id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.22401 | Richard Guest | Richard Guest, Eva Lievens, Martin Sas, Elena Botoeva, Temitope Adeyemo, Valerie Verdoodt, Elora Fernandes, Chris Allgrove | Facial Age Estimation: A Research Roadmap for Technological and Legal Development and Deployment | null | null | null | null | cs.CY | http://creativecommons.org/licenses/by-sa/4.0/ | Automated facial age assessment systems operate in either estimation mode - predicting age based on facial traits, or verification mode - confirming a claimed age. These systems support access control to age-restricted goods, services, and content, and can be used in areas like e-commerce, social media, forensics, and refugee support. They may also personalise services in healthcare, finance, and advertising. While improving technological accuracy is essential, deployment must consider legal, ethical, sociological, alongside technological factors. This white paper reviews the current challenges in deploying such systems, outlines the relevant legal and regulatory landscape, and explores future research for fair, robust, and ethical age estimation technologies. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:28:31 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Guest",
"Richard",
""
],
[
"Lievens",
"Eva",
""
],
[
"Sas",
"Martin",
""
],
[
"Botoeva",
"Elena",
""
],
[
"Adeyemo",
"Temitope",
""
],
[
"Verdoodt",
"Valerie",
""
],
[
"Fernandes",
"Elora",
""
],
[
"Allgrove",
"Chris",
""
]
] |
2505.22403 | Vladimir Shpilrain | Vladimir Shpilrain | Knot invariants from representations of braids by automorphisms of a free group | 11 pages. Comments are welcome | null | null | null | math.GT math.GR | http://creativecommons.org/licenses/by/4.0/ | We describe an alternative way of computing Alexander polynomials of knots/links, based on the Artin representation of the corresponding braids by automorphisms of a free group. Then we apply the same method to other representations of braid groups discovered by Wada and compare the corresponding isotopic invariants to Alexander polynomials. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:29:28 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Shpilrain",
"Vladimir",
""
]
] |
2505.22404 | Stef Cuyckens | Stef Cuyckens, Xiaoling Yi, Nitish Satya Murthy, Chao Fang, Marian Verhelst | Efficient Precision-Scalable Hardware for Microscaling (MX) Processing in Robotics Learning | To appear in 2025 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED 2025) | null | null | null | cs.AR cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Autonomous robots require efficient on-device learning to adapt to new environments without cloud dependency. For this edge training, Microscaling (MX) data types offer a promising solution by combining integer and floating-point representations with shared exponents, reducing energy consumption while maintaining accuracy. However, the state-of-the-art continuous learning processor, namely Dacapo, faces limitations with its MXINT-only support and inefficient vector-based grouping during backpropagation. In this paper, we present, to the best of our knowledge, the first work that addresses these limitations with two key innovations: (1) a precision-scalable arithmetic unit that supports all six MX data types by exploiting sub-word parallelism and unified integer and floating-point processing; and (2) support for square shared exponent groups to enable efficient weight handling during backpropagation, removing storage redundancy and quantization overhead. We evaluate our design against Dacapo under iso-peak-throughput on four robotics workloads in TSMC 16nm FinFET technology at 500MHz, reaching a 25.6% area reduction, a 51% lower memory footprint, and 4x higher effective training throughput while achieving comparable energy-efficiency, enabling efficient robotics continual learning at the edge. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:34:32 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Cuyckens",
"Stef",
""
],
[
"Yi",
"Xiaoling",
""
],
[
"Murthy",
"Nitish Satya",
""
],
[
"Fang",
"Chao",
""
],
[
"Verhelst",
"Marian",
""
]
] |
2505.22405 | Jevg\=enijs Vihrovs | Jevg\=enijs Vihrovs | Quantum Search on Computation Trees | null | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show a simple generalization of the quantum walk algorithm for search in backtracking trees by Montanaro (ToC 2018) to the case where vertices can have different times of computation. If a vertex $v$ in the tree of depth $D$ is computed in $t_v$ steps from its parent, then we show that detection of a marked vertex requires $\text{O}(\sqrt{TD})$ queries to the steps of the computing procedures, where $T = \sum_v t_v^2$. This framework provides an easy and convenient way to re-obtain a number of other quantum frameworks like variable time search, quantum divide & conquer and bomb query algorithms. The underlying algorithm is simple, explicitly constructed, and has low poly-logarithmic factors in the complexity.
