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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2504.15167 | Micha Christoph | Simona Boyadzhiyska, Micha Christoph, Tibor Szab\'o | Almost-perfect colorful matchings in three-edge-colored bipartite graphs | 15 pages | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove that, for positive integers $n,a_1, a_2, a_3$ satisfying
$a_1+a_2+a_3 = n-1$, it holds that any bipartite graph $G$ which is the union
of three perfect matchings $M_1$, $M_2$, and $M_3$ on $2n$ vertices contains a
matching $M$ such that $|M\cap M_i| =a_i$ for $i= 1,2,$ and $3$. The bound
$n-1$ on the sum is best possible in general. Our result verifies the
multiplicity extension of the Ryser-Brualdi-Stein Conjecture, proposed recently
by Anastos, Fabian, M\"uyesser, and Szab\'o, for three colors.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:22:13 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Boyadzhiyska",
"Simona",
""
],
[
"Christoph",
"Micha",
""
],
[
"Szabó",
"Tibor",
""
]
] |
2504.15168 | Qilin Tian | Qilin Tian | On true empty category | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | According to Chomsky (1981, 1986), empty categories consist of PRO, pro,
trace, and variable. However, some empty object positions seem to be
incompatible with extant empty categories. Given this, Li (2007a, 2007b, 2014)
and Li & Wei (2014) raise the true empty category hypothesis, which holds that
true empty category is only an empty position with category and Case features.
As a last resort option, it is used mainly to meet the subcatgorization of a
verb. This assumption is ingenious, and if proved to be true, it will exert a
great impact on the study of UG. In this paper, we evaluate their evidence from
topicalization and demonstrate that it can be accounted for without invoking
true empty category.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:22:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Tian",
"Qilin",
""
]
] |
2504.15169 | Jie Xu | Xiaodan Wei, Shuyuan Zhang, Boyang Ma, Zehua Feng, chunhan Yu, Linyuan
Peng, Zijian Hua, Jie Xu | Improving efficiency and stability for perovskite solar cell with
diethylene glycol dimethacrylate modification | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | The humidity resistance is the key challenges that hinder the commercial
application of perovskite solar cells (PSCs). Herein, we propose an ultra-thin
acrylate polymer (diethylene glycol dimethacrylate, DGDMA) into perovskite
films to investigate the influence of polymerized networks on stability. The
monomer molecules containing acrylate and carbonyl groups were selected, and
the effects of the polymerized were quantified with different concentration.
The experimental results show that, when the concentration of DGDMA is 1 mg/ml,
the PCE increases from 18.06% to 21.82%, which is optimum. The monomer
molecules with carbonyl groups polymerize, they can chelate with uncoordinated
Pb2+ in perovskite films to improve the film quality, reduce the surface defect
density to decrease non-radiative recombination, and also significantly enhance
the humidity stability of PSCs.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:23:51 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wei",
"Xiaodan",
""
],
[
"Zhang",
"Shuyuan",
""
],
[
"Ma",
"Boyang",
""
],
[
"Feng",
"Zehua",
""
],
[
"Yu",
"chunhan",
""
],
[
"Peng",
"Linyuan",
""
],
[
"Hua",
"Zijian",
""
],
[
"Xu",
"Jie",
""
]
] |
2504.15170 | Chengxi Han | Chengxi Han, Xiaoyu Su, Zhiqiang Wei, Meiqi Hu, Yichu Xu | HSANET: A Hybrid Self-Cross Attention Network For Remote Sensing Change
Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The remote sensing image change detection task is an essential method for
large-scale monitoring. We propose HSANet, a network that uses hierarchical
convolution to extract multi-scale features. It incorporates hybrid
self-attention and cross-attention mechanisms to learn and fuse global and
cross-scale information. This enables HSANet to capture global context at
different scales and integrate cross-scale features, refining edge details and
improving detection performance. We will also open-source our model code:
https://github.com/ChengxiHAN/HSANet.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:23:59 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Han",
"Chengxi",
""
],
[
"Su",
"Xiaoyu",
""
],
[
"Wei",
"Zhiqiang",
""
],
[
"Hu",
"Meiqi",
""
],
[
"Xu",
"Yichu",
""
]
] |
2504.15171 | Meng Cui | Meng Cui, Xianghu Yue, Xinyuan Qian, Jinzheng Zhao, Haohe Liu, Xubo
Liu, Daoliang Li, Wenwu Wang | Audio-Visual Class-Incremental Learning for Fish Feeding intensity
Assessment in Aquaculture | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Fish Feeding Intensity Assessment (FFIA) is crucial in industrial aquaculture
management. Recent multi-modal approaches have shown promise in improving FFIA
robustness and efficiency. However, these methods face significant challenges
when adapting to new fish species or environments due to catastrophic
forgetting and the lack of suitable datasets. To address these limitations, we
first introduce AV-CIL-FFIA, a new dataset comprising 81,932 labelled
audio-visual clips capturing feeding intensities across six different fish
species in real aquaculture environments. Then, we pioneer audio-visual class
incremental learning (CIL) for FFIA and demonstrate through benchmarking on
AV-CIL-FFIA that it significantly outperforms single-modality methods. Existing
CIL methods rely heavily on historical data. Exemplar-based approaches store
raw samples, creating storage challenges, while exemplar-free methods avoid
data storage but struggle to distinguish subtle feeding intensity variations
across different fish species. To overcome these limitations, we introduce
HAIL-FFIA, a novel audio-visual class-incremental learning framework that
bridges this gap with a prototype-based approach that achieves exemplar-free
efficiency while preserving essential knowledge through compact feature
representations. Specifically, HAIL-FFIA employs hierarchical representation
learning with a dual-path knowledge preservation mechanism that separates
general intensity knowledge from fish-specific characteristics. Additionally,
it features a dynamic modality balancing system that adaptively adjusts the
importance of audio versus visual information based on feeding behaviour
stages. Experimental results show that HAIL-FFIA is superior to SOTA methods on
AV-CIL-FFIA, achieving higher accuracy with lower storage needs while
effectively mitigating catastrophic forgetting in incremental fish species
learning.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:24:34 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cui",
"Meng",
""
],
[
"Yue",
"Xianghu",
""
],
[
"Qian",
"Xinyuan",
""
],
[
"Zhao",
"Jinzheng",
""
],
[
"Liu",
"Haohe",
""
],
[
"Liu",
"Xubo",
""
],
[
"Li",
"Daoliang",
""
],
[
"Wang",
"Wenwu",
""
]
] |
2504.15172 | Oliver Fox | Stephon Alexander, Roger Andrews, Oliver Fox, Sarben Sarkar | Quantum Excitation Transfer in an Artificial Photosynthetic
Light-Harvesting System | 17 pages, 6 figures | null | null | null | physics.chem-ph quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We analytically derive transfer probabilities and efficiencies for an
artificial light-harvesting photosynthetic system, which consists of a ring
coupled to a central acceptor. For an incident photon pair, we find
near-perfect single excitation transfer efficiency with negligible double
excitation transfer in the weak coupling regime. In the strong coupling regime,
single excitation transfer efficiency was greater than 90%, while the double
excitation efficiency was approximately 50%. We have found that the three main
factors which determine high transfer efficiencies are large acceptor
probabilities, long acceptor decay times, and strong photon-ring coupling. A
possible implementation of the theoretical framework to bio-inspired solar
energy devices is also discussed.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:26:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Alexander",
"Stephon",
""
],
[
"Andrews",
"Roger",
""
],
[
"Fox",
"Oliver",
""
],
[
"Sarkar",
"Sarben",
""
]
] |
2504.15173 | Francesco Costanzo | Francesco Costanzo (1), Mohammad Jannesari (1), Beatrice Ghitti (1 and
2) ((1) Center for Neural Engineering, Engineering Science and Mechanics
Department, Penn State University, (2) Auckland Bioengineering Institute, The
University of Auckland) | Poroelastic flow across a permeable interface: a Hamilton's principle
approach and its finite element implementation | null | null | null | null | math.NA cs.NA physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | We consider fluid flow across a permeable interface within a deformable
porous medium. We use mixture theory. The mixture's constituents are assumed to
be incompressible in their pure form. We use Hamilton's principle to obtain the
governing equations, and we propose a corresponding finite element
implementation. The filtration velocity and the pore pressure are allowed to be
discontinuous across the interface while some control of these discontinuities
is built into the interfacial constitutive behavior. To facilitate the
practical implementation of the formulation in a finite element scheme, we
introduce a Lagrange multiplier field over the interface for the explicit
enforcement of the jump condition of the balance of mass. Our formulation
appears to recover some basic results from the literature. The novelty of the
work is the formulation of an approach that can accommodate specific
constitutive assumptions pertaining to the behavior of the interface that do
not necessarily imply the continuity of the filtration velocity and/or of the
pore pressure across it.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:26:35 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Costanzo",
"Francesco",
"",
"1 and\n 2"
],
[
"Jannesari",
"Mohammad",
"",
"1 and\n 2"
],
[
"Ghitti",
"Beatrice",
"",
"1 and\n 2"
]
] |
2504.15174 | Marek Rogatko | Marek Rogatko | Uniqueness of electrically and magnetically charged phantom wormholes | 17 pages, RevTex, to be published in Phys. Rev. D15 | null | null | null | gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | We prove the uniqueness theorem for static spherically symmetric traversable
wormholes with two asymptotically flat ends, constituting the solutions of
Einstein-phantom-electric-magnetic equations of motion. For the completeness of
the results, we also consider the case of phantom U(1)- gauge field. The key
role in the proof is bounded with conformal positive energy theorem.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:27:54 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Rogatko",
"Marek",
""
]
] |
2504.15175 | Robert H. Jonsson | Maximilian Goll, Robert H. Jonsson | Geometric speed limit of state preparation and curved control spaces | 7 pages, 5 figures | null | null | NORDITA 2025-021 | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The preparation of quantum many-body systems faces the difficulty that in a
realistic scenario only few control parameters of the system may be accessible.
In this context, an interesting connection between the energy fluctuations
during state preparation and its geometric length as measured by the
Fubini-Study metric was discussed in Bukov et al., "Geometric Speed Limit of
Accessible Many-Body State Preparation", Phys. Rev. X 9, 011034 (2019). An
inspiring conjecture was put forward lower bounding the energy fluctuations by
the minimal geometric length of all accessible state preparation protocols. We
here show that the conjecture holds if the accessible parameter space has no
extrinsic curvature, when embedded into the space of all dynamically accessible
states. If the parameter space has extrinsic curvature a weakened version of
the conjecture applies. We discuss instructive examples for a qubit system and
harmonic oscillators.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:32:29 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Goll",
"Maximilian",
""
],
[
"Jonsson",
"Robert H.",
""
]
] |
2504.15176 | Miaomiao Cai | Miaomiao Cai, Simiao Li, Wei Li, Xudong Huang, Hanting Chen, Jie Hu,
Yunhe Wang | DSPO: Direct Semantic Preference Optimization for Real-World Image
Super-Resolution | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in diffusion models have improved Real-World Image
Super-Resolution (Real-ISR), but existing methods lack human feedback
integration, risking misalignment with human preference and may leading to
artifacts, hallucinations and harmful content generation. To this end, we are
the first to introduce human preference alignment into Real-ISR, a technique
that has been successfully applied in Large Language Models and Text-to-Image
tasks to effectively enhance the alignment of generated outputs with human
preferences. Specifically, we introduce Direct Preference Optimization (DPO)
into Real-ISR to achieve alignment, where DPO serves as a general alignment
technique that directly learns from the human preference dataset. Nevertheless,
unlike high-level tasks, the pixel-level reconstruction objectives of Real-ISR
are difficult to reconcile with the image-level preferences of DPO, which can
lead to the DPO being overly sensitive to local anomalies, leading to reduced
generation quality. To resolve this dichotomy, we propose Direct Semantic
Preference Optimization (DSPO) to align instance-level human preferences by
incorporating semantic guidance, which is through two strategies: (a) semantic
instance alignment strategy, implementing instance-level alignment to ensure
fine-grained perceptual consistency, and (b) user description feedback
strategy, mitigating hallucinations through semantic textual feedback on
instance-level images. As a plug-and-play solution, DSPO proves highly
effective in both one-step and multi-step SR frameworks.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:35:48 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cai",
"Miaomiao",
""
],
[
"Li",
"Simiao",
""
],
[
"Li",
"Wei",
""
],
[
"Huang",
"Xudong",
""
],
[
"Chen",
"Hanting",
""
],
[
"Hu",
"Jie",
""
],
[
"Wang",
"Yunhe",
""
]
] |
2504.15177 | Matthew Zahr | Huijing Dong and Masayuki Yano and Tianci Huang and Matthew J. Zahr | An $rp$-adaptive method for accurate resolution of shock-dominated
viscous flow based on implicit shock tracking | 43 pages, 35 figures, | null | null | null | math.NA cs.NA math.OC physics.flu-dyn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work introduces an optimization-based $rp$-adaptive numerical method to
approximate solutions of viscous, shock-dominated flows using implicit shock
tracking and a high-order discontinuous Galerkin discretization on
traditionally coarse grids without nonlinear stabilization (e.g., artificial
viscosity or limiting). The proposed method adapts implicit shock tracking
methods, originally developed to align mesh faces with solution
discontinuities, to compress elements into viscous shocks and boundary layers,
functioning as a novel approach to aggressive $r$-adaptation. This form of
$r$-adaptation is achieved naturally as the minimizer of the enriched residual
with respect to the discrete flow variables and coordinates of the nodes of the
grid. Several innovations to the shock tracking optimization solver are
proposed to ensure sufficient mesh compression at viscous features to render
stabilization unnecessary, including residual weighting, step constraints and
modifications, and viscosity-based continuation. Finally, $p$-adaptivity is
used to locally increase the polynomial degree with three clear benefits: (1)
lessens the mesh compression requirements near shock waves and boundary layers,
(2) reduces the error in regions where $r$-adaptivity is not sufficient with
the given grid topology, and (3) reduces computational cost by performing a
majority of the $r$-adaptivity iterations on the coarsest discretization. A
series of numerical experiments show the proposed method effectively resolves
viscous, shock-dominated flows, including accurate prediction of heat flux
profiles produced by hypersonic flow over a cylinder, and compares favorably in
terms of accuracy per degree of freedom to $h$-adaptation with a high-order
discretization.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:39:42 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Dong",
"Huijing",
""
],
[
"Yano",
"Masayuki",
""
],
[
"Huang",
"Tianci",
""
],
[
"Zahr",
"Matthew J.",
""
]
] |
2504.15178 | Jinhai Hu | Jinhai Hu | Time-Series Analysis on Edge-AI Hardware for Healthcare Monitoring | 38 pages, 20 figures, Progress report for qualification cum PhD
confirmation exercise | null | null | null | eess.SP | http://creativecommons.org/licenses/by/4.0/ | This project addresses the need for efficient, real-time analysis of
biomedical signals such as electrocardiograms (ECG) and electroencephalograms
(EEG) for continuous health monitoring. Traditional methods rely on
long-duration data recording followed by offline analysis, which is
power-intensive and delays responses to critical symptoms such as arrhythmia.
To overcome these limitations, a time-domain ECG analysis model based on a
novel dynamically-biased Long Short-Term Memory (DB-LSTM) neural network is
proposed. This model supports simultaneous ECG forecasting and classification
with high performance-achieving over 98% accuracy and a normalized mean square
error below 1e-3 for forecasting, and over 97% accuracy with faster convergence
and fewer training parameters for classification. To enable edge deployment,
the model is hardware-optimized by quantizing weights to INT4 or INT3 formats,
resulting in only a 2% and 6% drop in classification accuracy during training
and inference, respectively, while maintaining full accuracy for forecasting.
Extensive simulations using multiple ECG datasets confirm the model's
robustness. Future work includes implementing the algorithm on FPGA and CMOS
circuits for practical cardiac monitoring, as well as developing a digital
hardware platform that supports flexible neural network configurations and
on-chip online training for personalized healthcare applications.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:40:10 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Hu",
"Jinhai",
""
]
] |
2504.15179 | Fei Yin | Fei Yin and Mallikarjun B R and Chun-Han Yao and Rafa{\l} Mantiuk and
Varun Jampani | FaceCraft4D: Animated 3D Facial Avatar Generation from a Single Image | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a novel framework for generating high-quality, animatable 4D
avatar from a single image. While recent advances have shown promising results
in 4D avatar creation, existing methods either require extensive multiview data
or struggle with shape accuracy and identity consistency. To address these
limitations, we propose a comprehensive system that leverages shape, image, and
video priors to create full-view, animatable avatars. Our approach first
obtains initial coarse shape through 3D-GAN inversion. Then, it enhances
multiview textures using depth-guided warping signals for cross-view
consistency with the help of the image diffusion model. To handle expression
animation, we incorporate a video prior with synchronized driving signals
across viewpoints. We further introduce a Consistent-Inconsistent training to
effectively handle data inconsistencies during 4D reconstruction. Experimental
results demonstrate that our method achieves superior quality compared to the
prior art, while maintaining consistency across different viewpoints and
expressions.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:40:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yin",
"Fei",
""
],
[
"R",
"Mallikarjun B",
""
],
[
"Yao",
"Chun-Han",
""
],
[
"Mantiuk",
"Rafał",
""
],
[
"Jampani",
"Varun",
""
]
] |
2504.15180 | Yoshitaro Tanaka | Hiroshi Ishii and Yoshitaro Tanaka | Reaction-diffusion approximation of nonlocal interactions in
high-dimensional Euclidean space | null | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In various phenomena such as pattern formation, neural firing in the brain
and cell migration, interactions that can affect distant objects globally in
space can be observed. These interactions are referred to as nonlocal
interactions and are often modeled using spatial convolution with an
appropriate integral kernel. Many evolution equations incorporating nonlocal
interactions have been proposed. In such equations, the behavior of the system
and the patterns it generates can be controlled by modifying the shape of the
integral kernel. However, the presence of nonlocality poses challenges for
mathematical analysis. To address these difficulties, we develop an
approximation method that converts nonlocal effects into spatially local
dynamics using reaction-diffusion systems. In this paper, we present an
approximation method for nonlocal interactions in evolution equations based on
a linear sum of solutions to a reaction-diffusion system in high-dimensional
Euclidean space up to three dimensions. The key aspect of this approach is
identifying a class of integral kernels that can be approximated by a linear
combination of specific Green functions in the case of high-dimensional spaces.
