracking under noise and artifacts by implicitly learning
potential distant semantic cues. By combining with cross-map interleaved scan
(CIS), local pixel-wise interaction with positional inductive bias can also be
introduced to SSMX-Corr. The implicit low-level motion descriptor is proposed
as a non-visual prompt to enhance tracking robustness, addressing the
intermittent tip visibility problem. Extensive experiments on a dataset with
motorized needle insertion in both phantom and tissue samples demonstrate that
the proposed tracker outperforms other state-of-the-art trackers while ablation
studies further highlight the effectiveness of each proposed tracking module.
| [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 07:27:56 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 03:51:04 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Zhang",
"Yuelin",
""
],
[
"Lei",
"Long",
""
],
[
"Yan",
"Wanquan",
""
],
[
"Zhang",
"Tianyi",
""
],
[
"Tang",
"Raymond Shing-Yan",
""
],
[
"Cheng",
"Shing Shin",
""
]
] |
2411.08561 | Wei Guan | Wei Guan, Jian Cao, Shiyou Qian, Jianqi Gao, Chun Ouyang | LogLLM: Log-based Anomaly Detection Using Large Language Models | null | null | null | null | cs.SE cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Software systems often record important runtime information in logs to help
with troubleshooting. Log-based anomaly detection has become a key research
area that aims to identify system issues through log data, ultimately enhancing
the reliability of software systems. Traditional deep learning methods often
struggle to capture the semantic information embedded in log data, which is
typically organized in natural language. In this paper, we propose LogLLM, a
log-based anomaly detection framework that leverages large language models
(LLMs). LogLLM employs BERT for extracting semantic vectors from log messages,
while utilizing Llama, a transformer decoder-based model, for classifying log
sequences. Additionally, we introduce a projector to align the vector
representation spaces of BERT and Llama, ensuring a cohesive understanding of
log semantics. Unlike conventional methods that require log parsers to extract
templates, LogLLM preprocesses log messages with regular expressions,
streamlining the entire process. Our framework is trained through a novel
three-stage procedure designed to enhance performance and adaptability.
Experimental results across four public datasets demonstrate that LogLLM
outperforms state-of-the-art methods. Even when handling unstable logs, it
effectively captures the semantic meaning of log messages and detects anomalies
accurately.
| [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 12:18:00 GMT"
},
{
"version": "v2",
"created": "Tue, 21 Jan 2025 06:12:26 GMT"
},
{
"version": "v3",
"created": "Wed, 29 Jan 2025 15:41:55 GMT"
},
{
"version": "v4",
"created": "Sun, 9 Feb 2025 15:15:12 GMT"
},
{
"version": "v5",
"created": "Mon, 14 Apr 2025 02:52:50 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Guan",
"Wei",
""
],
[
"Cao",
"Jian",
""
],
[
"Qian",
"Shiyou",
""
],
[
"Gao",
"Jianqi",
""
],
[
"Ouyang",
"Chun",
""
]
] |
2411.08763 | Christian Laudag\'e | Christian Laudag\'e, Felix-Benedikt Liebrich, J\"orn Sass | Multi-asset return risk measures | null | null | null | null | q-fin.MF | http://creativecommons.org/licenses/by/4.0/ | We revisit the recently introduced concept of return risk measures (RRMs) and
extend it by incorporating risk management via multiple so-called eligible
assets. The resulting new class of risk measures, termed multi-asset return
risk measures (MARRMs), introduces a novel economic model for multiplicative
risk sharing. We analyze properties of these risk measures. In particular, we
prove that a positively homogeneous MARRM is quasi-convex if and only if it is
convex. Furthermore, we provide conditions to avoid inconsistent risk
evaluations. Then, we point out the connection between MARRMs and the
well-known concept of multi-asset risk measures (MARMs). This is used to obtain
various dual representations of MARRMs. Moreover, we conduct a series of case
studies, in which we use typical continuous-time financial markets and
different notions of acceptability of losses to compare RRMs, MARMs, and MARRMs
and draw conclusions about the cost of risk mitigation.
| [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 16:49:23 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Apr 2025 08:47:00 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Laudagé",
"Christian",
""
],
[
"Liebrich",
"Felix-Benedikt",
""
],
[
"Sass",
"Jörn",
""
]
] |
2411.09227 | Shigeki Matsutani | Shigeki Matsutani | Euler's original derivation of elastica equation | null | null | null | null | math-ph math.DG math.MP nlin.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Euler derived the differential equations of elastica by the variational
method in 1744, but his original derivation has never been properly interpreted
or explained in terms of modern mathematics. We elaborate Euler's original
derivation of elastica and show that Euler used Noether's theorem concerning
the translational symmetry of elastica, although Noether published her theorem
in 1918. It is also shown that his equation is essentially the static modified
KdV equation which is obtained by the isometric and isoenergy conditions, known
as the Goldstein-Petrich scheme.
| [
{
"version": "v1",
"created": "Thu, 14 Nov 2024 06:43:59 GMT"
},
{
"version": "v2",
"created": "Sun, 17 Nov 2024 07:34:48 GMT"
},
{
"version": "v3",
"created": "Mon, 14 Apr 2025 14:01:17 GMT"
}
] | 2025-04-15T00:00:00 | [
[
"Matsutani",
"Shigeki",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.