As a corollary, this gives a quantum algorithm for variable time search with unknown times with optimal query complexity $\text{O}(\sqrt{T \log \min(n,t_{\max})})$, where $T = \sum_i t_i^2$ and $t_{\max} = \max_i t_i$ if $t_i$ is the number of steps required to compute the $i$-th variable. This resolves the open question of the query complexity of variable time search, as the matching lower bound was recently shown by Ambainis, Kokainis and Vihrovs (TQC'23). As another result, we obtain an $\widetilde{\text{O}}(n)$ time algorithm for the geometric task of determining if any three lines among $n$ given intersect at the same point, improving the $\text{O}(n^{1+\text{o}(1)})$ algorithm of Ambainis and Larka (TQC'20). | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:35:18 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Vihrovs",
"Jevgēnijs",
""
]
] |
2505.22406 | Miguel Dovale \'Alvarez | Gerhard Heinzel, Javier \'Alvarez-Vizoso, Miguel Dovale-\'Alvarez | Complete Catalog of Laser Locking Configurations for LISA | 8 pages, 4 figures | null | null | null | physics.ins-det astro-ph.IM gr-qc physics.space-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Laser Interferometer Space Antenna (LISA) will enable direct observations of low-frequency gravitational waves, offering unprecedented insight into astrophysical and cosmological phenomena. LISA's heterodyne interferometric measurement system requires phase-locking five of its six onboard lasers with tunable frequency offsets to ensure that all beatnotes remain within the metrology system's operational range, despite Doppler-induced frequency shifts. The selection of these offset frequencies -- collectively forming a frequency plan -- is a complex optimization problem constrained by the spacecraft's orbital dynamics and instrument limitations. While previous work established an algorithmic solution for deriving time-dependent frequency plans, this study takes a complementary approach by systematically analyzing and cataloging all possible laser locking configurations. We present an automated method to explore, validate, and classify viable locking schemes, identifying 36 unique non-frequency-swapping configurations and 72 additional frequency-swapping configurations for an arbitrary choice of primary laser. This exhaustive classification provides a foundation for frequency planning across the full range of operational scenarios. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:37:13 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Heinzel",
"Gerhard",
""
],
[
"Álvarez-Vizoso",
"Javier",
""
],
[
"Dovale-Álvarez",
"Miguel",
""
]
] |
2505.22407 | Jiadong Pan | Jiadong Pan, Zhiyuan Ma, Kaiyan Zhang, Ning Ding, Bowen Zhou | Self-Reflective Reinforcement Learning for Diffusion-based Image Reasoning Generation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion models have recently demonstrated exceptional performance in image generation task. However, existing image generation methods still significantly suffer from the dilemma of image reasoning, especially in logic-centered image generation tasks. Inspired by the success of Chain of Thought (CoT) and Reinforcement Learning (RL) in LLMs, we propose SRRL, a self-reflective RL algorithm for diffusion models to achieve reasoning generation of logical images by performing reflection and iteration across generation trajectories. The intermediate samples in the denoising process carry noise, making accurate reward evaluation difficult. To address this challenge, SRRL treats the entire denoising trajectory as a CoT step with multi-round reflective denoising process and introduces condition guided forward process, which allows for reflective iteration between CoT steps. Through SRRL-based iterative diffusion training, we introduce image reasoning through CoT into generation tasks adhering to physical laws and unconventional physical phenomena for the first time. Notably, experimental results of case study exhibit that the superior performance of our SRRL algorithm even compared with GPT-4o. The project page is https://jadenpan0.github.io/srrl.github.io/. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:37:21 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Pan",