This enables us to demonstrate that any nonlocal interactions can be
approximated by a linear sum of auxiliary diffusive substances. Our results
establish a connection between a broad class of nonlocal interactions and
diffusive chemical reactions in dynamical systems.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:41:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ishii",
"Hiroshi",
""
],
[
"Tanaka",
"Yoshitaro",
""
]
] |
2504.15181 | Rokas Gipi\v{s}kis | Lily Stelling, Mick Yang, Rokas Gipi\v{s}kis, Leon Staufer, Ze Shen
Chin, Sim\'eon Campos, and Michael Chen | Existing Industry Practice for the EU AI Act's General-Purpose AI Code
of Practice Safety and Security Measures | 158 pages | null | null | null | cs.CY cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This report provides a detailed comparison between the measures proposed in
the EU AI Act's General-Purpose AI (GPAI) Code of Practice (Third Draft) and
current practices adopted by leading AI companies. As the EU moves toward
enforcing binding obligations for GPAI model providers, the Code of Practice
will be key to bridging legal requirements with concrete technical commitments.
Our analysis focuses on the draft's Safety and Security section which is only
relevant for the providers of the most advanced models (Commitments II.1-II.16)
and excerpts from current public-facing documents quotes that are relevant to
each individual measure.
We systematically reviewed different document types - including companies'
frontier safety frameworks and model cards - from over a dozen companies,
including OpenAI, Anthropic, Google DeepMind, Microsoft, Meta, Amazon, and
others. This report is not meant to be an indication of legal compliance nor
does it take any prescriptive viewpoint about the Code of Practice or
companies' policies. Instead, it aims to inform the ongoing dialogue between
regulators and GPAI model providers by surfacing evidence of precedent.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:44:01 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Stelling",
"Lily",
""
],
[
"Yang",
"Mick",
""
],
[
"Gipiškis",
"Rokas",
""
],
[
"Staufer",
"Leon",
""
],
[
"Chin",
"Ze Shen",
""
],
[
"Campos",
"Siméon",
""
],
[
"Chen",
"Michael",
""
]
] |
2504.15182 | Xianpan Zhou | Xianpan Zhou | Tiger200K: Manually Curated High Visual Quality Video Dataset from UGC
Platform | Project page: https://tinytigerpan.github.io/tiger200k/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The recent surge in open-source text-to-video generation models has
significantly energized the research community, yet their dependence on
proprietary training datasets remains a key constraint. While existing open
datasets like Koala-36M employ algorithmic filtering of web-scraped videos from
early platforms, they still lack the quality required for fine-tuning advanced
video generation models. We present Tiger200K, a manually curated high visual
quality video dataset sourced from User-Generated Content (UGC) platforms. By
prioritizing visual fidelity and aesthetic quality, Tiger200K underscores the
critical role of human expertise in data curation, and providing high-quality,
temporally consistent video-text pairs for fine-tuning and optimizing video
generation architectures through a simple but effective pipeline including shot
boundary detection, OCR, border detecting, motion filter and fine bilingual
caption. The dataset will undergo ongoing expansion and be released as an
open-source initiative to advance research and applications in video generative
models. Project page: https://tinytigerpan.github.io/tiger200k/
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:44:06 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhou",
"Xianpan",
""
]
] |
2504.15183 | Ger\'onimo Sequeiros | Ger\'onimo Sequeiros, Claudia M. S\'anchez, Lisandro Buljubasich, Ana
K. Chattah, Horacio M. Pastawski, Rodolfo H. Acosta | Multiple Quantum Many-Body Clustering Probed by Dynamical Decoupling | 7 pages, 4 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | The manipulation of quantum information in large systems requires precise
control of quantum systems that are out-of-equilibrium. As the size of the
system increases, its fragility in response to external perturbations and
intrinsic decoherence processes also increases. The degradation of the system
response makes accurate measurements a challenging and time-consuming task.
However, quantum information lifetime enhancement can be achieved by dynamical
decoupling techniques (DD), where an external drive with a frequency much
higher than the system's internal evolution renders signal acquisition with
decay times greater than 1000-fold. In this study, we demonstrate that the
system response during a prethermal period, subject to Floquet control, can be
utilized to probe the multiple quantum evolution of dense and highly connected
spin systems. This approach exhibits an enhanced sensitivity at a reduced
experimental time. The enhanced signal-to-noise ratio achieved enabled the use
of numerical inversion strategies to model the evolution of the excited
multiple quantum coherences, which describe the number of correlated spins
within a cluster. We observed for the first time, to the best of our knowledge,
that the increase in the number of correlated spins with multiple quantum
evolution is accompanied by an increase in the distribution of spin cluster
sizes, which follows a quadratic law.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:46:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sequeiros",
"Gerónimo",
""
],
[
"Sánchez",
"Claudia M.",
""
],
[
"Buljubasich",
"Lisandro",
""
],
[
"Chattah",
"Ana K.",
""
],
[
"Pastawski",
"Horacio M.",
""
],
[
"Acosta",
"Rodolfo H.",
""
]
] |
2504.15184 | Stanislav Semenov | Stanislav Semenov | Wave Arithmetic: A Smooth Integral Representation of Number Theory | 26 pages, submitted to arXiv | null | null | null | math.LO math.NT | http://creativecommons.org/licenses/by/4.0/ | We introduce Wave Arithmetic, a smooth analytical framework in which natural,
integer, and rational numbers are represented not as discrete entities, but as
integrals of smooth, compactly supported or periodic kernel functions. In this
formulation, each number arises as the accumulated amplitude of a structured
waveform -- an interference pattern encoded by carefully designed kernels.
Arithmetic operations such as addition, multiplication, and exponentiation are
realized as geometric and tensorial constructions over multidimensional
integration domains. Rational numbers emerge through amplitude scaling, and
negative values through sign inversion, preserving all classical arithmetic
identities within a continuous and differentiable structure. This
representation embeds number theory into the realm of smooth analysis, enabling
new interpretations of primality, factorization, and divisibility as geometric
and spectral phenomena. Beyond technical formulation, Wave Arithmetic proposes
a paradigm shift: numbers as the collapsed states of harmonic processes --
analytic resonances rather than atomic symbols.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:46:44 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Semenov",
"Stanislav",
""
]
] |
2504.15185 | Andy Wanna | Andy Wanna (1), Hanqiu Chen (1), Cong Hao (1) ((1) Georgia Institute
of Technology) | ForgeBench: A Machine Learning Benchmark Suite and Auto-Generation
Framework for Next-Generation HLS Tools | null | null | null | null | cs.AR | http://creativecommons.org/licenses/by/4.0/ | Although High-Level Synthesis (HLS) has attracted considerable interest in
hardware design, it has not yet become mainstream due to two primary
challenges. First, current HLS hardware design benchmarks are outdated as they
do not cover modern machine learning (ML) applications, preventing the rigorous
development of HLS tools on ML-focused hardware design. Second, existing HLS
tools are outdated because they predominantly target individual accelerator
designs and lack an architecture-oriented perspective to support common
hardware module extraction and reuse, limiting their adaptability and broader
applicability. Motivated by these two limitations, we propose ForgeBench, an
ML-focused benchmark suite with a hardware design auto-generation framework for
next-generation HLS tools. In addition to the auto-generation framework, we
provide two ready-to-use benchmark suites. The first contains over 6,000
representative ML HLS designs. We envision future HLS tools being
architecture-oriented, capable of automatically identifying common
computational modules across designs, and supporting flexible dataflow and
control. Accordingly, the second benchmark suite includes ML HLS designs with
possible resource sharing manually implemented to highlight the necessity of
architecture-oriented design, ensuring it is future-HLS ready. ForgeBench is
open-sourced at https://github.com/hchen799/ForgeBench .
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:49:27 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wanna",
"Andy",
""
],
[
"Chen",
"Hanqiu",
""
],
[
"Hao",
"Cong",
""
]
] |
2504.15186 | Therar Kadri | Therrar Kadri, Rahil Omairi, Khaled Smaili, Seifedine Kadry | Sum of Independent XGamma Distributions | 14 pages, 1 figure | null | null | null | math.ST math.PR stat.TH | http://creativecommons.org/publicdomain/zero/1.0/ | The XGamma distribution is a generated distribution from a mixture of
Exponential and Gamma distributions. It is found that in many cases the XGamma
has more flexibility than the Exponential distribution. In this paper we
consider the sum of independent XGamma distributions with different parameters.
We showed that the probability density function of this distribution is a sum
of the probability density function of the Erlang distributions. As a
consequence, we find exact closed expressions of the other related statistical
functions. Next, we examine the estimation of the parameters by maximum
likelihood estimators. We observe in an applications a real data set which
shows that this model provides better fit to the data as compared to the sum of
the Exponential distributions, the Hypoexponential models.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:55:29 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kadri",
"Therrar",
""
],
[
"Omairi",
"Rahil",
""
],
[
"Smaili",
"Khaled",
""
],
[
"Kadry",
"Seifedine",
""
]
] |
2504.15187 | John Stenger | John P. T. Stenger, Gloria Bazargan, Nicholas T. Bronn, Daniel
Gunlycke | Method for simulating open-system dynamics using mid-circuit
measurements on a quantum computer | 7 pages, figures, 1 table | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | We present a method for simulating the dynamics of an open electronic system
on a quantum computer. This approach entails mid-circuit measurements and
resets to simulate the addition or removal of fermions from the system and
provides a way to apply non-reversible operations to a quantum computer. Using
this method, we simulate the dynamics of an open electronic system consisting
of a fermionic chain positioned between two conductive leads on the
$ibm\_torino$ quantum computer.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 15:57:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Stenger",
"John P. T.",
""
],
[
"Bazargan",
"Gloria",
""
],
[
"Bronn",
"Nicholas T.",
""
],
[
"Gunlycke",
"Daniel",
""
]
] |
2504.15191 | Charles Klewicki | Charles Klewicki, Ari Schenkman, Geoffrey Spedding | Aerodynamic Control of Laminar Separation on a Wall-Bounded Airfoil at
Transitional Reynolds Numbers | 32 pages, 28 figures | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | Experiments were conducted in a low-turbulence wind tunnel to investigate the
efficacy of localised acoustic forcing upon the dynamics and stability of the
flow on a cambered, wall-bounded airfoil over a range of Reynolds numbers (Re)
where the flow state can switch between two limits -- a low-lift state (SI)
where separation continues beyond the trailing edge and a high-lift state (SII)
where the separated flow is closed off to form a laminar separation bubble. The
switching between SI and SII can occur close to a critical angle of attack
($\alpha_{\textrm{crit}}$) which varies with $\textrm{Re}$. The most effective
forcing frequencies are found at a constant value of a rescaled Strouhal
number, $\textrm{St}^* = \textrm{St}/\textrm{Re}^{1/2}= 0.027$, which indicates
that though the primary unstable modes of the separated shear layer are of the
inviscid, Kelvin-Helmholtz type, these modes are seeded by length scales that
originate in the laminar (viscous) boundary layer. The most effective chordwise
forcing location varies with $\textrm{St}/\textrm{Re}^{1/2}$ and incidence
angle, $\alpha$, and is always upstream of the separation point. Although the
boundary layer flows are far from two-dimensional, forcing at a fixed chord
location across all spanwise locations is effective in controlling the SI --
SII transition. Strategies for active and passive feedback control are
suggested.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:01:41 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Klewicki",
"Charles",
""
],
[
"Schenkman",
"Ari",
""
],
[
"Spedding",
"Geoffrey",
""
]
] |
2504.15192 | Yaqian Chen | Yaqian Chen, Lin Li, Hanxue Gu, Haoyu Dong, Derek L. Nguyen, Allan D.
Kirk, Maciej A. Mazurowski, E. Shelley Hwang | Breast density in MRI: an AI-based quantification and relationship to
assessment in mammography | 13 pages, 5 figures | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Mammographic breast density is a well-established risk factor for breast
cancer. Recently there has been interest in breast MRI as an adjunct to
mammography, as this modality provides an orthogonal and highly quantitative
assessment of breast tissue. However, its 3D nature poses analytic challenges
related to delineating and aggregating complex structures across slices. Here,
we applied an in-house machine-learning algorithm to assess breast density on
normal breasts in three MRI datasets. Breast density was consistent across
different datasets (0.104 - 0.114). Analysis across different age groups also
demonstrated strong consistency across datasets and confirmed a trend of
decreasing density with age as reported in previous studies. MR breast density
was correlated with mammographic breast density, although some notable
differences suggest that certain breast density components are captured only on
MRI. Future work will determine how to integrate MR breast density with current
tools to improve future breast cancer risk prediction.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:01:51 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Yaqian",
""
],
[
"Li",
"Lin",
""
],
[
"Gu",
"Hanxue",
""
],
[
"Dong",
"Haoyu",
""
],
[
"Nguyen",
"Derek L.",
""
],
[
"Kirk",
"Allan D.",
""
],
[
"Mazurowski",
"Maciej A.",
""
],
[
"Hwang",
"E. Shelley",
""
]
] |
2504.15193 | Neelesh Kumar | Neelesh Kumar, Oya Aran | Automated Measurement of Eczema Severity with Self-Supervised Learning | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Automated diagnosis of eczema using images acquired from digital camera can
enable individuals to self-monitor their recovery. The process entails first
segmenting out the eczema region from the image and then measuring the severity
of eczema in the segmented region. The state-of-the-art methods for automated
eczema diagnosis rely on deep neural networks such as convolutional neural
network (CNN) and have shown impressive performance in accurately measuring the
severity of eczema. However, these methods require massive volume of annotated
data to train which can be hard to obtain. In this paper, we propose a
self-supervised learning framework for automated eczema diagnosis under limited
training data regime. Our framework consists of two stages: i) Segmentation,
where we use an in-context learning based algorithm called SegGPT for few-shot
segmentation of eczema region from the image; ii) Feature extraction and
classification, where we extract DINO features from the segmented regions and
feed it to a multi-layered perceptron (MLP) for 4-class classification of
eczema severity. When evaluated on a dataset of annotated "in-the-wild" eczema
images, we show that our method outperforms (Weighted F1: 0.67 $\pm$ 0.01) the
state-of-the-art deep learning methods such as finetuned Resnet-18 (Weighted
F1: 0.44 $\pm$ 0.16) and Vision Transformer (Weighted F1: 0.40 $\pm$ 0.22). Our
results show that self-supervised learning can be a viable solution for
automated skin diagnosis where labeled data is scarce.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:02:26 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kumar",
"Neelesh",
""
],
[
"Aran",
"Oya",
""
]
] |
2504.15194 | Lvzhou Li | Guanzhong Li, Lvzhou Li, and Jingquan Luo | Quantum phase discrimination with applications to quantum search on
graphs | null | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | We study the phase discrimination problem, in which we want to decide whether
the eigenphase $\theta\in(-\pi,\pi]$ of a given eigenstate $|\psi\rangle$ with
eigenvalue $e^{i\theta}$ is zero or not, using applications of the unitary $U$
provided as a black box oracle.We propose a quantum algorithm named {\it
quantum phase discrimination(QPD)} for this task, with optimal query complexity
$\Theta(\frac{1}{\lambda}\log\frac{1}{\delta})$ to the oracle $U$, where
$\lambda$ is the gap between zero and non-zero eigenphases and $\delta$ the
allowed one-sided error. The quantum circuit is simple, consisting of only one
ancillary qubit and a sequence of controlled-$U$ interleaved with single qubit
$Y$ rotations, whose angles are given by a simple analytical formula. Quantum
phase discrimination could become a fundamental subroutine in other quantum
algorithms, as we present two applications to quantum search on graphs:
i) Spatial search on graphs. Inspired by the structure of QPD, we propose a
new quantum walk model, and based on them we tackle the spatial search problem,
obtaining a novel quantum search algorithm. For any graph with any number of
marked vertices, the quantum algorithm that can find a marked vertex with
probability $\Omega(1)$ in total evolution time $ O(\frac{1}{\lambda
\sqrt{\varepsilon}})$ and query complexity $ O(\frac{1}{\sqrt{\varepsilon}})$,
where $\lambda$ is the gap between the zero and non-zero eigenvalues of the
graph Laplacian and $\varepsilon$ is a lower bound on the proportion of marked
vertices.
ii) Path-finding on graphs.} By using QPD, we reduce the query complexity of
a path-finding algorithm proposed by Li and Zur [arxiv: 2311.07372] from
$\tilde{O}(n^{11})$ to $\tilde{O}(n^8)$, in a welded-tree circuit graph with
$\Theta(n2^n)$ vertices.
Besides these two applications, we argue that more quantum algorithms might
benefit from QPD.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:06:43 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Li",
"Guanzhong",
""
],
[
"Li",
"Lvzhou",
""
],
[
"Luo",
"Jingquan",
""
]
] |
2504.15195 | Ruadha\'i Dervan | Ruadha\'i Dervan | Arc K-semistability is a very general property | 13 pages | null | null | null | math.AG math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove that arc K-semistability is a very general property in flat families
of polarised varieties, and prove a similar result for uniform arc K-stability.
This can be used to produce the only current examples of smooth uniformly arc
K-stable varieties which are not known to admit a constant scalar curvature
K\"ahler metric. Our technique is to prove a general result stating that
semistability of a pair in the sense of Paul is a Zariski open property, and to
employ prior work with Reboulet relating arc K-semistability to semistability
of an associated pair.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:07:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Dervan",
"Ruadhaí",
""
]
] |
2504.15196 | Diyako Ghaderyan | Diyako Ghaderyan, Stefan Werner | Fully Adaptive Stepsizes: Which System Benefit More -- Centralized or
Decentralized? | null | null | null | null | math.OC cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In decentralized optimization, the choice of stepsize plays a critical role
in algorithm performance. A common approach is to use a shared stepsize across
all agents to ensure convergence. However, selecting an optimal stepsize often
requires careful tuning, which can be time-consuming and may lead to slow
convergence, especially when there is significant variation in the smoothness
(L-smoothness) of local objective functions across agents. Individually tuning
stepsizes per agent is also impractical, particularly in large-scale networks.
To address these limitations, we propose AdGT, an adaptive gradient tracking
method that enables each agent to adjust its stepsize based on the smoothness
of its local objective. We prove that AdGT generates a sequence of iterates
that converges to the optimal consensus solution. Through numerical
experiments, we compare AdGT with fixed-stepsize gradient tracking methods and
demonstrate its superior performance. Additionally, we compare AdGT with
adaptive gradient descent (AdGD) in a centralized setting and observe that
fully adaptive stepsizes offer greater benefits in decentralized networks than
in centralized ones.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:07:32 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ghaderyan",
"Diyako",
""
],
[
"Werner",
"Stefan",
""
]
] |
2504.15197 | Luca Di Carlo | Luca Di Carlo, Francesca Mignacco, Christopher W. Lynn and William
Bialek | Extended mean-field theories for networks of real neurons | null | null | null | null | physics.bio-ph cond-mat.stat-mech q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | If the behavior of a system with many degrees of freedom can be captured by a
small number of collective variables, then plausibly there is an underlying
mean-field theory. We show that simple versions of this idea fail to describe
the patterns of activity in networks of real neurons. An extended mean-field
theory that matches the distribution of collective variables is at least
consistent, though shows signs that these networks are poised near a critical
point, in agreement with other observations. These results suggest a path to
analysis of emerging data on ever larger numbers of neurons.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:11:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Di Carlo",
"Luca",
""
],
[
"Mignacco",
"Francesca",
""
],
[
"Lynn",
"Christopher W.",
""
],
[
"Bialek",
"William",
""
]
] |
2504.15198 | Khandaker Akramul Haque | Khandaker Akramul Haque, Shining Sun, Xiang Huo, Ana E. Goulart and
Katherine R. Davis | Scalable Discrete Event Simulation Tool for Large-Scale Cyber-Physical
Energy Systems: Advancing System Efficiency and Scalability | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Modern power systems face growing risks from cyber-physical attacks,
necessitating enhanced resilience due to their societal function as critical
infrastructures. The challenge is that defense of large-scale
systems-of-systems requires scalability in their threat and risk assessment
environment for cyber physical analysis including cyber-informed transmission
planning, decision-making, and intrusion response. Hence, we present a scalable
discrete event simulation tool for analysis of energy systems, called DESTinE.
The tool is tailored for largescale cyber-physical systems, with a focus on
power systems. It supports faster-than-real-time traffic generation and models
packet flow and congestion under both normal and adversarial conditions. Using
three well-established power system synthetic cases with 500, 2000, and 10,000
buses, we overlay a constructed cyber network employing star and radial
topologies. Experiments are conducted to identify critical nodes within a
communication network in response to a disturbance. The findings are
incorporated into a constrained optimization problem to assess the impact of
the disturbance on a specific node and its cascading effects on the overall
network. Based on the solution of the optimization problem, a new hybrid
network topology is also derived, combining the strengths of star and radial
structures to improve network resilience. Furthermore, DESTinE is integrated
with a virtual server and a hardware-in-the-loop (HIL) system using Raspberry
Pi 5.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:14:45 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Haque",
"Khandaker Akramul",
""
],
[
"Sun",
"Shining",
""
],
[
"Huo",
"Xiang",
""
],
[
"Goulart",
"Ana E.",
""
],
[
"Davis",
"Katherine R.",
""
]
] |
2504.15199 | Yassir Benhammou | Yassir Benhammou, Alessandro Tiberio, Gabriel Trautmann, Suman Kalyan | Zero-Shot, But at What Cost? Unveiling the Hidden Overhead of MILS's
LLM-CLIP Framework for Image Captioning | 9 pages, 2 tables, 1 figure | null | null | null | cs.CV cs.AI cs.LG cs.PF | http://creativecommons.org/licenses/by/4.0/ | MILS (Multimodal Iterative LLM Solver) is a recently published framework that
claims "LLMs can see and hear without any training" by leveraging an iterative,
LLM-CLIP based approach for zero-shot image captioning. While this MILS
approach demonstrates good performance, our investigation reveals that this
success comes at a hidden, substantial computational cost due to its expensive
multi-step refinement process. In contrast, alternative models such as BLIP-2
and GPT-4V achieve competitive results through a streamlined, single-pass
approach. We hypothesize that the significant overhead inherent in MILS's
iterative process may undermine its practical benefits, thereby challenging the
narrative that zero-shot performance can be attained without incurring heavy
resource demands. This work is the first to expose and quantify the trade-offs
between output quality and computational cost in MILS, providing critical
insights for the design of more efficient multimodal models.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:16:19 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Benhammou",
"Yassir",
""
],
[
"Tiberio",
"Alessandro",
""
],
[
"Trautmann",
"Gabriel",
""
],
[
"Kalyan",
"Suman",
""
]
] |
2504.15200 | Ramakrishna Nanduri Dr | Ramakrishna Nanduri and Tapas Kumar Roy | On robust toric ideals of weighted oriented graphs | Comments are welcome ! | null | null | null | math.AC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this work, we study the equivalence of robustness, strongly robustness,
generalized robustness, and weakly robustness properties of toric ideals of
weighted oriented graphs. For any weighted oriented graph $D$, if its toric
ideal $I_D$ is generalized robust or weakly robust, then we show that $D$ has
no subgraphs of certain structures.
We prove the equality of Graver basis, Universal Gr\"obner basis, reduced
Gr\"obner basis with respect to degree lexicographic order of the toric ideals
$I_D$ of weighted oriented graphs $D=\C_1\cup_P \cdots \cup_P \C_n$ consist of
cycles $\C_1\ldots,\C_n$ that share a path $P$. As a consequence, we show that
(i) $I_{D}$ is robust iff $I_{D}$ is strongly robust; (ii) $I_{D}$ is
generalized robust iff $I_{D}$ is weakly robust. If at most two of the cycles
$\C_i$ in $D$ are unbalanced, then the following statements are equivalent: (i)
$I_{D}$ is strongly robust; (ii) $I_{D}$ is robust; (iii) $I_{D}$ is
generalized robust; (iv) $I_{D}$ is weakly robust; (v) $D$ has no subgraphs of
types $D_{1}$ and $D_{2}$, where $D_1$ is a weighted oriented graph consisting
of two balanced cycles share an edge in $D$, and $D_2$ is a weighted oriented
graph consisting of three cycles that share an edge\ such that one cycle is
balanced and the rest two are unbalanced cycles, as in figure \ref{fig2}. We
explicitly determine the Graver basis of the toric ideal of two balanced cycles
sharing a path.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:17:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Nanduri",
"Ramakrishna",
""
],
[
"Roy",
"Tapas Kumar",
""
]
] |
2504.15201 | Annalisa Quaini | Maxim Olshanskii and Annalisa Quaini | Phase-separated lipid vesicles: continuum modeling, simulation, and
validation | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | The paper presents a complete research cycle comprising continuum-based
modeling, computational framework development, and validation setup to predict
phase separation and surface hydrodynamics in lipid bilayer membranes. We
starting with an overview of the key physical characteristics of lipid
bilayers, including their composition, mechanical properties, and
thermodynamics, and then discuss continuum models of multi-component bilayers.
The most complex model is a Navier--Stokes--Cahn--Hilliard (NSCH) type system,
describing the coupling of incompressible surface fluid dynamics with
phase-field dynamics on arbitrarily curved geometries. It is discretized using
trace finite element methods, which offer geometric flexibility and stability
in representing surface PDEs. Numerical studies are conducted to examine
physical features such as coarsening rates and interfacial dynamics. The
computational results obtained from the NSCH model are compared against
experimental data for membrane compositions with distinct phase behaviors,
demonstrating that including both phase-field models and surface hydrodynamics
is essential to accurately reproduce domain evolution observed in
epi-fluorescence microscopy. Lastly, we extend the model to incorporate
external forces that enable the simulation of vesicles containing cationic
lipids, used to enhance membrane fusion.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:17:39 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Olshanskii",
"Maxim",
""
],
[
"Quaini",
"Annalisa",
""
]
] |
2504.15202 | Therar Kadri | Jana Hamza, Mohammad EL Hindi, Seifeddine Kadri, Therrar Kadri, Yahya
Awad | Extending the ElGamal Cryptosystem to the Third Group of Units of
$\Z_{n}$ | 11 pages, 1 figure | null | null | null | cs.CR cs.IT math.GR math.IT math.PR math.RA | http://creativecommons.org/publicdomain/zero/1.0/ | In this paper, we extend the ElGamal cryptosystem to the third group of units
of the ring $\Z_{n}$, which we prove to be more secure than the previous
extensions. We describe the arithmetic needed in the new setting. We also
provide some numerical simulations that shows the security and efficiency of
our proposed cryptosystem.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:17:53 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Hamza",
"Jana",
""
],
[
"Hindi",
"Mohammad EL",
""
],
[
"Kadri",
"Seifeddine",
""
],
[
"Kadri",
"Therrar",
""
],
[
"Awad",
"Yahya",
""
]
] |
2504.15204 | Tim Janz | Tim Janz, Simon Oberm\"uller, Andreas Zunker, Stephan ten Brink | Soft-Output from Covered Space Decoding of Product Codes | 6 pages, 2 figures, 2 tables | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we propose a new soft-in soft-out decoder called soft-output
from covered space (SOCS) decoder. It estimates the a posteriori reliability
based on the space explored by a list decoder, i.e., the set of vectors for
which the list decoder knows whether they are codewords. This approach enables
a more accurate calculation of the a posteriori reliability and results in
gains of up to 0.25$\,$dB for turbo product decoding with SOCS decoding
compared to Chase-Pyndiah decoding.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:20:13 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Janz",
"Tim",
""
],
[
"Obermüller",
"Simon",
""
],
[
"Zunker",
"Andreas",
""
],
[
"Brink",
"Stephan ten",
""
]
] |
2504.15205 | Nandan Thakur | Nandan Thakur, Ronak Pradeep, Shivani Upadhyay, Daniel Campos, Nick
Craswell, Jimmy Lin | Support Evaluation for the TREC 2024 RAG Track: Comparing Human versus
LLM Judges | Accepted at SIGIR 2025 (short) | null | null | null | cs.CL cs.AI cs.IR | http://creativecommons.org/licenses/by-sa/4.0/ | Retrieval-augmented generation (RAG) enables large language models (LLMs) to
generate answers with citations from source documents containing "ground
truth", thereby reducing system hallucinations. A crucial factor in RAG
evaluation is "support", whether the information in the cited documents
supports the answer. To this end, we conducted a large-scale comparative study
of 45 participant submissions on 36 topics to the TREC 2024 RAG Track,
comparing an automatic LLM judge (GPT-4o) against human judges for support
assessment. We considered two conditions: (1) fully manual assessments from
scratch and (2) manual assessments with post-editing of LLM predictions. Our
results indicate that for 56% of the manual from-scratch assessments, human and
GPT-4o predictions match perfectly (on a three-level scale), increasing to 72%
in the manual with post-editing condition. Furthermore, by carefully analyzing
the disagreements in an unbiased study, we found that an independent human
judge correlates better with GPT-4o than a human judge, suggesting that LLM
judges can be a reliable alternative for support assessment. To conclude, we
provide a qualitative analysis of human and GPT-4o errors to help guide future
iterations of support assessment.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:20:43 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Thakur",
"Nandan",
""
],
[
"Pradeep",
"Ronak",
""
],
[
"Upadhyay",
"Shivani",
""
],
[
"Campos",
"Daniel",
""
],
[
"Craswell",
"Nick",
""
],
[
"Lin",
"Jimmy",
""
]
] |
2504.15206 | S\'ilvia Casacuberta | S\'ilvia Casacuberta, Parikshit Gopalan, Varun Kanade, Omer Reingold | How Global Calibration Strengthens Multiaccuracy | null | null | null | null | cs.LG cs.CC | http://creativecommons.org/licenses/by/4.0/ | Multiaccuracy and multicalibration are multigroup fairness notions for
prediction that have found numerous applications in learning and computational
complexity. They can be achieved from a single learning primitive: weak
agnostic learning. Here we investigate the power of multiaccuracy as a learning
primitive, both with and without the additional assumption of calibration. We
find that multiaccuracy in itself is rather weak, but that the addition of
global calibration (this notion is called calibrated multiaccuracy) boosts its
power substantially, enough to recover implications that were previously known
only assuming the stronger notion of multicalibration.
We give evidence that multiaccuracy might not be as powerful as standard weak
agnostic learning, by showing that there is no way to post-process a
multiaccurate predictor to get a weak learner, even assuming the best
hypothesis has correlation $1/2$. Rather, we show that it yields a restricted
form of weak agnostic learning, which requires some concept in the class to
have correlation greater than $1/2$ with the labels. However, by also requiring
the predictor to be calibrated, we recover not just weak, but strong agnostic
learning.
A similar picture emerges when we consider the derivation of hardcore
measures from predictors satisfying multigroup fairness notions. On the one
hand, while multiaccuracy only yields hardcore measures of density half the
optimal, we show that (a weighted version of) calibrated multiaccuracy achieves
optimal density.
Our results yield new insights into the complementary roles played by
multiaccuracy and calibration in each setting. They shed light on why
multiaccuracy and global calibration, although not particularly powerful by
themselves, together yield considerably stronger notions.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:22:44 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Casacuberta",
"Sílvia",
""
],
[
"Gopalan",
"Parikshit",
""
],
[
"Kanade",
"Varun",
""
],
[
"Reingold",
"Omer",
""
]
] |
2504.15207 | Dylan Cant | Filip Bro\'ci\'c and Dylan Cant | Parametric Gromov width of Liouville domains | 97 pages | null | null | null | math.SG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The classical Gromov width measures the largest symplectic ball embeddable
into a symplectic manifold; inspired by the symplectic camel problem, we
generalize this to ask how large a symplectic ball can be embedded as a family
over a parameter space $N$. Given a smooth map $f: N \to \Omega$, where
$\Omega$ is a symplectic manifold, we define the parametric Gromov width
$\mathrm{Gr}(f,\Omega)$ as the supremum of capacities $a$ for which there
exists $F: N \times B(a) \to \Omega$ with $F(\eta, 0) = f(\eta)$ and which
restricts to a symplectic embedding on each ball $\{ \eta \} \times B(a)$,
where $B(a) \subset \mathbb{C}^n$ is the closed ball of capacity $a$. For
Liouville domains $\Omega$, we establish upper bounds on
$\mathrm{Gr}(f,\Omega)$ using the Floer cohomology persistence module
associated to $\Omega$. Specializing to fiberwise starshaped domains in the
cotangent bundle $T^*M$, we derive computable bounds via filtered string
topology. Specific examples of $\Omega$ -- including disk cotangent bundles of
thin ellipsoids, open books, and tori -- demonstrate our bounds, and reveal
constraints on parameterized symplectic embeddings beyond the classical Gromov
width.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:26:10 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Broćić",
"Filip",
""
],
[
"Cant",
"Dylan",
""
]
] |
2504.15208 | Marc Finzi | Marc Finzi and Sanyam Kapoor and Diego Granziol and Anming Gu and
Christopher De Sa and J. Zico Kolter and Andrew Gordon Wilson | Compute-Optimal LLMs Provably Generalize Better With Scale | ICLR 2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Why do larger language models generalize better? To investigate this
question, we develop generalization bounds on the pretraining objective of
large language models (LLMs) in the compute-optimal regime, as described by the
Chinchilla scaling laws. We introduce a novel, fully empirical Freedman-type
martingale concentration inequality that tightens existing bounds by accounting
for the variance of the loss function. This generalization bound can be
decomposed into three interpretable components: the number of parameters per
token, the loss variance, and the quantization error at a fixed bitrate. As
compute-optimal language models are scaled up, the number of parameters per
data point remains constant; however, both the loss variance and the
quantization error decrease, implying that larger models should have smaller
generalization gaps. We examine why larger models tend to be more quantizable
from an information theoretic perspective, showing that the rate at which they
can integrate new information grows more slowly than their capacity on the
compute-optimal frontier. From these findings we produce a scaling law for the
generalization gap, with bounds that become predictably stronger with scale.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:26:56 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Finzi",
"Marc",
""
],
[
"Kapoor",
"Sanyam",
""
],
[
"Granziol",
"Diego",
""
],
[
"Gu",
"Anming",
""
],
[
"De Sa",
"Christopher",
""
],
[
"Kolter",
"J. Zico",
""
],
[
"Wilson",
"Andrew Gordon",
""
]
] |
2504.15209 | Xin Liao | Xin Liao, Bing Yang, Tan Dongli and Cai Yu | A Causal Convolutional Low-rank Representation Model for Imputation of
Water Quality Data | 9 pages, 3 figures | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | The monitoring of water quality is a crucial part of environmental
protection, and a large number of monitors are widely deployed to monitor water
quality. Due to unavoidable factors such as data acquisition breakdowns,
sensors and communication failures, water quality monitoring data suffers from
missing values over time, resulting in High-Dimensional and Sparse (HDS) Water
Quality Data (WQD). The simple and rough filling of the missing values leads to
inaccurate results and affects the implementation of relevant measures.
Therefore, this paper proposes a Causal convolutional Low-rank Representation
(CLR) model for imputing missing WQD to improve the completeness of the WQD,
which employs a two-fold idea: a) applying causal convolutional operation to
consider the temporal dependence of the low-rank representation, thus
incorporating temporal information to improve the imputation accuracy; and b)
implementing a hyperparameters adaptation scheme to automatically adjust the
best hyperparameters during model training, thereby reducing the tedious manual
adjustment of hyper-parameters. Experimental studies on three real-world water
quality datasets demonstrate that the proposed CLR model is superior to some of
the existing state-of-the-art imputation models in terms of imputation accuracy
and time cost, as well as indicating that the proposed model provides more
reliable decision support for environmental monitoring.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:27:16 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Liao",
"Xin",
""
],
[
"Yang",
"Bing",
""
],
[
"Dongli",
"Tan",
""
],
[
"Yu",
"Cai",
""
]
] |
2504.15210 | Marina Sakharova | Marina Sakharova and Abhinav Anand and Mira Mezini | Integrating Symbolic Execution into the Fine-Tuning of Code-Generating
LLMs | null | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | Code-generating Large Language Models (LLMs) have become essential tools in
modern software development, enhancing productivity and accelerating
development. This paper aims to investigate the fine-tuning of code-generating
LLMs using Reinforcement Learning and Direct Preference Optimization, further
improving their performance. To achieve this, we enhance the training data for
the reward model with the help of symbolic execution techniques, ensuring more
comprehensive and objective data. With symbolic execution, we create a custom
dataset that better captures the nuances in code evaluation. Our reward models,
fine-tuned on this dataset, demonstrate significant improvements over the
baseline, CodeRL, in estimating the quality of generated code. Our
code-generating LLMs, trained with the help of reward model feedback, achieve
similar results compared to the CodeRL benchmark.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:29:07 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sakharova",
"Marina",
""
],
[
"Anand",
"Abhinav",
""
],
[
"Mezini",
"Mira",
""
]
] |
2504.15211 | Yanan Long | Yanan Long | Position: Bayesian Statistics Facilitates Stakeholder Participation in
Evaluation of Generative AI | To be presented at ACM CHI 2025 workshop STAIG | null | null | null | cs.AI stat.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The evaluation of Generative AI (GenAI) systems plays a critical role in
public policy and decision-making, yet existing methods are often limited by
reliance on benchmark-driven, point-estimate comparisons that fail to capture
uncertainty and broader societal impacts. This paper argues for the use of
Bayesian statistics as a principled framework to address these challenges.