"Jiadong",
""
],
[
"Ma",
"Zhiyuan",
""
],
[
"Zhang",
"Kaiyan",
""
],
[
"Ding",
"Ning",
""
],
[
"Zhou",
"Bowen",
""
]
] |
2505.22408 | Victor Enescu | Victor Enescu, Hichem Sahbi | Frugal Incremental Generative Modeling using Variational Autoencoders | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Continual or incremental learning holds tremendous potential in deep learning with different challenges including catastrophic forgetting. The advent of powerful foundation and generative models has propelled this paradigm even further, making it one of the most viable solution to train these models. However, one of the persisting issues lies in the increasing volume of data particularly with replay-based methods. This growth introduces challenges with scalability since continuously expanding data becomes increasingly demanding as the number of tasks grows. In this paper, we attenuate this issue by devising a novel replay-free incremental learning model based on Variational Autoencoders (VAEs). The main contribution of this work includes (i) a novel incremental generative modelling, built upon a well designed multi-modal latent space, and also (ii) an orthogonality criterion that mitigates catastrophic forgetting of the learned VAEs. The proposed method considers two variants of these VAEs: static and dynamic with no (or at most a controlled) growth in the number of parameters. Extensive experiments show that our method is (at least) an order of magnitude more ``memory-frugal'' compared to the closely related works while achieving SOTA accuracy scores. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:37:57 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Enescu",
"Victor",
""
],
[
"Sahbi",
"Hichem",
""
]
] |
2505.22409 | Jiaming Li | Shuai Peng, Sijia Peng, Lijun Ren, Shaokun Liu, Bin Liu, Jiaming Li, and Le Luo | Precision Measurement of Spin-Dependent Dipolar Splitting in $^6$Li p-Wave Feshbach Resonances | null | null | null | null | cond-mat.quant-gas | http://creativecommons.org/licenses/by/4.0/ | The magnetic dipolar splitting of a p-wave Feshbach resonance is governed by the spin-orbital configuration of the valence electrons in the triplet molecular state. We perform high-resolution trap loss spectroscopy on ultracold 6Li atoms to resolve this splitting with sub-milligauss precision. By comparing spin-polarized (|mS| = 1) and spin-mixture (mS = 0) configurations of the triplet state, we observe a clear spin-dependent reversal in the splitting structure, confirmed via momentumresolved absorption imaging. This behavior directly reflects the interplay between electron spin projection mS and orbital angular momentum ml in the molecular states. Our results provide a stringent benchmark for dipole-dipole interaction models and lay the groundwork for controlling the pairing in p-wave superfluid systems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:38:21 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Peng",
"Shuai",
""
],
[
"Peng",
"Sijia",
""
],
[
"Ren",
"Lijun",
""
],
[
"Liu",
"Shaokun",
""
],
[
"Liu",
"Bin",
""
],
[
"Li",
"Jiaming",
""
],
[
"Luo",
"Le",
""
]
] |
2505.22410 | Cornelius Brand | Cornelius Brand, Radu Curticapean, Baitian Li, Kevin Pratt | Faster Convolutions: Yates and Strassen Revisited | null | null | null | null | cs.DS cs.CC | http://creativecommons.org/licenses/by/4.0/ | Given two vectors $u,v \in \mathbb{Q}^D$ over a finite domain $D$ and a function $f : D\times D\to D$, the convolution problem asks to compute the vector $w \in \mathbb{Q}^D$ whose entries are defined by $w(d) = \sum_{\substack{x,y \in D \\ f(x,y)=d}} u(x)v(y).$ In parameterized and exponential-time algorithms, convolutions on product domains are particularly prominent: Here, a finite domain $B$ and a function $h : B \times B \to B$ are fixed, and convolution is done over the product domain $D = B^k$, using the function $h^k :D \times D\to D$ that applies $h$ coordinate-wise to its input tuples.