Bayesian methods enable the integration of domain expertise through prior
elicitation, allow for continuous learning from new data, and provide robust
uncertainty quantification via posterior inference. We demonstrate how Bayesian
inference can be applied to GenAI evaluation, particularly in incorporating
stakeholder perspectives to enhance fairness, transparency, and reliability.
Furthermore, we discuss Bayesian workflows as an iterative process for model
validation and refinement, ensuring robust assessments of GenAI systems in
dynamic, real-world contexts.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:31:15 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Long",
"Yanan",
""
]
] |
2504.15212 | Pandelis Dodos | Dylan J. Altschuler, Pandelis Dodos, Konstantin Tikhomirov and
Konstantinos Tyros | A universal threshold for geometric embeddings of trees | null | null | null | null | math.CO math.FA math.MG math.PR | http://creativecommons.org/licenses/by/4.0/ | A graph $G=(V,E)$ is geometrically embeddable into a normed space $X$ when
there is a mapping $\zeta: V\to X$ such that $\|\zeta(v)-\zeta(w)\|_X\leqslant
1$ if and only if $\{v,w\}\in E$, for all distinct $v,w\in V$. Our result is
the following universal threshold for the embeddability of trees. Let $\Delta
\geqslant 3$, and let $N$ be sufficiently large in terms of $\Delta$. Every
$N$--vertex tree of maximal degree at most $\Delta$ is embeddable into any
normed space of dimension at least $64\,\frac{\log N}{\log\log N}$, and
complete trees are non-embeddable into any normed space of dimension less than
$\frac{1}{2}\,\frac{\log N}{\log\log N}$. In striking contrast, spectral
expanders and random graphs are known to be non-embeddable in sublogarithmic
dimension. Our result is based on a randomized embedding whose analysis
utilizes the recent breakthroughs on Bourgain's slicing problem.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:33:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Altschuler",
"Dylan J.",
""
],
[
"Dodos",
"Pandelis",
""
],
[
"Tikhomirov",
"Konstantin",
""
],
[
"Tyros",
"Konstantinos",
""
]
] |
2504.15213 | Bianka Kov\'acs | Bianka Kov\'acs, Barnab\'as Benedek, Gergely Palla | Community detection in hypergraphs through hyperedge percolation | 19 pages, 8 figures | null | null | null | physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Complex networks, representing the connections between the constituents of
large complex systems, are often characterised by a community structure, with
communities corresponding to denser sub-graphs in which nodes are closely
linked. When modelling systems where interactions extend beyond node pairs to
involve arbitrary numbers of nodes, hypergraphs become necessary, creating a
need for specialised community detection methods. Here, we adapt the classical
$k$-clique percolation method to hypergraphs, building communities from
hyperedges containing at least $k$ nodes, defining hyperedge adjacency
similarly to clique adjacency in the original algorithm. Although the analogy
between the proposed hyperedge percolation method and the classical clique
percolation algorithm is evident, we show that communities obtained directly
from the hyperedges can differ from those identified by the clique percolation
method in the pairwise projection of the hypergraph. We also propose an
alternative way for uniting hyperedges into communities, where instead of
limiting the cardinality of the considered hyperedges from below, we restrict
the size of the largest hyperedges that can be added to a community. This
alternative algorithm is better suited to hypergraphs where larger edges
realise weaker linkages between the nodes. After comparing the suggested two
approaches on simple synthetic hypergraphs constructed to highlight their
differences, we test them on hypergraphs generated with a newly proposed
geometric process on the hyperbolic plane, as well as on some real-world
examples.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:35:24 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kovács",
"Bianka",
""
],
[
"Benedek",
"Barnabás",
""
],
[
"Palla",
"Gergely",
""
]
] |
2504.15215 | Brendan Cross | Boleslaw K. Szymanski, Brendan Cross, John Hulton, James Flamino,
Chris Gaiteri, Jonathan Z. Bakdash | An experimental study of the influence of anonymous information on
social media users | null | null | null | null | cs.SI | http://creativecommons.org/licenses/by/4.0/ | Increasingly, people use social media for their day-to-day interactions and
as a source of information, even though much of this information is practically
anonymous. This raises the question: does anonymous information influence its
recipients? We conducted an online, two-phase, preregistered experiment using a
nationally representative sample of participants from the U.S. to find the
answer. To avoid biases of opinions among participants, in the first phase,
each participant examines ten Rorschach inkblots and chooses one of four
opinions assigned to each inkblot. In the second phase, the participants are
randomly assigned to one of four distinct information conditions and are asked
to revisit their opinions for the same ten inkblots. Conditions ranged from
repeating phase one to receiving anonymous comments about certain opinions.
Results were consistent with the preregistration. Importantly, anonymous
comments shown in phase two influence up to half of the participants' opinion
selections. To better understand the role of anonymous comments in influencing
the selections of opinions, we implemented agent-based modeling (ABM). ABM
results suggest that a straightforward mechanism can explain the impact of such
information. Overall, our results indicate that even anonymous information can
have a significant impact on its recipients, potentially altering their
popularity rankings. However, the strength of such influence weakens when
recipients' confidence in their selections increases. Additionally, we found
that participants' confidence in the first phase is inversely related to the
number of change opinions.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:39:17 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Szymanski",
"Boleslaw K.",
""
],
[
"Cross",
"Brendan",
""
],
[
"Hulton",
"John",
""
],
[
"Flamino",
"James",
""
],
[
"Gaiteri",
"Chris",
""
],
[
"Bakdash",
"Jonathan Z.",
""
]
] |
2504.15217 | Nicholas J. Bryan | Yatong Bai and Jonah Casebeer and Somayeh Sojoudi and Nicholas J.
Bryan | DRAGON: Distributional Rewards Optimize Diffusion Generative Models | null | null | null | null | cs.SD cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present Distributional RewArds for Generative OptimizatioN (DRAGON), a
versatile framework for fine-tuning media generation models towards a desired
outcome. Compared with traditional reinforcement learning with human feedback
(RLHF) or pairwise preference approaches such as direct preference optimization
(DPO), DRAGON is more flexible. It can optimize reward functions that evaluate
either individual examples or distributions of them, making it compatible with
a broad spectrum of instance-wise, instance-to-distribution, and
distribution-to-distribution rewards. Leveraging this versatility, we construct
novel reward functions by selecting an encoder and a set of reference examples
to create an exemplar distribution. When cross-modality encoders such as CLAP
are used, the reference examples may be of a different modality (e.g., text
versus audio). Then, DRAGON gathers online and on-policy generations, scores
them to construct a positive demonstration set and a negative set, and
leverages the contrast between the two sets to maximize the reward. For
evaluation, we fine-tune an audio-domain text-to-music diffusion model with 20
different reward functions, including a custom music aesthetics model, CLAP
score, Vendi diversity, and Frechet audio distance (FAD). We further compare
instance-wise (per-song) and full-dataset FAD settings while ablating multiple
FAD encoders and reference sets. Over all 20 target rewards, DRAGON achieves an
81.45% average win rate. Moreover, reward functions based on exemplar sets
indeed enhance generations and are comparable to model-based rewards. With an
appropriate exemplar set, DRAGON achieves a 60.95% human-voted music quality
win rate without training on human preference annotations. As such, DRAGON
exhibits a new approach to designing and optimizing reward functions for
improving human-perceived quality. Sound examples at
https://ml-dragon.github.io/web.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:41:40 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bai",
"Yatong",
""
],
[
"Casebeer",
"Jonah",
""
],
[
"Sojoudi",
"Somayeh",
""
],
[
"Bryan",
"Nicholas J.",
""
]
] |
2504.15218 | Qing Gao | Qing Gao, Yungui Gong, Zhu Yi and Fengge Zhang | Non-minimal coupling in light of ACT | 5 pages, 2 figures | null | null | null | astro-ph.CO | http://creativecommons.org/publicdomain/zero/1.0/ | The latest ACT data release disfavors the attractor $n_s=1-2/N$. In
inflationary models with nonminimal coupling, such attractors typically arise
in the strong coupling limit. To align with observational constraints, we focus
on nonminimal coupling models with small coupling constants. For the model with
the coupling function $\Omega(\phi) = 1 + \xi f(\phi)$ and the potential
$V(\phi) = \lambda^2 f^2(\phi)$, we find that observational data constrain the
parameters as $0.1 \lesssim \xi \lesssim 35$ and $0 \lesssim k \lesssim 1.5$
for $f(\phi) = \phi^k$ at the $1\sigma$ confidence level. With the help of the
nonmiminal coupling $\Omega(\phi) = 1 + \xi \phi^2$, the hilltop inflation and
power-law inflation models with power indices $2/3$ and $1/3$ can be consistent
with observational data within the $1\sigma$ range. We also give the viable
parameter regions for $\xi$ for these three models.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:43:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Gao",
"Qing",
""
],
[
"Gong",
"Yungui",
""
],
[
"Yi",
"Zhu",
""
],
[
"Zhang",
"Fengge",
""
]
] |
2504.15219 | Manya Wadhwa | Manya Wadhwa, Zayne Sprague, Chaitanya Malaviya, Philippe Laban, Junyi
Jessy Li, Greg Durrett | EvalAgent: Discovering Implicit Evaluation Criteria from the Web | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Evaluation of language model outputs on structured writing tasks is typically
conducted with a number of desirable criteria presented to human evaluators or
large language models (LLMs). For instance, on a prompt like "Help me draft an
academic talk on coffee intake vs research productivity", a model response may
be evaluated for criteria like accuracy and coherence. However, high-quality
responses should do more than just satisfy basic task requirements. An
effective response to this query should include quintessential features of an
academic talk, such as a compelling opening, clear research questions, and a
takeaway. To help identify these implicit criteria, we introduce EvalAgent, a
novel framework designed to automatically uncover nuanced and task-specific
criteria. EvalAgent first mines expert-authored online guidance. It then uses
this evidence to propose diverse, long-tail evaluation criteria that are
grounded in reliable external sources. Our experiments demonstrate that the
grounded criteria produced by EvalAgent are often implicit (not directly stated
in the user's prompt), yet specific (high degree of lexical precision).
Further, EvalAgent criteria are often not satisfied by initial responses but
they are actionable, such that responses can be refined to satisfy them.
Finally, we show that combining LLM-generated and EvalAgent criteria uncovers
more human-valued criteria than using LLMs alone.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:43:50 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wadhwa",
"Manya",
""
],
[
"Sprague",
"Zayne",
""
],
[
"Malaviya",
"Chaitanya",
""
],
[
"Laban",
"Philippe",
""
],
[
"Li",
"Junyi Jessy",
""
],
[
"Durrett",
"Greg",
""
]
] |
2504.15220 | Juli\'an Cendrero | Juli\'an Cendrero, Julio Gonzalo, Ivar Zapata | Fully Bayesian Approaches to Topics over Time | 25 pages | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Topics over Time (ToT) model captures thematic changes in timestamped
datasets by explicitly modeling publication dates jointly with word
co-occurrence patterns. However, ToT was not approached in a fully Bayesian
fashion, a flaw that makes it susceptible to stability problems. To address
this issue, we propose a fully Bayesian Topics over Time (BToT) model via the
introduction of a conjugate prior to the Beta distribution. This prior acts as
a regularization that prevents the online version of the algorithm from
unstable updates when a topic is poorly represented in a mini-batch. The
characteristics of this prior to the Beta distribution are studied here for the
first time. Still, this model suffers from a difference in scale between the
single-time observations and the multiplicity of words per document. A
variation of BToT, Weighted Bayesian Topics over Time (WBToT), is proposed as a
solution. In WBToT, publication dates are repeated a certain number of times
per document, which balances the relative influence of words and timestamps
along the inference process. We have tested our models on two datasets: a
collection of over 200 years of US state-of-the-union (SOTU) addresses and a
large-scale COVID-19 Twitter corpus of 10 million tweets. The results show that
WBToT captures events better than Latent Dirichlet Allocation and other SOTA
topic models like BERTopic: the median absolute deviation of the topic presence
over time is reduced by $51\%$ and $34\%$, respectively. Our experiments also
demonstrate the superior coherence of WBToT over BToT, which highlights the
importance of balancing the time and word modalities. Finally, we illustrate
the stability of the online optimization algorithm in WBToT, which allows the
application of WBToT to problems that are intractable for standard ToT.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:46:07 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cendrero",
"Julián",
""
],
[
"Gonzalo",
"Julio",
""
],
[
"Zapata",
"Ivar",
""
]
] |
2504.15221 | Adam Chudecki Dr | Adam Chudecki | On Walker and para-Hermite Einstein spaces | null | null | null | null | math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A special class of (complex) para-Hermite Einstein spaces is analyzed. For
this class of spaces the self-dual Weyl tensor is type-[D] in the
Petrov-Penrose classification. The anti-self-dual Weyl tensor is algebraically
degenerate, equivalently, there exists an anti-self-dual congruence of null
strings. It is assumed that this congruence is parallely propagated. Thus, the
spaces are not only para-Hermite but also Walker. A classification of the
spaces according to three criteria is given. Finally, explicit metrics of all
admitted Petrov-Penrose types are found.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:47:54 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chudecki",
"Adam",
""
]
] |
2504.15222 | Deng Wang | Deng Wang, David Mota | Did DESI DR2 truly reveal dynamical dark energy? | 9 pages, 5 figures | null | null | null | astro-ph.CO astro-ph.HE gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A fundamental question in cosmology is whether dark energy evolves over time,
a topic that has gained prominence since the discovery of cosmic acceleration.
Recently, the DESI collaboration has reported increasing evidence for evolving
dark energy using combinations of cosmic microwave background (CMB), type Ia
supernova (SN), and their new measurements of baryon acoustic oscillations
(BAO). However, our analysis reveals that these combinations are problematic
due to clear tensions among the CMB, BAO and SN datasets. Consequently, DESI's
claim of dynamical dark energy (DDE) is not robust. A more reliable approach
involves constraining the evolution of dark energy using each dataset
independently. Through a statistical comparison for each dataset, on average,
we find that DDE is strongly preferred over the $\Lambda$CDM model. This
suggests that DDE likely exists, although its real parameter space remains
elusive due to weak constraints on the dark energy equation of state and
inconsistencies among the datasets. Interestingly, when considering DDE, none
of the individual datasets -- including CMB, DESI DR2, Pantheon+, Union3, and
DESY5 -- can independently detect cosmic acceleration at a significant level.
Our findings not only clarify the current understanding of the nature of dark
energy but also challenge the established discovery of cosmic acceleration and
the long-held notion that dark energy exerts negative pressure. Both individual
and combined datasets suggest that the ultimate fate of the universe is likely
to be dominated by matter rather than dark energy.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:50:54 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wang",
"Deng",
""
],
[
"Mota",
"David",
""
]
] |
2504.15223 | Tao Yang | Tao Yang, Yu Cheng, Yaokun Ren, Yujia Lou, Minggu Wei, Honghui Xin | A Deep Learning Framework for Sequence Mining with Bidirectional LSTM
and Multi-Scale Attention | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the challenges of mining latent patterns and modeling
contextual dependencies in complex sequence data. A sequence pattern mining
algorithm is proposed by integrating Bidirectional Long Short-Term Memory
(BiLSTM) with a multi-scale attention mechanism. The BiLSTM captures both
forward and backward dependencies in sequences, enhancing the model's ability
to perceive global contextual structures. At the same time, the multi-scale
attention module assigns adaptive weights to key feature regions under
different window sizes. This improves the model's responsiveness to both local
and global important information. Extensive experiments are conducted on a
publicly available multivariate time series dataset. The proposed model is
compared with several mainstream sequence modeling methods. Results show that
it outperforms existing models in terms of accuracy, precision, and recall.
This confirms the effectiveness and robustness of the proposed architecture in
complex pattern recognition tasks. Further ablation studies and sensitivity
analyses are carried out to investigate the effects of attention scale and
input sequence length on model performance. These results provide empirical
support for structural optimization of the model.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:53:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yang",
"Tao",
""
],
[
"Cheng",
"Yu",
""
],
[
"Ren",
"Yaokun",
""
],
[
"Lou",
"Yujia",
""
],
[
"Wei",
"Minggu",
""
],
[
"Xin",
"Honghui",
""
]
] |
2504.15224 | Victor Daniel Mendoza Rubio | Rafael Holanda, Victor H. Jorge-P\'erez, Victor D. Mendoza-Rubio | On Extension modules of finite homological dimension | 22 pages. Comments and suggestions are welcome | null | null | null | math.AC | http://creativecommons.org/licenses/by/4.0/ | We explore the implications of the finiteness of homological dimensions for
Ext modules, focusing on projective dimension, injective dimension, and their
Gorenstein counterpart. In this direction, we establish several finiteness
criteria for homological dimensions. Our results include the consequences of
the finiteness of the Gorenstein (injective) dimension of the deficiency
modules as well as a duality for certain Ext modules, all of finite
(Gorenstein) projective dimension.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:54:57 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Holanda",
"Rafael",
""
],
[
"Jorge-Pérez",
"Victor H.",
""
],
[
"Mendoza-Rubio",
"Victor D.",
""
]
] |
2504.15225 | Sarah Alnegheimish | Sarah Alnegheimish, Zelin He, Matthew Reimherr, Akash Chandrayan,
Abhinav Pradhan, Luca D'Angelo | M$^2$AD: Multi-Sensor Multi-System Anomaly Detection through Global
Scoring and Calibrated Thresholding | Accepted at AISTATS 2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | With the widespread availability of sensor data across industrial and
operational systems, we frequently encounter heterogeneous time series from
multiple systems. Anomaly detection is crucial for such systems to facilitate
predictive maintenance. However, most existing anomaly detection methods are
designed for either univariate or single-system multivariate data, making them
insufficient for these complex scenarios. To address this, we introduce
M$^2$AD, a framework for unsupervised anomaly detection in multivariate time
series data from multiple systems. M$^2$AD employs deep models to capture
expected behavior under normal conditions, using the residuals as indicators of
potential anomalies. These residuals are then aggregated into a global anomaly
score through a Gaussian Mixture Model and Gamma calibration. We theoretically
demonstrate that this framework can effectively address heterogeneity and
dependencies across sensors and systems. Empirically, M$^2$AD outperforms
existing methods in extensive evaluations by 21% on average, and its
effectiveness is demonstrated on a large-scale real-world case study on 130
assets in Amazon Fulfillment Centers. Our code and results are available at
https://github.com/sarahmish/M2AD.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:57:46 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Alnegheimish",
"Sarah",
""
],
[
"He",
"Zelin",
""
],
[
"Reimherr",
"Matthew",
""
],
[
"Chandrayan",
"Akash",
""
],
[
"Pradhan",
"Abhinav",
""
],
[
"D'Angelo",
"Luca",
""
]
] |
2504.15226 | Nathan Steffen | Nathan Steffen, Wilhelm Louw, Nicholas Ernest, Timothy Arnett, and
Kelly Cohen | A Genetic Fuzzy-Enabled Framework on Robotic Manipulation for In-Space
Servicing | null | null | null | null | cs.RO cs.AI cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automation of robotic systems for servicing in cislunar space is becoming
extremely important as the number of satellites in orbit increases. Safety is
critical in performing satellite maintenance, so the control techniques
utilized must be trusted in addition to being highly efficient. In this work,
Genetic Fuzzy Trees are combined with the widely used LQR control scheme via
Thales' TrUE AI Toolkit to create a trusted and efficient controller for a
two-degree-of-freedom planar robotic manipulator that would theoretically be
used to perform satellite maintenance. It was found that Genetic Fuzzy-LQR is
18.5% more performant than optimal LQR on average, and that it is incredibly
robust to uncertainty.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:57:56 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Steffen",
"Nathan",
""
],
[
"Louw",
"Wilhelm",
""
],
[
"Ernest",
"Nicholas",
""
],
[
"Arnett",
"Timothy",
""
],
[
"Cohen",
"Kelly",
""
]
] |
2504.15227 | Karel Kn\'i\v{z}ek | Jan Zich, Tom\'a\v{s} Plech\'a\v{c}ek, Anton\'in Sojka, Petr
Levinsk\'y, Ji\v{r}\'i Navr\'atil, Pavl\'ina Ruleov\'a, Stanislav \v{S}lang,
Karel Kn\'i\v{z}ek, Ji\v{r}\'i Hejtm\'anek, Vojt\v{e}ch Ne\v{c}ina, and
\v{C}estm\'ir Dra\v{s}ar | The preparation and properties of polycrystalline Bi$_2$O$_2$Se --
pitfalls and difficulties with reproducibility and charge transport limiting
parameters | Manuscript: 20 pages, 9 figures. Electronic Supplementary Material:
21 pages, 16 figures, 11 tables | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Thermoelectric materials allow the direct conversion of waste heat into
electricity, and novel materials are being investigated for this purpose.
Recently, doped Bi$_2$O$_2$Se has shown high application potential. In this
study, we discuss causes for large variation in reported transport properties
of pure Bi$_2$O$_2$Se and present a preparation method that improves the
reproducibility of undoped polycrystalline samples and improves their stability
under thermal cycling. Key steps of this method include calcination of the
Bi$_2$O$_3$ precursor, purification of the synthesized material in a
temperature gradient, use of a coarse particle fraction and compaction of the
powders in a Si3N4 die instead of a graphite die. The resulting polycrystalline
material exhibits improved reproducibility and enhanced resistance to thermal
cycling. It has room temperature electrical conductivity {\sigma}RT ~ 500 S.m-1
and Seebeck coefficient S ~ -300 $\mu$V.K$^{-1}$. These properties make it
suitable as a reference material for future doping studies. The presented
synthesis approach may provide a more reliable platform for investigating the
intrinsic behavior and doping response of Bi$_2$O$_2$Se in thermoelectric
applications.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:58:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zich",
"Jan",
""
],
[
"Plecháček",
"Tomáš",
""
],
[
"Sojka",
"Antonín",
""
],
[
"Levinský",
"Petr",
""
],
[
"Navrátil",
"Jiří",
""
],
[
"Ruleová",
"Pavlína",
""
],
[
"Šlang",
"Stanislav",
""
],
[
"Knížek",
"Karel",
""
],
[
"Hejtmánek",
"Jiří",
""
],
[
"Nečina",
"Vojtěch",
""
],
[
"Drašar",
"Čestmír",
""
]
] |
2504.15228 | Maxime Robeyns | Maxime Robeyns, Martin Szummer, Laurence Aitchison | A Self-Improving Coding Agent | Published at an ICLR 2025 workshop on Scaling Self-Improving
Foundation Models | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | We demonstrate that an LLM coding agent, equipped with basic coding tools,
can autonomously edit itself, and thereby improve its performance on benchmark
tasks. We find performance gains from 17% to 53% on a random subset of SWE
Bench Verified, with additional performance gains on LiveCodeBench, as well as
synthetically generated agent benchmarks. Our work represents an advancement in
the automated and open-ended design of agentic systems, and provides a
reference agent framework for those seeking to post-train LLMs on tool use and
other agentic tasks.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 16:58:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Robeyns",
"Maxime",
""
],
[
"Szummer",
"Martin",
""
],
[
"Aitchison",
"Laurence",
""
]
] |
2504.15229 | Takuya Boehringer | Takuya Boehringer, Jonathan Embley-Riches, Karim Hammoud, Valerio
Modugno, Dimitrios Kanoulas | Immersive Teleoperation Framework for Locomanipulation Tasks | CASE2025, 8 pages, 9 figures | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in robotic loco-manipulation have leveraged Virtual
Reality (VR) to enhance the precision and immersiveness of teleoperation
systems, significantly outperforming traditional methods reliant on 2D camera
feeds and joystick controls. Despite these advancements, challenges remain,
particularly concerning user experience across different setups. This paper
introduces a novel VR-based teleoperation framework designed for a robotic
manipulator integrated onto a mobile platform. Central to our approach is the
application of Gaussian splatting, a technique that abstracts the manipulable
scene into a VR environment, thereby enabling more intuitive and immersive
interactions. Users can navigate and manipulate within the virtual scene as if
interacting with a real robot, enhancing both the engagement and efficacy of
teleoperation tasks. An extensive user study validates our approach,
demonstrating significant usability and efficiency improvements. Two-thirds
(66%) of participants completed tasks faster, achieving an average time
reduction of 43%. Additionally, 93% preferred the Gaussian Splat interface
overall, with unanimous (100%) recommendations for future use, highlighting
improvements in precision, responsiveness, and situational awareness. Finally,
we demonstrate the effectiveness of our framework through real-world
experiments in two distinct application scenarios, showcasing the practical
capabilities and versatility of the Splat-based VR interface.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:00:31 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Boehringer",
"Takuya",
""
],
[
"Embley-Riches",
"Jonathan",
""
],
[
"Hammoud",
"Karim",
""
],
[
"Modugno",
"Valerio",
""
],
[
"Kanoulas",
"Dimitrios",
""
]
] |
2504.15230 | Tista Banerjee | Mainak Pal, Tista Banerjee | Rydberg Atoms in a Ladder Geometry: Quench Dynamics and Floquet
Engineering | 27 pages, 28 figures | null | null | null | quant-ph cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | In recent days, Rydberg atom quantum simulator platforms have emerged as
novel quantum simulators for physical systems ranging from condensed matter to
particle physics. On a fundamental level, these platforms allow for a direct
test of our understanding of the emergence of quantum statistical mechanics
starting from the laws of quantum dynamics. In this paper, we investigate the
fate of quantum dynamics in a model of Rydberg atoms arranged in a square
ladder geometry, with a Rabi frequency $2\Omega$ and a detuning profile which
is staggered along the longer direction with amplitude $\Delta$. As the
staggering strength $\Delta$ is tuned from $\Delta/\Omega=0\rightarrow\infty$,
the model exhibits a wide class of dynamical phenomena, ranging from (i)
quantum many-body scars (QMBS) ($\Delta/\Omega \sim 0,1$), (ii) integrability
induced slow dynamics and approximate Krylov fractures ($\Delta/\Omega \gg 1$)
. Additionally, by leveraging the underlying chiral nature of the spectrum of
this model Hamiltonian, it is possible to design Floquet protocols leading to
dynamical signatures reminiscent of discrete time-crystalline order and exact
Floquet flat bands. Finally, we study the robustness of these dynamical
features against imperfections in the implementation of the Floquet protocols,
long-range van der Waals interactions and inevitable influences from the
environment in the form of pure dephasing and the finite lifetime of the
Rydberg excited state.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:00:44 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Pal",
"Mainak",
""
],
[
"Banerjee",
"Tista",
""
]
] |
2504.15231 | Kanat Abdukhalikov | Kanat Abdukhalikov, Duy Ho, San Ling, Gyanendra K. Verma | Linear Complementary Pairs of Quasi-Cyclic and Quasi-Twisted Codes | null | null | null | null | cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | In this paper, we provide a polynomial characterization of linear
complementary pairs of quasi-cyclic and quasi-twisted codes of index 2. We also
give several examples of linear complementary pairs of quasi-cyclic and
quasi-twisted codes with (almost) optimal security parameters.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:08:28 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Abdukhalikov",
"Kanat",
""
],
[
"Ho",
"Duy",
""
],
[
"Ling",
"San",
""
],
[
"Verma",
"Gyanendra K.",
""
]
] |
2504.15232 | Xiaoyu Han | Xiaoyu Han, Shunyuan Zheng, Zonglin Li, Chenyang Wang, Xin Sun,
Quanling Meng | Shape-Guided Clothing Warping for Virtual Try-On | Accepted by ACM MM 2024. The code is available at
https://github.com/xyhanHIT/SCW-VTON | null | 10.1145/3664647.3680756 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image-based virtual try-on aims to seamlessly fit in-shop clothing to a
person image while maintaining pose consistency. Existing methods commonly
employ the thin plate spline (TPS) transformation or appearance flow to deform
in-shop clothing for aligning with the person's body. Despite their promising
performance, these methods often lack precise control over fine details,
leading to inconsistencies in shape between clothing and the person's body as
well as distortions in exposed limb regions. To tackle these challenges, we
propose a novel shape-guided clothing warping method for virtual try-on, dubbed
SCW-VTON, which incorporates global shape constraints and additional limb
textures to enhance the realism and consistency of the warped clothing and
try-on results. To integrate global shape constraints for clothing warping, we
devise a dual-path clothing warping module comprising a shape path and a flow
path. The former path captures the clothing shape aligned with the person's
body, while the latter path leverages the mapping between the pre- and
post-deformation of the clothing shape to guide the estimation of appearance
flow. Furthermore, to alleviate distortions in limb regions of try-on results,
we integrate detailed limb guidance by developing a limb reconstruction network
based on masked image modeling. Through the utilization of SCW-VTON, we are
able to generate try-on results with enhanced clothing shape consistency and
precise control over details. Extensive experiments demonstrate the superiority
of our approach over state-of-the-art methods both qualitatively and
quantitatively. The code is available at https://github.com/xyhanHIT/SCW-VTON.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:08:36 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Han",
"Xiaoyu",
""
],
[
"Zheng",
"Shunyuan",
""
],
[
"Li",
"Zonglin",
""
],
[
"Wang",
"Chenyang",
""
],
[
"Sun",
"Xin",
""
],
[
"Meng",
"Quanling",
""
]
] |
2504.15233 | Mayank Raikwar | Mayank Raikwar | A Review on Privacy in DAG-Based DLTs | A short version of the paper has been published at BRAINS'24
conference | null | null | null | cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Directed Acyclic Graph (DAG)-based Distributed Ledger Technologies (DLTs)
have emerged as a promising solution to the scalability issues inherent in
traditional blockchains. However, amidst the focus on scalability, the crucial
aspect of privacy within DAG-based DLTs has been largely overlooked. This paper
seeks to address this gap by providing a comprehensive examination of privacy
notions and challenges within DAG-based DLTs. We delve into potential
methodologies to enhance privacy within these systems, while also analyzing the
associated hurdles and real-world implementations within state-of-the-art
DAG-based DLTs. By exploring these methodologies, we not only illuminate the
current landscape of privacy in DAG-based DLTs but also outline future research
directions in this evolving field.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:08:45 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Raikwar",
"Mayank",
""
]
] |
2504.15234 | Vasu Tewari | Nantel Bergeron, Lucas Gagnon, Philippe Nadeau, Hunter Spink, Vasu
Tewari | Equivariant quasisymmetry and noncrossing partitions | null | null | null | null | math.CO math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a definition of ``equivariant quasisymmetry'' for polynomials in
two sets of variables. Using this definition we define quasisymmetric
generalizations of the theory of double Schur and double Schubert polynomials
that we call double fundamental polynomials and double forest polynomials,
where the subset of ``noncrossing partitions'' plays the role of $S_n$. In
subsequent work we will show this combinatorics is governed by a new geometric
construction we call the ``quasisymmetric flag variety'' which plays the same
role for equivariant quasisymmetry as the usual flag variety plays in the
classical story.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:09:52 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bergeron",
"Nantel",
""
],
[
"Gagnon",
"Lucas",
""
],
[
"Nadeau",
"Philippe",
""
],
[
"Spink",
"Hunter",
""
],
[
"Tewari",
"Vasu",
""
]
] |
2504.15235 | Kaustubh Joshi | Kaustubh Joshi, Tianchen Liu, Nikhil Chopra | Cascade IPG Observer for Underwater Robot State Estimation | ICRA 2025 (International Conference on Robotics and Automation),
Atlanta, USA | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel cascade nonlinear observer framework for inertial
state estimation. It tackles the problem of intermediate state estimation when
external localization is unavailable or in the event of a sensor outage. The
proposed observer comprises two nonlinear observers based on a recently
developed iteratively preconditioned gradient descent (IPG) algorithm. It takes
the inputs via an IMU preintegration model where the first observer is a
quaternion-based IPG. The output for the first observer is the input for the
second observer, estimating the velocity and, consequently, the position. The
proposed observer is validated on a public underwater dataset and a real-world
experiment using our robot platform. The estimation is compared with an
extended Kalman filter (EKF) and an invariant extended Kalman filter (InEKF).
Results demonstrate that our method outperforms these methods regarding better
positional accuracy and lower variance.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:10:10 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Joshi",
"Kaustubh",
""
],
[
"Liu",
"Tianchen",
""
],
[
"Chopra",
"Nikhil",
""
]
] |
2504.15236 | Esin Durmus | Saffron Huang and Esin Durmus and Miles McCain and Kunal Handa and
Alex Tamkin and Jerry Hong and Michael Stern and Arushi Somani and Xiuruo
Zhang and Deep Ganguli | Values in the Wild: Discovering and Analyzing Values in Real-World
Language Model Interactions | 44 pages | null | null | null | cs.CL cs.AI cs.CY cs.LG | http://creativecommons.org/licenses/by/4.0/ | AI assistants can impart value judgments that shape people's decisions and
worldviews, yet little is known empirically about what values these systems
rely on in practice. To address this, we develop a bottom-up,
privacy-preserving method to extract the values (normative considerations
stated or demonstrated in model responses) that Claude 3 and 3.5 models exhibit
in hundreds of thousands of real-world interactions. We empirically discover
and taxonomize 3,307 AI values and study how they vary by context. We find that
Claude expresses many practical and epistemic values, and typically supports
prosocial human values while resisting values like "moral nihilism". While some
values appear consistently across contexts (e.g. "transparency"), many are more
specialized and context-dependent, reflecting the diversity of human
interlocutors and their varied contexts. For example, "harm prevention" emerges
when Claude resists users, "historical accuracy" when responding to queries
about controversial events, "healthy boundaries" when asked for relationship
advice, and "human agency" in technology ethics discussions. By providing the
first large-scale empirical mapping of AI values in deployment, our work
creates a foundation for more grounded evaluation and design of values in AI
systems.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:13:16 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Huang",
"Saffron",
""
],
[
"Durmus",
"Esin",
""
],
[
"McCain",
"Miles",
""
],
[
"Handa",
"Kunal",
""
],
[
"Tamkin",
"Alex",
""
],
[
"Hong",
"Jerry",
""
],
[
"Stern",
"Michael",
""
],
[
"Somani",
"Arushi",
""
],
[
"Zhang",
"Xiuruo",
""
],
[
"Ganguli",
"Deep",
""
]
] |
2504.15237 | Robin Chisholm | Robin Chisholm, Elena D'Onghia, Noam Libeskind, Scott Lucchini, Andrew
J. Fox, Matthias Steinmetz | Coronal Gas in Magellanic-Analog Dwarfs: Insights from the
\textsc{HESTIA} Simulations | 7 pages, 2 figures, 1 table, Submitted to AAS Journals | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We characterize the warm circumgalactic medium (CGM) of a dwarf galaxy pair
with properties similar to the Magellanic Clouds in the \textsc{Hestia}
cosmological simulations. The system consists of a massive dwarf ($M_{\rm halo}
\sim 10^{11.5} M_{\odot}$) and a lower-mass companion ($M_{\rm halo} \sim
10^{10} M_{\odot}$), dynamically evolving in isolation before infall into a
Milky Way-mass halo. The massive dwarf hosts a warm coronal gas envelope with a
temperature of $T \sim 3 \times 10^5$ K, consistent with expectations for
virialized CGM in dwarf halos. Tidal interactions produce a neutral gas stream
that extends over $\sim 150$ kpc, with an \ion{H}{1} mass of $M_{\rm HI} \sim
10^8 M_{\odot}$, similar to the Magellanic Stream. Furthermore, in the
\textsc{Hestia} simulation suite, we find that coronal gas is ubiquitous in all
halos with $M_{\rm halo} > 10^{11} M_{\odot}$, implying that massive dwarfs
generically develop extended gaseous envelopes prior to accretion. This result
has significant implications for the survival of neutral tidal structures, and
suggests that current and future high-ion UV absorption-line observations are
indicative of warm coronae surrounding LMC-mass dwarfs, independent of their
environment.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:13:24 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chisholm",
"Robin",
""
],
[
"D'Onghia",
"Elena",
""
],
[
"Libeskind",
"Noam",
""
],
[
"Lucchini",
"Scott",
""
],
[
"Fox",
"Andrew J.",
""
],
[
"Steinmetz",
"Matthias",
""
]
] |
2504.15238 | Jeppe C. Dyre | Till B\"ohmer, Jeppe C. Dyre, and Lorenzo Costigliola | Swap Monte Carlo for diatomic molecules | null | null | null | null | cond-mat.soft | http://creativecommons.org/licenses/by/4.0/ | In recent years the Swap Monte Carlo algorithm has led to remarkable progress
in equilibrating supercooled model liquids at low temperatures. Applications
have so far been limited to systems composed of spherical particles, however,
whereas most real-world supercooled liquids are molecular. We here introduce a
simple size-polydisperse molecular model that allows for efficient thermal
equilibration \textit{in silico} with the Swap Monte Carlo method, resulting in
an estimated speedup of $10^3-10^6$ at moderate polydispersity (5-10 %).