We present a new perspective on product-domain convolutions through multilinear algebra. This viewpoint streamlines the presentation and analysis of existing algorithms, such as those by van Rooij et al. (ESA 2009). Moreover, using established results from the theory of fast matrix multiplication, we derive improved $O^\ast(|B|^{2\omega/3 \cdot k}) = O(|D|^{1.582})$ time algorithms, improving upon previous upper bounds by Esmer et al. (Algorithmica 86(1), 2024) of the form $c^k |B|^{2k}$ for $c < 1$. Using the setup described in this note, Strassen's asymptotic rank conjecture from algebraic complexity theory would imply quasi-linear $|D|^{1+o(1)}$ time algorithms. This conjecture has recently gained attention in the algorithms community. (Bj\"orklund-Kaski and Pratt, STOC 2024, Bj\"orklund et al., SODA 2025)
Our paper is intended as a self-contained exposition for an algorithms audience, and it includes all essential mathematical prerequisites with explicit coordinate-based notation. In particular, we assume no knowledge in abstract algebra. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:38:22 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Brand",
"Cornelius",
""
],
[
"Curticapean",
"Radu",
""
],
[
"Li",
"Baitian",
""
],
[
"Pratt",
"Kevin",
""
]
] |
2505.22411 | Yao Huang | Yao Huang, Huanran Chen, Shouwei Ruan, Yichi Zhang, Xingxing Wei, Yinpeng Dong | Mitigating Overthinking in Large Reasoning Models via Manifold Steering | 17 pages, 7 figures | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recent advances in Large Reasoning Models (LRMs) have demonstrated remarkable capabilities in solving complex tasks such as mathematics and coding. However, these models frequently exhibit a phenomenon known as overthinking during inference, characterized by excessive validation loops and redundant deliberation, leading to substantial computational overheads. In this paper, we aim to mitigate overthinking by investigating the underlying mechanisms from the perspective of mechanistic interpretability. We first showcase that the tendency of overthinking can be effectively captured by a single direction in the model's activation space and the issue can be eased by intervening the activations along this direction. However, this efficacy soon reaches a plateau and even deteriorates as the intervention strength increases. We therefore systematically explore the activation space and find that the overthinking phenomenon is actually tied to a low-dimensional manifold, which indicates that the limited effect stems from the noises introduced by the high-dimensional steering direction. Based on this insight, we propose Manifold Steering, a novel approach that elegantly projects the steering direction onto the low-dimensional activation manifold given the theoretical approximation of the interference noise. Extensive experiments on DeepSeek-R1 distilled models validate that our method reduces output tokens by up to 71% while maintaining and even improving the accuracy on several mathematical benchmarks. Our method also exhibits robust cross-domain transferability, delivering consistent token reduction performance in code generation and knowledge-based QA tasks. Code is available at: https://github.com/Aries-iai/Manifold_Steering. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:39:26 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Huang",
"Yao",
""
],
[
"Chen",
"Huanran",
""
],
[
"Ruan",
"Shouwei",
""
],
[
"Zhang",
"Yichi",
""
],
[
"Wei",
"Xingxing",
""
],
[
"Dong",
"Yinpeng",
""
]
] |
2505.22412 | Yu-Ting Rong | Yu-Ting Rong, Ping-Mo Liu, Dan Yang, Rong An | Neutron Magic Numbers in $sd$ Shell from Nuclear Charge Radii within Neutron-Proton Correction around the Fermi Surface | 10 pages, 4 figures | null | null | null | nucl-th | http://creativecommons.org/licenses/by/4.0/ | Charge radii are sensitive indicators to identify the nuclear structure phenomena throughout the whole nuclide chart. In particular, the shrunken trend of changes of charge radii along a long isotopic chain is intimately associated with the shell quenching effect. In this work, the systematic evolution of charge radii along the proton numbers $Z=8$, $10$, $12$, $14$, $18$ isotopes is investigated by a relativistic Hartree Bogoliubov model. A ansatz about neutron-proton correlation around Fermi surface is considered for describing the abnormal behavior of nuclear charge radii. Our results show that the neutron-proton pairing corrections around the Fermi surface lead to a sudden strengthening of the charge radii of these isotopic chains at $N=8$, 20 and 28, reflecting the fact that this correction enhances the shell closure across $N=8$, 20 and 28. The reproduction of the $N=14$ charge radius in the Mg isotopes is affected by the way in which pairing correlations are handled, with BCS theory overestimating the shell effect of $N=14$, and the Bogoliubov quasiparticle transformation suggests a stronger pairing correlation near the proton Fermi surface, which is more consistent with experimental results. An analysis of the deviations from the theoretical and available experimental data for the charge radii of the 24 selected even-even nuclei shows that the neutron-proton pairing correction around the Fermi surface has an improved effect on the calculation of the charge {radii} using the meson-exchange effective interactions, but it does not help to significantly improve the results calculated by the density-dependent effective interactions. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:39:48 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Rong",