Despite being polydisperse, the model exhibits little difference between the
size-resolved orientational time-autocorrelation functions. Our results
demonstrate the possibility of designing molecular models that can be simulated
close to the calorimetric glass transition.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:13:36 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Böhmer",
"Till",
""
],
[
"Dyre",
"Jeppe C.",
""
],
[
"Costigliola",
"Lorenzo",
""
]
] |
2504.15239 | Ghazaleh Asghari | Hicham Arroussi, Ghazaleh Asghari, Jani Virtanen | Toeplitz operators on large vector-valued Fock spaces | null | null | null | null | math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We characterize boundedness and compactness of Toeplitz operators on large
vector-valued Fock spaces with Dall'Ara's weights [Adv.\ Math., 285 (2015)
1706--1740] in terms of generalized Berezin transforms, averaging functions,
and Carleson measures. To determine Schatten class Toeplitz operators, we
introduce the operator-valued Berezin transform and averaging functions.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:13:50 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Arroussi",
"Hicham",
""
],
[
"Asghari",
"Ghazaleh",
""
],
[
"Virtanen",
"Jani",
""
]
] |
2504.15240 | Amirhossein Mollaali | Amirhossein Mollaali, Christian Bolivar Moya, Amanda A. Howard,
Alexander Heinlein, Panos Stinis, Guang Lin | Conformalized-KANs: Uncertainty Quantification with Coverage Guarantees
for Kolmogorov-Arnold Networks (KANs) in Scientific Machine Learning | 17 pages, 8 figures, | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper explores uncertainty quantification (UQ) methods in the context of
Kolmogorov-Arnold Networks (KANs). We apply an ensemble approach to KANs to
obtain a heuristic measure of UQ, enhancing interpretability and robustness in
modeling complex functions. Building on this, we introduce Conformalized-KANs,
which integrate conformal prediction, a distribution-free UQ technique, with
KAN ensembles to generate calibrated prediction intervals with guaranteed
coverage. Extensive numerical experiments are conducted to evaluate the
effectiveness of these methods, focusing particularly on the robustness and
accuracy of the prediction intervals under various hyperparameter settings. We
show that the conformal KAN predictions can be applied to recent extensions of
KANs, including Finite Basis KANs (FBKANs) and multifideilty KANs (MFKANs). The
results demonstrate the potential of our approaches to improve the reliability
and applicability of KANs in scientific machine learning.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:14:05 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Mollaali",
"Amirhossein",
""
],
[
"Moya",
"Christian Bolivar",
""
],
[
"Howard",
"Amanda A.",
""
],
[
"Heinlein",
"Alexander",
""
],
[
"Stinis",
"Panos",
""
],
[
"Lin",
"Guang",
""
]
] |
2504.15241 | Yahan Yang | Yahan Yang, Soham Dan, Shuo Li, Dan Roth, Insup Lee | MR. Guard: Multilingual Reasoning Guardrail using Curriculum Learning | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) are susceptible to adversarial attacks such as
jailbreaking, which can elicit harmful or unsafe behaviors. This vulnerability
is exacerbated in multilingual setting, where multilingual safety-aligned data
are often limited. Thus, developing a guardrail capable of detecting and
filtering unsafe content across diverse languages is critical for deploying
LLMs in real-world applications. In this work, we propose an approach to build
a multilingual guardrail with reasoning. Our method consists of: (1) synthetic
multilingual data generation incorporating culturally and linguistically
nuanced variants, (2) supervised fine-tuning, and (3) a curriculum-guided Group
Relative Policy Optimization (GRPO) framework that further improves
performance. Experimental results demonstrate that our multilingual guardrail
consistently outperforms recent baselines across both in-domain and
out-of-domain languages. The multilingual reasoning capability of our guardrail
enables it to generate multilingual explanations, which are particularly useful
for understanding language-specific risks and ambiguities in multilingual
content moderation.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:15:06 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yang",
"Yahan",
""
],
[
"Dan",
"Soham",
""
],
[
"Li",
"Shuo",
""
],
[
"Roth",
"Dan",
""
],
[
"Lee",
"Insup",
""
]
] |
2504.15242 | Pooja Gupta | Pooja Gupta | Modified Kantorovich-type Sampling Series in Orlicz Space Frameworks | null | null | null | null | math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study examines a modified Kantorovich approach applied to generalized
sampling series. The paper establishes that the approximation order to a
function using these modified operators is atleast as good as that achieved by
classical methods by using some graphs. The analysis focuses on these series
within the context of Orlicz space \( L^{\eta}(\mathbb{R}) \), specifically
looking at irregularly spaced samples. This is crucial for real-world
applications, especially in fields like signal processing and computational
mathematics, where samples are often not uniformly spaced. The paper also
establishes a result on modular convergence for functions \( g \in
L^{\eta}(\mathbb{R}) \), which includes specific cases like convergence in \(
L^{p}(\mathbb{R}) \)-spaces, \( L \log L \)-spaces, and exponential spaces. The
study then explores practical applications of the modified sampling series,
notably for discontinuous functions and provides graphs to illustrate the
results.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:15:44 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Gupta",
"Pooja",
""
]
] |
2504.15243 | Ming Yang | Ming Yang, Gang Li, Quanqi Hu, Qihang Lin, Tianbao Yang | Single-loop Algorithms for Stochastic Non-convex Optimization with
Weakly-Convex Constraints | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Constrained optimization with multiple functional inequality constraints has
significant applications in machine learning. This paper examines a crucial
subset of such problems where both the objective and constraint functions are
weakly convex. Existing methods often face limitations, including slow
convergence rates or reliance on double-loop algorithmic designs. To overcome
these challenges, we introduce a novel single-loop penalty-based stochastic
algorithm. Following the classical exact penalty method, our approach employs a
{\bf hinge-based penalty}, which permits the use of a constant penalty
parameter, enabling us to achieve a {\bf state-of-the-art complexity} for
finding an approximate Karush-Kuhn-Tucker (KKT) solution. We further extend our
algorithm to address finite-sum coupled compositional objectives, which are
prevalent in artificial intelligence applications, establishing improved
complexity over existing approaches. Finally, we validate our method through
experiments on fair learning with receiver operating characteristic (ROC)
fairness constraints and continual learning with non-forgetting constraints.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:15:48 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Yang",
"Ming",
""
],
[
"Li",
"Gang",
""
],
[
"Hu",
"Quanqi",
""
],
[
"Lin",
"Qihang",
""
],
[
"Yang",
"Tianbao",
""
]
] |
2504.15244 | Lisheng Ren | Ilias Diakonikolas, Daniel M. Kane, Lisheng Ren | Faster Algorithms for Agnostically Learning Disjunctions and their
Implications | null | null | null | null | cs.LG cs.DS stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the algorithmic task of learning Boolean disjunctions in the
distribution-free agnostic PAC model. The best known agnostic learner for the
class of disjunctions over $\{0, 1\}^n$ is the $L_1$-polynomial regression
algorithm, achieving complexity $2^{\tilde{O}(n^{1/2})}$. This complexity bound
is known to be nearly best possible within the class of Correlational
Statistical Query (CSQ) algorithms. In this work, we develop an agnostic
learner for this concept class with complexity $2^{\tilde{O}(n^{1/3})}$. Our
algorithm can be implemented in the Statistical Query (SQ) model, providing the
first separation between the SQ and CSQ models in distribution-free agnostic
learning.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:16:14 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Diakonikolas",
"Ilias",
""
],
[
"Kane",
"Daniel M.",
""
],
[
"Ren",
"Lisheng",
""
]
] |
2504.15245 | Chunjian Zhang | Chunjian Zhang, Jiangyong Jia, Jinhui Chen, Chun Shen, Lumeng Liu | Probing the octupole deformation of $^{238}$U in high-energy nuclear
collisions | 6 pages, 4 figures | null | null | null | nucl-th nucl-ex | http://creativecommons.org/licenses/by/4.0/ | Some atomic nuclei exhibit ``pear" shapes arising from octupole deformation
($\beta_3$), though direct experimental evidences for such exotic shapes
remains scarce. Low-energy model studies suggest $^{238}$U may have a modest
octupole deformation arising from collective vibrational degrees of freedom, in
addition to a large prolate shape. We investigated the impact of this modest
octupole shape on observables involving triangular flow ($v_3$) in high-energy
nuclear collisions. Using a hydrodynamic framework, we show $v_3$ and its
correlation with mean transverse momentum, $\langle v_3^2 \delta\pT \rangle$,
exhibit strong sensitivity to $\beta_3$. We found that $\langle v_3^2\rangle$
follows a linear increase with $\beta_3^2$, while $\langle v_3^2 \delta\pT
\rangle$ is suppressed in the presence of $\beta_3$. Our findings show that the
collective-flow-assisted nuclear imaging method in high-energy nuclear
collisions, when compared with experimental data, can provide unique
constraints on higher-order deformations.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:17:33 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Chunjian",
""
],
[
"Jia",
"Jiangyong",
""
],
[
"Chen",
"Jinhui",
""
],
[
"Shen",
"Chun",
""
],
[
"Liu",
"Lumeng",
""
]
] |
2504.15246 | James Bailie | James Bailie and Ruobin Gong and Xiao-Li Meng | A Refreshment Stirred, Not Shaken (III): Can Swapping Be Differentially
Private? | 27 pages, 1 figure | null | null | null | cs.CR stat.OT | http://creativecommons.org/licenses/by/4.0/ | The quest for a precise and contextually grounded answer to the question in
the present paper's title resulted in this stirred-not-shaken triptych, a
phrase that reflects our desire to deepen the theoretical basis, broaden the
practical applicability, and reduce the misperception of differential privacy
(DP)$\unicode{x2014}$all without shaking its core foundations. Indeed, given
the existence of more than 200 formulations of DP (and counting), before even
attempting to answer the titular question one must first precisely specify what
it actually means to be DP. Motivated by this observation, a theoretical
investigation into DP's fundamental essence resulted in Part I of this trio,
which introduces a five-building-block system explicating the who, where, what,
how and how much aspects of DP. Instantiating this system in the context of the
United States Decennial Census, Part II then demonstrates the broader
applicability and relevance of DP by comparing a swapping strategy like that
used in 2010 with the TopDown Algorithm$\unicode{x2014}$a DP method adopted in
the 2020 Census. This paper provides nontechnical summaries of the preceding
two parts as well as new discussion$\unicode{x2014}$for example, on how greater
awareness of the five building blocks can thwart privacy theatrics; how our
results bridging traditional SDC and DP allow a data custodian to reap the
benefits of both these fields; how invariants impact disclosure risk; and how
removing the implicit reliance on aleatoric uncertainty could lead to new
generalizations of DP.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:19:57 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Bailie",
"James",
""
],
[
"Gong",
"Ruobin",
""
],
[
"Meng",
"Xiao-Li",
""
]
] |
2504.15247 | Weston Pace | Weston Pace, Chang She, Lei Xu, Will Jones, Albert Lockett, Jun Wang,
Raunak Shah | Lance: Efficient Random Access in Columnar Storage through Adaptive
Structural Encodings | null | null | null | null | cs.DB | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The growing interest in artificial intelligence has created workloads that
require both sequential and random access. At the same time, NVMe-backed
storage solutions have emerged, providing caching capability for large columnar
datasets in cloud storage. Current columnar storage libraries fall short of
effectively utilizing an NVMe device's capabilities, especially when it comes
to random access. Historically, this has been assumed an implicit weakness in
columnar storage formats, but this has not been sufficiently explored. In this
paper, we examine the effectiveness of popular columnar formats such as Apache
Arrow, Apache Parquet, and Lance in both random access and full scan tasks
against NVMe storage.
We argue that effective encoding of a column's structure, such as the
repetition and validity information, is the key to unlocking the disk's
performance. We show that Parquet, when configured correctly, can achieve over
60x better random access performance than default settings. We also show that
this high random access performance requires making minor trade-offs in scan
performance and RAM utilization. We then describe the Lance structural encoding
scheme, which alternates between two different structural encodings based on
data width, and achieves better random access performance without making
trade-offs in scan performance or RAM utilization.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:22:18 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Pace",
"Weston",
""
],
[
"She",
"Chang",
""
],
[
"Xu",
"Lei",
""
],
[
"Jones",
"Will",
""
],
[
"Lockett",
"Albert",
""
],
[
"Wang",
"Jun",
""
],
[
"Shah",
"Raunak",
""
]
] |
2504.15249 | Matteo Clerici | Thomas Dickinson, Ivi Afxenti, Giedre Astrauskaite, Lennart Hirsch,
Samuel Nerenberg, Ottavia Jedrkiewicz, Daniele Faccio, Caroline
M\"ullenbroich, Alessandra Gatti, Matteo Clerici, Lucia Caspani | Quantum-enhanced second harmonic generation beyond the photon pairs
regime | null | null | null | null | quant-ph physics.optics | http://creativecommons.org/licenses/by/4.0/ | Two-photon processes are crucial in applications like microscopy and
microfabrication, but their low cross-section requires intense illumination and
limits, e.g., the penetration depth in nonlinear microscopy. Entangled states
have been proposed to enhance the efficiency of two-photon interactions and
have shown effectiveness at low intensities. This quantum enhancement is
generally believed to be lost at high intensities, for more than one photon per
mode, raising doubts about its usefulness. We explored experimentally and
theoretically two-photon processes driven by entangled photons at intensities
beyond this threshold and compared the results with the classical case. We
found that a quantum advantage can still be observed at nearly one order of
magnitude higher intensities than previously assumed. Our findings show a
potential path for exploiting quantum-enhanced two-photon processes in
practical applications.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:27:26 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Dickinson",
"Thomas",
""
],
[
"Afxenti",
"Ivi",
""
],
[
"Astrauskaite",
"Giedre",
""
],
[
"Hirsch",
"Lennart",
""
],
[
"Nerenberg",
"Samuel",
""
],
[
"Jedrkiewicz",
"Ottavia",
""
],
[
"Faccio",
"Daniele",
""
],
[
"Müllenbroich",
"Caroline",
""
],
[
"Gatti",
"Alessandra",
""
],
[
"Clerici",
"Matteo",
""
],
[
"Caspani",
"Lucia",
""
]
] |
2504.15250 | Ritwick Sarkar | Ritwick Sarkar and Ion Santra | Tracer dynamics in an interacting active bath: fluctuations and energy
partition | 27 pages, 10 figures | null | null | null | cond-mat.stat-mech cond-mat.soft | http://creativecommons.org/licenses/by/4.0/ | We investigate the dynamics of a massive tracer particle coupled to an
interacting active bath, modeled as a harmonic chain of overdamped active
particles analytically, with an aim to understand the impact of bath
interactions and activity on the nonequilibrium fluctuations of the tracer.
From the microscopic equations, we derive the tracer particle's effective
Langevin equation, obtaining the dissipative and stochastic forces from the
bath. We analyze the friction kernel, revealing power-law tails in the weak
coupling limit and exponential decay in the strong coupling regime. Due to the
interplay between bath interactions, probe-bath coupling, and activity, the
mean squared displacement, velocity, and stationary velocity correlations
exhibit different dynamical regimes, which we characterize analytically. Under
harmonic confinement, we find that energy equipartition holds at low activity
but breaks down at higher activity, with the kinetic energy exhibiting a
non-monotonic dependence on the activity of the bath.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:27:41 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Sarkar",
"Ritwick",
""
],
[
"Santra",
"Ion",
""
]
] |
2504.15251 | Thanasis Pittas | Ilias Diakonikolas, Daniel M. Kane, Sushrut Karmalkar, Jasper C.H.
Lee, Thanasis Pittas | On Learning Parallel Pancakes with Mostly Uniform Weights | null | null | null | null | cs.LG cs.DS math.ST stat.ML stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the complexity of learning $k$-mixtures of Gaussians ($k$-GMMs) on
$\mathbb{R}^d$. This task is known to have complexity $d^{\Omega(k)}$ in full
generality. To circumvent this exponential lower bound on the number of
components, research has focused on learning families of GMMs satisfying
additional structural properties. A natural assumption posits that the
component weights are not exponentially small and that the components have the
same unknown covariance. Recent work gave a $d^{O(\log(1/w_{\min}))}$-time
algorithm for this class of GMMs, where $w_{\min}$ is the minimum weight. Our
first main result is a Statistical Query (SQ) lower bound showing that this
quasi-polynomial upper bound is essentially best possible, even for the special
case of uniform weights. Specifically, we show that it is SQ-hard to
distinguish between such a mixture and the standard Gaussian. We further
explore how the distribution of weights affects the complexity of this task.
Our second main result is a quasi-polynomial upper bound for the aforementioned
testing task when most of the weights are uniform while a small fraction of the
weights are potentially arbitrary.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:31:55 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Diakonikolas",
"Ilias",
""
],
[
"Kane",
"Daniel M.",
""
],
[
"Karmalkar",
"Sushrut",
""
],
[
"Lee",
"Jasper C. H.",
""
],
[
"Pittas",
"Thanasis",
""
]
] |
2504.15252 | Lakshay Sharma | Tue Vo, Lakshay Sharma, Tuan Dinh, Khuong Dinh, Trang Nguyen, Trung
Phan, Minh Do, Duong Vu | SuoiAI: Building a Dataset for Aquatic Invertebrates in Vietnam | Published as a workshop paper at "Tackling Climate Change with
Machine Learning", ICLR 2025 | null | null | null | cs.AI cs.CV cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Understanding and monitoring aquatic biodiversity is critical for ecological
health and conservation efforts. This paper proposes SuoiAI, an end-to-end
pipeline for building a dataset of aquatic invertebrates in Vietnam and
employing machine learning (ML) techniques for species classification. We
outline the methods for data collection, annotation, and model training,
focusing on reducing annotation effort through semi-supervised learning and
leveraging state-of-the-art object detection and classification models. Our
approach aims to overcome challenges such as data scarcity, fine-grained
classification, and deployment in diverse environmental conditions.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:33:02 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Vo",
"Tue",
""
],
[
"Sharma",
"Lakshay",
""
],
[
"Dinh",
"Tuan",
""
],
[
"Dinh",
"Khuong",
""
],
[
"Nguyen",
"Trang",
""
],
[
"Phan",
"Trung",
""
],
[
"Do",
"Minh",
""
],
[
"Vu",
"Duong",
""
]
] |
2504.15253 | Yilun Zhou | Yilun Zhou, Austin Xu, Peifeng Wang, Caiming Xiong, Shafiq Joty | Evaluating Judges as Evaluators: The JETTS Benchmark of LLM-as-Judges as
Test-Time Scaling Evaluators | The first two authors contributed equally. The codebase is at
https://github.com/SalesforceAIResearch/jetts-benchmark | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Scaling test-time computation, or affording a generator large language model
(LLM) extra compute during inference, typically employs the help of external
non-generative evaluators (i.e., reward models). Concurrently, LLM-judges,
models trained to generate evaluations and critiques (explanations) in natural
language, are becoming increasingly popular in automatic evaluation. Despite
judge empirical successes, their effectiveness as evaluators in test-time
scaling settings is largely unknown. In this paper, we introduce the Judge
Evaluation for Test-Time Scaling (JETTS) benchmark, which evaluates judge
performance in three domains (math reasoning, code generation, and instruction
following) under three task settings: response reranking, step-level beam
search, and critique-based response refinement. We evaluate 10 different judge
models (7B-70B parameters) for 8 different base generator models (6.7B-72B
parameters). Our benchmark shows that while judges are competitive with outcome
reward models in reranking, they are consistently worse than process reward
models in beam search procedures. Furthermore, though unique to LLM-judges,
their natural language critiques are currently ineffective in guiding the
generator towards better responses.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:33:23 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhou",
"Yilun",
""
],
[
"Xu",
"Austin",
""
],
[
"Wang",
"Peifeng",
""
],
[
"Xiong",
"Caiming",
""
],
[
"Joty",
"Shafiq",
""
]
] |
2504.15254 | Anirudh Khatry | Anirudh Khatry, Robert Zhang, Jia Pan, Ziteng Wang, Qiaochu Chen, Greg
Durrett, Isil Dillig | CRUST-Bench: A Comprehensive Benchmark for C-to-safe-Rust Transpilation | null | null | null | null | cs.SE cs.CL | http://creativecommons.org/licenses/by/4.0/ | C-to-Rust transpilation is essential for modernizing legacy C code while
enhancing safety and interoperability with modern Rust ecosystems. However, no
dataset currently exists for evaluating whether a system can transpile C into
safe Rust that passes a set of test cases. We introduce CRUST-Bench, a dataset
of 100 C repositories, each paired with manually-written interfaces in safe
Rust as well as test cases that can be used to validate correctness of the
transpilation. By considering entire repositories rather than isolated
functions, CRUST-Bench captures the challenges of translating complex projects
with dependencies across multiple files. The provided Rust interfaces provide
explicit specifications that ensure adherence to idiomatic, memory-safe Rust
patterns, while the accompanying test cases enforce functional correctness. We
evaluate state-of-the-art large language models (LLMs) on this task and find
that safe and idiomatic Rust generation is still a challenging problem for
various state-of-the-art methods and techniques. We also provide insights into
the errors LLMs usually make in transpiling code from C to safe Rust. The best
performing model, OpenAI o1, is able to solve only 15 tasks in a single-shot
setting. Improvements on CRUST-Bench would lead to improved transpilation
systems that can reason about complex scenarios and help in migrating legacy
codebases from C into languages like Rust that ensure memory safety. You can
find the dataset and code at https://github.com/anirudhkhatry/CRUST-bench.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:33:33 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Khatry",
"Anirudh",
""
],
[
"Zhang",
"Robert",
""
],
[
"Pan",
"Jia",
""
],
[
"Wang",
"Ziteng",
""
],
[
"Chen",
"Qiaochu",
""
],
[
"Durrett",
"Greg",
""
],
[
"Dillig",
"Isil",
""
]
] |
2504.15255 | Bingjie Wang | Bingjie Wang, Joel Leja, Hakim Atek, Rachel Bezanson, Emilie Burnham,
Pratika Dayal, Robert Feldmann, Jenny E. Greene, Benjamin D. Johnson, Ivo
Labbe, Michael V. Maseda, Themiya Nanayakkara, Sedona H. Price, Katherine A.
Suess, John R. Weaver, Katherine E. Whitaker | Population Models for Star Formation Timescales in Early Galaxies: The
First Step Towards Solving Outshining in Star Formation History Inference | Submitted to ApJ on 12/18/24; this version corresponds to a revised
version after a reviewer's report. 27 pages, 17 figures | null | null | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | JWST have revealed temporarily-quenched and ultraviolet-luminous galaxies in
the early universe, suggesting enhanced star formation stochasticity. Verifying
this hypothesis is critical, yet challenging; outshining, wherein light from
young stars dominates the spectral energy distribution, represents perhaps the
greatest challenge in inferring the formation histories of unresolved galaxies.
In this paper, we take a simple model of burstiness and show that
state-of-the-art inference methods with flexible star formation histories
(SFHs) and neutral priors, while recovering average star formation rates (SFRs;
$\sim0.1$ dex median offset), fail to recover the complexities of fluctuations
on tens of Myr timescales, and typically underestimate masses in bursty systems
($\sim0.15$ dex). Surprisingly, detailed SFH recovery is still sensitive to
priors even when data quality is optimal, e.g., including high signal-to-noise
($\rm20~pixel^{-1}$) spectroscopy with wide coverage (rest-frame
$0.12-1.06~\mu$m). Crucially, however, refitting the same data with a prior
correctly encoding the bursty expectation eliminates these biases: median
offsets in mass and SFRs decrease to $\sim 0.04$ dex and $\sim 0.05$ dex,
respectively. Under the assumption that current population burstiness predicts
past SFH, the solution to outshining in modeling statistical samples is
empirically measuring recent galaxy SFHs with population modeling. A prototype
is H$\alpha$/UV: while helpful, it is insufficient to constrain the expected
complex burstiness. To this end, we introduce a more complete, quantitative
population-level approach and demonstrate that it promises to recover the
typical amplitude, timescale, and slope of the recent SFH to high accuracy.
This approach thus has the strong potential to solve outshining using
observations from JWST.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:34:55 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Wang",
"Bingjie",
""
],
[
"Leja",
"Joel",
""
],
[
"Atek",
"Hakim",
""
],
[
"Bezanson",
"Rachel",
""
],
[
"Burnham",
"Emilie",
""
],
[
"Dayal",
"Pratika",
""
],
[
"Feldmann",
"Robert",
""
],
[
"Greene",
"Jenny E.",
""
],
[
"Johnson",
"Benjamin D.",
""
],
[
"Labbe",
"Ivo",
""
],
[
"Maseda",
"Michael V.",
""
],
[
"Nanayakkara",
"Themiya",
""
],
[
"Price",
"Sedona H.",
""
],
[
"Suess",
"Katherine A.",
""
],
[
"Weaver",
"John R.",
""
],
[
"Whitaker",
"Katherine E.",
""
]
] |
2504.15256 | Mustafa Kad{\i}lar | Mustafa Can Kadilar, Ersin Topta\c{s} and Gazi Akg\"un | Impulsive pattern recognition of a myoelectric hand via Dynamic Time
Warping | 6 pages, 3 tables, 8 figures. Part of this work was presented at the
International Journal of Advanced Natural Sciences and Engineering Researches
(ICIAS) 2024 | null | null | null | physics.bio-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Although myoelectric prosthetic hands provide amputees with intuitive
control, their reliance on many EMG sensors limits accessibility and makes them
complex and expensive. To address this problem, this work presents a different
perspective that makes use of a single EMG sensor and brief impulse signals in
conjunction with Dynamic Time Warping (DTW) for accurate pattern detection.
Conventional techniques rely on real-time data from multiple sensors, which can
be costly and bulky. The method presents high accuracy while lowering hardware
complexity and expense. A DTW-based system that reliably identifies muscle
activation patterns from short EMG signals was created and tested. Results show
that this single-sensor approach obtained an accuracy rate of 92%, which is
similar to that of conventional multi-sensor systems. This research provides a
more straightforward and economical approach that can be used to obtain
enhanced myoelectric control. These findings provide a different perspective on
more easily accessible and user-friendly prosthetic devices, which will be
especially helpful in disaster-affected areas where quick deployment is
essential. Future improvements would investigate this system's dependability
over time and wider implementations in real situations, to take prosthetic
technology one step further.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:35:12 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Kadilar",
"Mustafa Can",
""
],
[
"Toptaş",
"Ersin",
""
],
[
"Akgün",
"Gazi",
""
]
] |
2504.15257 | Yue Liu | Hongcheng Gao, Yue Liu, Yufei He, Longxu Dou, Chao Du, Zhijie Deng,
Bryan Hooi, Min Lin, Tianyu Pang | FlowReasoner: Reinforcing Query-Level Meta-Agents | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes a query-level meta-agent named FlowReasoner to automate
the design of query-level multi-agent systems, i.e., one system per user query.
Our core idea is to incentivize a reasoning-based meta-agent via external
execution feedback. Concretely, by distilling DeepSeek R1, we first endow the
basic reasoning ability regarding the generation of multi-agent systems to
FlowReasoner. Then, we further enhance it via reinforcement learning (RL) with
external execution feedback. A multi-purpose reward is designed to guide the RL
training from aspects of performance, complexity, and efficiency. In this
manner, FlowReasoner is enabled to generate a personalized multi-agent system
for each user query via deliberative reasoning. Experiments on both engineering
and competition code benchmarks demonstrate the superiority of FlowReasoner.
Remarkably, it surpasses o1-mini by 10.52% accuracy across three benchmarks.
The code is available at https://github.com/sail-sg/FlowReasoner.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:35:42 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Gao",
"Hongcheng",
""
],
[
"Liu",
"Yue",
""
],
[
"He",
"Yufei",
""
],
[
"Dou",
"Longxu",
""
],
[
"Du",
"Chao",
""
],
[
"Deng",
"Zhijie",
""
],
[
"Hooi",
"Bryan",
""
],
[
"Lin",
"Min",
""
],
[
"Pang",
"Tianyu",
""
]
] |
2504.15258 | Cristian Bonato | Issam Belgacem, Pasquale Cilibrizzi, Muhammad Junaid Arshad, Daniel
White, Malte Kroj, Christiaan Bekker, Margherita Mazzera, Brian D. Gerardot,
Angelo C. Frangeskou, Gavin W. Morley, Nguyen Tien Son, Jawad Ul-Hassan,
Takeshi Ohshima, Hiroshi Abe, Lorenzo Vinco, Dario Polli, Giulio Cerullo,
Cristian Bonato | Broadband Fourier transform spectroscopy of quantum emitters
photoluminescence with sub-nanosecond temporal resolution | 16 pages, 9 figures | null | null | null | quant-ph cond-mat.mes-hall physics.chem-ph physics.optics | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The spectral characterization of quantum emitter luminescence over broad
wavelength ranges and fast timescales is important for applications ranging
from biophysics to quantum technologies. Here we present the application of
time-domain Fourier transform spectroscopy, based on a compact and stable
birefringent interferometer coupled to low-dark-count superconducting
single-photon detectors, to the study of quantum emitters. We experimentally
demonstrate that the system enables spectroscopy of quantum emitters over a
broad wavelength interval from the near-infrared to the telecom range, where
grating-based spectrometers coupled to InGaAs cameras are typically noisy and
inefficient. We further show that the high temporal resolution of single-photon
detectors, which can be on the order of tens of picoseconds, enables the
monitoring of spin-dependent spectral changes on sub-nanosecond timescales.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:38:20 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Belgacem",
"Issam",
""
],
[
"Cilibrizzi",
"Pasquale",
""
],
[
"Arshad",
"Muhammad Junaid",
""
],
[
"White",
"Daniel",
""
],
[
"Kroj",
"Malte",
""
],
[
"Bekker",
"Christiaan",
""
],
[
"Mazzera",
"Margherita",
""
],
[
"Gerardot",
"Brian D.",
""
],
[
"Frangeskou",
"Angelo C.",
""
],
[
"Morley",
"Gavin W.",
""
],
[
"Son",
"Nguyen Tien",
""
],
[
"Ul-Hassan",
"Jawad",
""
],
[
"Ohshima",
"Takeshi",
""
],
[
"Abe",
"Hiroshi",
""
],
[
"Vinco",
"Lorenzo",
""
],
[
"Polli",
"Dario",
""
],
[
"Cerullo",
"Giulio",
""
],
[
"Bonato",
"Cristian",
""
]
] |
2504.15259 | Yunxuan Cai | Yunxuan Cai, Sitao Xiang, Zongjian Li, Haiwei Chen, Yajie Zhao | Bringing Diversity from Diffusion Models to Semantic-Guided Face Asset
Generation | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Digital modeling and reconstruction of human faces serve various
applications. However, its availability is often hindered by the requirements
of data capturing devices, manual labor, and suitable actors. This situation
restricts the diversity, expressiveness, and control over the resulting models.
This work aims to demonstrate that a semantically controllable generative
network can provide enhanced control over the digital face modeling process. To
enhance diversity beyond the limited human faces scanned in a controlled
setting, we introduce a novel data generation pipeline that creates a
high-quality 3D face database using a pre-trained diffusion model. Our proposed
normalization module converts synthesized data from the diffusion model into
high-quality scanned data. Using the 44,000 face models we obtained, we further
developed an efficient GAN-based generator. This generator accepts semantic
attributes as input, and generates geometry and albedo. It also allows
continuous post-editing of attributes in the latent space. Our asset refinement
component subsequently creates physically-based facial assets. We introduce a
comprehensive system designed for creating and editing high-quality face
assets. Our proposed model has undergone extensive experiment, comparison and
evaluation. We also integrate everything into a web-based interactive tool. We
aim to make this tool publicly available with the release of the paper.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:38:50 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Cai",
"Yunxuan",
""
],
[
"Xiang",
"Sitao",
""
],
[
"Li",
"Zongjian",
""
],
[
"Chen",
"Haiwei",
""
],
[
"Zhao",
"Yajie",
""
]
] |
2504.15260 | Xuesong Liu | Xuesong Liu, Yansong Liu, Haoyu Tang, Fangzhou Zhao, Le Xia, and Yao
Sun | Joint Knowledge and Power Management for Secure Semantic Communication
Networks | null | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, semantic communication (SemCom) has shown its great superiorities
in resource savings and information exchanges. However, while its unique
background knowledge guarantees accurate semantic reasoning and recovery,
semantic information security-related concerns are introduced at the same time.
Since the potential eavesdroppers may have the same background knowledge to
accurately decrypt the private semantic information transmitted between legal
SemCom users, this makes the knowledge management in SemCom networks rather
challenging in joint consideration with the power control. To this end, this
paper focuses on jointly addressing three core issues of power allocation,
knowledge base caching (KBC), and device-to-device (D2D) user pairing (DUP) in
secure SemCom networks. We first develop a novel performance metric, namely
semantic secrecy throughput (SST), to quantify the information security level
that can be achieved at each pair of D2D SemCom users. Next, an SST
maximization problem is formulated subject to secure SemCom-related delay and
reliability constraints. Afterward, we propose a security-aware resource
management solution using the Lagrange primal-dual method and a two-stage
method. Simulation results demonstrate our proposed solution nearly doubles the
SST performance and realizes less than half of the queuing delay performance
compared to different benchmarks.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:39:59 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Liu",
"Xuesong",
""
],
[
"Liu",
"Yansong",
""
],
[
"Tang",
"Haoyu",
""
],
[
"Zhao",
"Fangzhou",
""
],
[
"Xia",
"Le",
""
],
[
"Sun",
"Yao",
""
]
] |
2504.15261 | Mohammad Beheshti | Mohammad Beheshti, Lovedeep Gondara, Iris Zachary | Leveraging Language Models for Automated Patient Record Linkage | null | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Objective: Healthcare data fragmentation presents a major challenge for
linking patient data, necessitating robust record linkage to integrate patient
records from diverse sources. This study investigates the feasibility of
leveraging language models for automated patient record linkage, focusing on
two key tasks: blocking and matching. Materials and Methods: We utilized
real-world healthcare data from the Missouri Cancer Registry and Research
Center, linking patient records from two independent sources using
probabilistic linkage as a baseline. A transformer-based model, RoBERTa, was
fine-tuned for blocking using sentence embeddings. For matching, several
language models were experimented under fine-tuned and zero-shot settings,
assessing their performance against ground truth labels. Results: The
fine-tuned blocking model achieved a 92% reduction in the number of candidate
pairs while maintaining near-perfect recall. In the matching task, fine-tuned
Mistral-7B achieved the best performance with only 6 incorrect predictions.
Among zero-shot models, Mistral-Small-24B performed best, with a total of 55
incorrect predictions. Discussion: Fine-tuned language models achieved strong
performance in patient record blocking and matching with minimal errors.
However, they remain less accurate and efficient than a hybrid rule-based and
probabilistic approach for blocking. Additionally, reasoning models like
DeepSeek-R1 are impractical for large-scale record linkage due to high
computational costs. Conclusion: This study highlights the potential of
language models for automating patient record linkage, offering improved
efficiency by eliminating the manual efforts required to perform patient record
linkage. Overall, language models offer a scalable solution that can enhance
data integration, reduce manual effort, and support disease surveillance and
research.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:41:15 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Beheshti",
"Mohammad",
""
],
[
"Gondara",
"Lovedeep",
""
],
[
"Zachary",
"Iris",
""
]
] |
2504.15262 | Brandon Zhao | Brandon Zhao, Aviad Levis, Liam Connor, Pratul P. Srinivasan,
Katherine L. Bouman | Revealing the 3D Cosmic Web through Gravitationally Constrained Neural
Fields | null | null | null | null | astro-ph.CO cs.CV | http://creativecommons.org/licenses/by/4.0/ | Weak gravitational lensing is the slight distortion of galaxy shapes caused
primarily by the gravitational effects of dark matter in the universe. In our
work, we seek to invert the weak lensing signal from 2D telescope images to
reconstruct a 3D map of the universe's dark matter field. While inversion
typically yields a 2D projection of the dark matter field, accurate 3D maps of
the dark matter distribution are essential for localizing structures of
interest and testing theories of our universe. However, 3D inversion poses
significant challenges. First, unlike standard 3D reconstruction that relies on
multiple viewpoints, in this case, images are only observed from a single
viewpoint. This challenge can be partially addressed by observing how galaxy
emitters throughout the volume are lensed. However, this leads to the second
challenge: the shapes and exact locations of unlensed galaxies are unknown, and
can only be estimated with a very large degree of uncertainty. This introduces
an overwhelming amount of noise which nearly drowns out the lensing signal
completely. Previous approaches tackle this by imposing strong assumptions
about the structures in the volume. We instead propose a methodology using a
gravitationally-constrained neural field to flexibly model the continuous
matter distribution. We take an analysis-by-synthesis approach, optimizing the
weights of the neural network through a fully differentiable physical forward
model to reproduce the lensing signal present in image measurements. We
showcase our method on simulations, including realistic simulated measurements
of dark matter distributions that mimic data from upcoming telescope surveys.