"Yu-Ting",
""
],
[
"Liu",
"Ping-Mo",
""
],
[
"Yang",
"Dan",
""
],
[
"An",
"Rong",
""
]
] |
2505.22413 | Nicola Pinamonti | Romeo Brunetti, Klaus Fredenhagen and Nicola Pinamonti | Thermodynamical aspects of fermions in external electromagnetic fields | 30 pages | null | null | null | math-ph hep-th math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The thermodynamics of Dirac fields under the influence of external electromagnetic fields is studied. For perturbations which act only for finite time, the influence of the perturbation can be described by an automorphism which can be unitarily implemented in the GNS representations of KMS states, a result long known for the Fock representation. For time-independent perturbations, however, the time evolution cannot be implemented in typical cases, so the standard methods of quantum statistical mechanics do not apply. Instead we show that a smooth switching on of the external potential allows a comparison of the free and the perturbed time evolution, and approach to equilibrium, a possible existence of non-equilibrium stationary states (NESS) and Araki's relative entropy can be investigated. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:40:50 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Brunetti",
"Romeo",
""
],
[
"Fredenhagen",
"Klaus",
""
],
[
"Pinamonti",
"Nicola",
""
]
] |
2505.22414 | Md Ehtesham-Ul-Haque | Md Ehtesham-Ul-Haque, Syed Masum Billah | ToPSen: Task-Oriented Priming and Sensory Alignment for Comparing Coding Strategies Between Sighted and Blind Programmers | Accepted at DIS'25 | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper examines how the coding strategies of sighted and blind programmers differ when working with audio feedback alone. The goal is to identify challenges in mixed-ability collaboration, particularly when sighted programmers work with blind peers or teach programming to blind students. To overcome limitations of traditional blindness simulation studies, we proposed Task-Oriented Priming and Sensory Alignment (ToPSen), a design framework that reframes sensory constraints as technical requirements rather than as a disability. Through a study of 12 blind and 12 sighted participants coding non-visually, we found that expert blind programmers maintain more accurate mental models and process more information in working memory than sighted programmers using ToPSen. Our analysis revealed that blind and sighted programmers process structural information differently, exposing gaps in current IDE designs. These insights inform our guidelines for improving the accessibility of programming tools and fostering effective mixed-ability collaboration. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:41:29 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Ehtesham-Ul-Haque",
"Md",
""
],
[
"Billah",
"Syed Masum",
""
]
] |
2505.22415 | Jian Bin Ben Chen | J. B. B. Chen E. Adli, P. Drobniak, O. G. Finnerud, E. H{\o}rlyk, D. Kalvik, C. A. Lindstr{\o}m, F. Pe\~na, K. Sjobak | ABEL: The Adaptable Beginning-to-End Linac simulation framework | null | null | null | null | physics.acc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce ABEL, the Adaptable Beginning-to-End Linac simulation framework developed for agile design studies of plasma-based accelerators and colliders. ABEL's modular architecture allows users to simulate particle acceleration across various beamline components. The framework supports specialised codes such as HiPACE++, Wake-T, ELEGANT, GUINEA-PIG, CLICopti and ImpactX, which facilitate precise modelling of complex machine components. Key features include simplified models for addressing transverse instabilities, radiation reactions, and ion motion, alongside comprehensive diagnostics and optimisation capabilities. Our simulation studies focus on the HALHF plasma linac, examining tolerances for drive beam jitter, including effects of self-correction mechanisms. Simulation results demonstrate ABEL's ability to model emittance growth due to transverse instability and ion motion, highlighting the framework's adaptability in balancing simulation fidelity with computational efficiency. The findings point towards ABEL's potential for advancing compact accelerator designs and contribute to the broader goals of enhancing control and precision in plasma-based acceleration. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:42:25 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Adli",
"J. B. B. Chen E.",
""
],
[
"Drobniak",
"P.",
""
],
[
"Finnerud",
"O. G.",
""
],
[
"Hørlyk",
"E.",
""
],
[
"Kalvik",
"D.",
""
],
[
"Lindstrøm",
"C. A.",
""
],
[
"Peña",
"F.",
""
],
[
"Sjobak",
"K.",
""
]
] |