Our results show that our method can not only outperform previous methods, but
importantly is also able to recover potentially surprising dark matter
structures.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:43:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhao",
"Brandon",
""
],
[
"Levis",
"Aviad",
""
],
[
"Connor",
"Liam",
""
],
[
"Srinivasan",
"Pratul P.",
""
],
[
"Bouman",
"Katherine L.",
""
]
] |
2504.15263 | Ehsan Ahmadi | Ehsan Ahmadi, Chao Wang | Interpretable Locomotion Prediction in Construction Using a
Memory-Driven LLM Agent With Chain-of-Thought Reasoning | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | Construction tasks are inherently unpredictable, with dynamic environments
and safety-critical demands posing significant risks to workers. Exoskeletons
offer potential assistance but falter without accurate intent recognition
across diverse locomotion modes. This paper presents a locomotion prediction
agent leveraging Large Language Models (LLMs) augmented with memory systems,
aimed at improving exoskeleton assistance in such settings. Using multimodal
inputs - spoken commands and visual data from smart glasses - the agent
integrates a Perception Module, Short-Term Memory (STM), Long-Term Memory
(LTM), and Refinement Module to predict locomotion modes effectively.
Evaluation reveals a baseline weighted F1-score of 0.73 without memory, rising
to 0.81 with STM, and reaching 0.90 with both STM and LTM, excelling with vague
and safety-critical commands. Calibration metrics, including a Brier Score drop
from 0.244 to 0.090 and ECE from 0.222 to 0.044, affirm improved reliability.
This framework supports safer, high-level human-exoskeleton collaboration, with
promise for adaptive assistive systems in dynamic industries.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:45:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Ahmadi",
"Ehsan",
""
],
[
"Wang",
"Chao",
""
]
] |
2504.15264 | Barnab\'as Janzer | Barnab\'as Janzer, Zhihan Jin, Benny Sudakov and Kewen Wu | Sunflowers and Ramsey problems for restricted intersections | 23 pages + 7-page appendix | null | null | null | math.CO cs.DM quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extremal problems on set systems with restricted intersections have been an
important part of combinatorics in the last 70 year. In this paper, we study
the following Ramsey version of these problems. Given a set $L\subseteq
\{0,\dots,k-1\}$ and a family $\mathcal{F}$ of $k$-element sets which does not
contain a sunflower with $m$ petals whose kernel size is in $L$, how large a
subfamily of $\mathcal{F}$ can we find in which no pair has intersection size
in $L$? We give matching upper and lower bounds, determining the dependence on
$m$ for all $k$ and $L$. This problem also finds applications in quantum
computing.
As an application of our techniques, we also obtain a variant of F\"uredi's
celebrated semilattice lemma, which is a key tool in the powerful delta-system
method. We prove that one cannot remove the double-exponential dependency on
the uniformity in F\"uredi's result, however, we provide an alternative with
significantly better, single-exponential dependency on the parameters, which is
still strong enough for most applications of the delta-system method.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:46:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Janzer",
"Barnabás",
""
],
[
"Jin",
"Zhihan",
""
],
[
"Sudakov",
"Benny",
""
],
[
"Wu",
"Kewen",
""
]
] |
2504.15265 | Tharrmashastha Sapv | Yash Saxena, Sagnik Chatterjee, Tharrmashastha Sapv | Realization of maximally-entangling two-qutrit gates using the
Cross-Resonance scheme | Experiments on actual backend are underway. Comments welcome! | null | null | null | quant-ph | http://creativecommons.org/licenses/by-sa/4.0/ | Three-level systems have natural advantages over their two-level counterparts
in quantum information and computation. Although generally used for qubits, the
existing superconducting transmon architecture can naturally be extended to
three levels, which allows the exploration of the entire qutrit Hilbert space.
Realizing a maximally entangling two-qutrit gate is necessary for universal
quantum computation using three-level systems. Our work extends the
state-of-the-art approach by providing a theoretical framework for realizing a
high-fidelity two-qutrit generalized CR gate using the cross-resonance
framework. Using our framework, we experimentally demonstrate two-qutrit
generalized controlled-$\sqrt{X}$ gates through simulations on QISKIT dynamics,
which in turn allow us to obtain a maximum extracted concurrence of
$97.2\pm0.1\%$, thereby demonstrating that entanglement has been achieved on
all three levels. Using the above gates, we also prepare a two-qutrit Bell
state with a fidelity of $96.9\pm 0.1\%$.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:47:16 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Saxena",
"Yash",
""
],
[
"Chatterjee",
"Sagnik",
""
],
[
"Sapv",
"Tharrmashastha",
""
]
] |
2504.15266 | Chen Henry Wu | Vaishnavh Nagarajan and Chen Henry Wu and Charles Ding and Aditi
Raghunathan | Roll the dice & look before you leap: Going beyond the creative limits
of next-token prediction | 37 pages | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | We design a suite of minimal algorithmic tasks that are a loose abstraction
of open-ended real-world tasks. This allows us to cleanly and controllably
quantify the creative limits of the present-day language model. Much like
real-world tasks that require a creative, far-sighted leap of thought, our
tasks require an implicit, open-ended stochastic planning step that either (a)
discovers new connections in an abstract knowledge graph (like in wordplay,
drawing analogies, or research) or (b) constructs new patterns (like in
designing math problems or new proteins). In these tasks, we empirically and
conceptually argue how next-token learning is myopic and memorizes excessively;
comparatively, multi-token approaches, namely teacherless training and
diffusion models, excel in producing diverse and original output. Secondly, in
our tasks, we find that to elicit randomness from the Transformer without
hurting coherence, it is better to inject noise right at the input layer (via a
method we dub hash-conditioning) rather than defer to temperature sampling from
the output layer. Thus, our work offers a principled, minimal test-bed for
analyzing open-ended creative skills, and offers new arguments for going beyond
next-token learning and softmax-based sampling. We make part of the code
available under https://github.com/chenwu98/algorithmic-creativity
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:47:46 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Nagarajan",
"Vaishnavh",
""
],
[
"Wu",
"Chen Henry",
""
],
[
"Ding",
"Charles",
""
],
[
"Raghunathan",
"Aditi",
""
]
] |
2504.15267 | Shaorong Zhang | Shaorong Zhang, Tamoghna Chattopadhyay, Sophia I. Thomopoulos,
Jose-Luis Ambite, Paul M. Thompson, Greg Ver Steeg | Diffusion Bridge Models for 3D Medical Image Translation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Diffusion tensor imaging (DTI) provides crucial insights into the
microstructure of the human brain, but it can be time-consuming to acquire
compared to more readily available T1-weighted (T1w) magnetic resonance imaging
(MRI). To address this challenge, we propose a diffusion bridge model for 3D
brain image translation between T1w MRI and DTI modalities. Our model learns to
generate high-quality DTI fractional anisotropy (FA) images from T1w images and
vice versa, enabling cross-modality data augmentation and reducing the need for
extensive DTI acquisition. We evaluate our approach using perceptual
similarity, pixel-level agreement, and distributional consistency metrics,
demonstrating strong performance in capturing anatomical structures and
preserving information on white matter integrity. The practical utility of the
synthetic data is validated through sex classification and Alzheimer's disease
classification tasks, where the generated images achieve comparable performance
to real data. Our diffusion bridge model offers a promising solution for
improving neuroimaging datasets and supporting clinical decision-making, with
the potential to significantly impact neuroimaging research and clinical
practice.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:49:06 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Zhang",
"Shaorong",
""
],
[
"Chattopadhyay",
"Tamoghna",
""
],
[
"Thomopoulos",
"Sophia I.",
""
],
[
"Ambite",
"Jose-Luis",
""
],
[
"Thompson",
"Paul M.",
""
],
[
"Steeg",
"Greg Ver",
""
]
] |
2504.15268 | J.D. Opdyke | JD Opdyke | Beating the Correlation Breakdown: Robust Inference, Flexible Scenarios,
and Stress Testing for Financial Portfolios | null | null | null | null | q-fin.RM q-fin.PM q-fin.ST stat.AP | http://creativecommons.org/licenses/by/4.0/ | We live in a multivariate world, and effective modeling of financial
portfolios, including their construction, allocation, forecasting, and risk
analysis, simply is not possible without explicitly modeling the dependence
structure of their assets. Dependence structure can drive portfolio results
more than many other parameters in investment and risk models, sometimes even
more than their combined effects, but the literature provides relatively little
to define the finite-sample distributions of dependence measures in useable and
useful ways under challenging, real-world financial data conditions. Yet this
is exactly what is needed to make valid inferences about their estimates, and
to use these inferences for a myriad of essential purposes, such as hypothesis
testing, dynamic monitoring, realistic and granular scenario and reverse
scenario analyses, and mitigating the effects of correlation breakdowns during
market upheavals (which is when we need valid inferences the most). This work
develops a new and straightforward method, Nonparametric Angles-based
Correlation (NAbC), for defining the finite-sample distributions of any
dependence measure whose matrix of pairwise associations is positive definite
(e.g. Pearsons, Kendalls Tau, Spearmans Rho, Chatterjees, Lancasters, Szekelys,
and their many variants). The solution remains valid under marginal asset
distributions characterized by notably different and varying degrees of serial
correlation, non-stationarity, heavy-tailedness, and asymmetry. Notably, NAbCs
p-values and confidence intervals remain analytically consistent at both the
matrix level and the pairwise cell level. Finally, NAbC maintains validity even
when selected cells in the matrix are frozen for a given scenario or stress
test, that is, unaffected by the scenario, thus enabling flexible, granular,
and realistic scenarios.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:52:36 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Opdyke",
"JD",
""
]
] |
2504.15269 | Changwoo Lee | Changwoo J. Lee, Benjamin K. Dahl, Otso Ovaskainen, David B. Dunson | Scalable and robust regression models for continuous proportional data | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Beta regression is used routinely for continuous proportional data, but it
often encounters practical issues such as a lack of robustness of regression
parameter estimates to misspecification of the beta distribution. We develop an
improved class of generalized linear models starting with the continuous
binomial (cobin) distribution and further extending to dispersion mixtures of
cobin distributions (micobin). The proposed cobin regression and micobin
regression models have attractive robustness, computation, and flexibility
properties. A key innovation is the Kolmogorov-Gamma data augmentation scheme,
which facilitates Gibbs sampling for Bayesian computation, including in
hierarchical cases involving nested, longitudinal, or spatial data. We
demonstrate robustness, ability to handle responses exactly at the boundary (0
or 1), and computational efficiency relative to beta regression in simulation
experiments and through analysis of the benthic macroinvertebrate multimetric
index of US lakes using lake watershed covariates.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:55:00 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Lee",
"Changwoo J.",
""
],
[
"Dahl",
"Benjamin K.",
""
],
[
"Ovaskainen",
"Otso",
""
],
[
"Dunson",
"David B.",
""
]
] |
2504.15270 | Ji Qi | Ji Qi, Yuan Yao, Yushi Bai, Bin Xu, Juanzi Li, Zhiyuan Liu, Tat-Seng
Chua | An LMM for Efficient Video Understanding via Reinforced Compression of
Video Cubes | null | null | null | null | cs.CV cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Multimodal Models (LMMs) uniformly perceive video frames, creating
computational inefficiency for videos with inherently varying temporal
information density. This paper present \textbf{Quicksviewer}, an LMM with new
perceiving paradigm that partitions a video of nonuniform density into varying
cubes using Gumbel Softmax, followed by a unified resampling for each cube to
achieve efficient video understanding. This simple and intuitive approach
dynamically compress video online based on its temporal density, significantly
reducing spatiotemporal redundancy (overall 45$\times$ compression rate), while
enabling efficient training with large receptive field. We train the model from
a language backbone through three progressive stages, each incorporating
lengthy videos on average of 420s/1fps thanks to the perceiving efficiency.
With only 0.8M total video-text samples for training, our model outperforms the
direct baseline employing a fixed partitioning strategy by a maximum of 8.72 in
accuracy, demonstrating the effectiveness in performance. On Video-MME,
Quicksviewer achieves SOTA under modest sequence lengths using just up to 5\%
of tokens per frame required by baselines. With this paradigm, scaling up the
number of input frames reveals a clear power law of the model capabilities. It
is also empirically verified that the segments generated by the cubing network
can help for analyzing continuous events in videos.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:57:21 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Qi",
"Ji",
""
],
[
"Yao",
"Yuan",
""
],
[
"Bai",
"Yushi",
""
],
[
"Xu",
"Bin",
""
],
[
"Li",
"Juanzi",
""
],
[
"Liu",
"Zhiyuan",
""
],
[
"Chua",
"Tat-Seng",
""
]
] |
2504.15271 | Chen Guo | Guo Chen, Zhiqi Li, Shihao Wang, Jindong Jiang, Yicheng Liu, Lidong
Lu, De-An Huang, Wonmin Byeon, Matthieu Le, Tuomas Rintamaki, Tyler Poon, Max
Ehrlich, Tuomas Rintamaki, Tyler Poon, Tong Lu, Limin Wang, Bryan Catanzaro,
Jan Kautz, Andrew Tao, Zhiding Yu, Guilin Liu | Eagle 2.5: Boosting Long-Context Post-Training for Frontier
Vision-Language Models | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | We introduce Eagle 2.5, a family of frontier vision-language models (VLMs)
for long-context multimodal learning. Our work addresses the challenges in long
video comprehension and high-resolution image understanding, introducing a
generalist framework for both tasks. The proposed training framework
incorporates Automatic Degrade Sampling and Image Area Preservation, two
techniques that preserve contextual integrity and visual details. The framework
also includes numerous efficiency optimizations in the pipeline for
long-context data training. Finally, we propose Eagle-Video-110K, a novel
dataset that integrates both story-level and clip-level annotations,
facilitating long-video understanding. Eagle 2.5 demonstrates substantial
improvements on long-context multimodal benchmarks, providing a robust solution
to the limitations of existing VLMs. Notably, our best model Eagle 2.5-8B
achieves 72.4% on Video-MME with 512 input frames, matching the results of
top-tier commercial model such as GPT-4o and large-scale open-source models
like Qwen2.5-VL-72B and InternVL2.5-78B.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:57:28 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Chen",
"Guo",
""
],
[
"Li",
"Zhiqi",
""
],
[
"Wang",
"Shihao",
""
],
[
"Jiang",
"Jindong",
""
],
[
"Liu",
"Yicheng",
""
],
[
"Lu",
"Lidong",
""
],
[
"Huang",
"De-An",
""
],
[
"Byeon",
"Wonmin",
""
],
[
"Le",
"Matthieu",
""
],
[
"Rintamaki",
"Tuomas",
""
],
[
"Poon",
"Tyler",
""
],
[
"Ehrlich",
"Max",
""
],
[
"Rintamaki",
"Tuomas",
""
],
[
"Poon",
"Tyler",
""
],
[
"Lu",
"Tong",
""
],
[
"Wang",
"Limin",
""
],
[
"Catanzaro",
"Bryan",
""
],
[
"Kautz",
"Jan",
""
],
[
"Tao",
"Andrew",
""
],
[
"Yu",
"Zhiding",
""
],
[
"Liu",
"Guilin",
""
]
] |
2504.15272 | Yuma Narita | Kohta Murase, Yuma Narita, and Wen Yin | Superheavy dark matter from the natural inflation in light of the
highest-energy astroparticle events | 36 pages, 5 figures | null | null | TU-1260 | hep-ph astro-ph.CO astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Superheavy dark matter has been attractive as a candidate of particle dark
matter. We propose a "natural" particle model, in which the dark matter serves
as the inflaton in natural inflation, while decaying to high-energy particles
at energies of $10^{9}-10^{13} \, \text{GeV}$ from the prediction of the
inflation. A scalar field responsible to dilute the dark matter abundance
revives the natural inflation. Since the dark matter must be a spin zero
scalar, we study carefully the galactic dark matter 3-body decay into fermions
and two body decays into a gluon pair, and point out relevant multi-messenger
bounds that constrain these decay modes. Interestingly, the predicted energy
scale may coincide with the AMATERASU event and/or the KM3NeT neutrino event,
KM3-230213A. We also point out particle models with dark baryon to further
alleviate $\gamma$-ray bounds. This scenario yields several testable
predictions for the UHECR observations, including the highest-energy neutrons
that are unaffected by magnetic fields, the tensor-to-scalar ratio, the running
of spectral indices, $\alpha_s\gtrsim\mathcal{O}(0.001)$, and the existence of
light new colored particles that could be accessible at future collider
experiments. Further measurements of high energy cosmic rays, including their
components and detailed coordinates may provide insight into not only the
origin of the cosmic rays but also inflation.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:57:50 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Murase",
"Kohta",
""
],
[
"Narita",
"Yuma",
""
],
[
"Yin",
"Wen",
""
]
] |
2504.15273 | Layla Parast | Rebecca Knowlton and Layla Parast | Efficient Testing Using Surrogate Information | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In modern clinical trials, there is immense pressure to use surrogate markers
in place of an expensive or long-term primary outcome to make more timely
decisions about treatment effectiveness. However, using a surrogate marker to
test for a treatment effect can be difficult and controversial. Existing
methods tend to either rely on fully parametric methods where strict
assumptions are made about the relationship between the surrogate and the
outcome, or assume the surrogate marker is valid for the entire study
population. In this paper, we develop a fully nonparametric method for
efficient testing using surrogate information (ETSI). Our approach is
specifically designed for settings where there is heterogeneity in the utility
of the surrogate marker, i.e., the surrogate is valid for certain patient
subgroups and not others. ETSI enables treatment effect estimation and
hypothesis testing via kernel-based estimation for a setting where the
surrogate is used in place of the primary outcome for individuals for whom the
surrogate is valid, and the primary outcome is purposefully only measured in
the remaining patients. In addition, we provide a framework for future study
design with power and sample size estimates based on our proposed testing
procedure. We demonstrate the performance of our methods via a simulation study
and application to two distinct HIV clinical trials.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 17:57:53 GMT"
}
] | 2025-04-22T00:00:00 | [
[
"Knowlton",
"Rebecca",
""
],
[
"Parast",
"Layla",
""
]
] |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.