2505.22416 | Sihun Cha | Sihun Cha, Serin Yoon, Kwanggyoon Seo, Junyong Noh | Neural Face Skinning for Mesh-agnostic Facial Expression Cloning | null | null | null | null | cs.GR cs.CV | http://creativecommons.org/licenses/by/4.0/ | Accurately retargeting facial expressions to a face mesh while enabling manipulation is a key challenge in facial animation retargeting. Recent deep-learning methods address this by encoding facial expressions into a global latent code, but they often fail to capture fine-grained details in local regions. While some methods improve local accuracy by transferring deformations locally, this often complicates overall control of the facial expression. To address this, we propose a method that combines the strengths of both global and local deformation models. Our approach enables intuitive control and detailed expression cloning across diverse face meshes, regardless of their underlying structures. The core idea is to localize the influence of the global latent code on the target mesh. Our model learns to predict skinning weights for each vertex of the target face mesh through indirect supervision from predefined segmentation labels. These predicted weights localize the global latent code, enabling precise and region-specific deformations even for meshes with unseen shapes. We supervise the latent code using Facial Action Coding System (FACS)-based blendshapes to ensure interpretability and allow straightforward editing of the generated animation. Through extensive experiments, we demonstrate improved performance over state-of-the-art methods in terms of expression fidelity, deformation transfer accuracy, and adaptability across diverse mesh structures. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:43:43 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Cha",
"Sihun",
""
],
[
"Yoon",
"Serin",
""
],
[
"Seo",
"Kwanggyoon",
""
],
[
"Noh",
"Junyong",
""
]
] |
2505.22417 | Bin Wu | Xinbing Kong, Bin Wu, Wuyi Ye | High-Dimensional Binary Variates: Maximum Likelihood Estimation with Nonstationary Covariates and Factors | null | null | null | null | math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces a high-dimensional binary variate model that accommodates nonstationary covariates and factors, and studies their asymptotic theory. This framework encompasses scenarios where single indices are nonstationary or cointegrated. For nonstationary single indices, the maximum likelihood estimator (MLE) of the coefficients has dual convergence rates and is collectively consistent under the condition $T^{1/2}/N\to0$, as both the cross-sectional dimension $N$ and the time horizon $T$ approach infinity. The MLE of all nonstationary factors is consistent when $T^{\delta}/N\to0$, where $\delta$ depends on the link function. The limiting distributions of the factors depend on time $t$, governed by the convergence of the Hessian matrix to zero. In the case of cointegrated single indices, the MLEs of both factors and coefficients converge at a higher rate of $\min(\sqrt{N},\sqrt{T})$. A distinct feature compared to nonstationary single indices is that the dual rate of convergence of the coefficients increases from $(T^{1/4},T^{3/4})$ to $(T^{1/2},T)$. Moreover, the limiting distributions of the factors do not depend on $t$ in the cointegrated case. Monte Carlo simulations verify the accuracy of the estimates. In an empirical application, we analyze jump arrivals in financial markets using this model, extract jump arrival factors, and demonstrate their efficacy in large-cross-section asset pricing. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:43:44 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Kong",
"Xinbing",
""
],
[
"Wu",
"Bin",
""
],
[
"Ye",
"Wuyi",
""
]
] |
2505.22418 | Jie Cai | Jeongwon Jo, He Zhang, Jie Cai, Nitesh Goyal | AI Trust Reshaping Administrative Burdens: Understanding Trust-Burden Dynamics in LLM-Assisted Benefits Systems | FAccT 2025 | null | 10.1145/3715275.3732077 | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | Supplemental Nutrition Assistance Program (SNAP) is an essential benefit support system provided by the US administration to 41 million federally determined low-income applicants. Through interviews with such applicants across a diverse set of experiences with the SNAP system, our findings reveal that new AI technologies like LLMs can alleviate traditional burdens but also introduce new burdens. We introduce new types of learning, compliance, and psychological costs that transform the administrative burden on applicants. We also identify how trust in AI across three dimensions--competence, integrity, and benevolence--is perceived to reduce administrative burdens, which may stem from unintended and untoward overt trust in the system. We discuss calibrating appropriate levels of user trust in LLM-based administrative systems, mitigating newly introduced burdens. In particular, our findings suggest that evidence-based information disclosure is necessary in benefits administration and propose directions for future research on trust-burden dynamics in AI-assisted administration systems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:43:55 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jo",
"Jeongwon",
""
],
[
"Zhang",
"He",
""
],
[
"Cai",
"Jie",
""
],
[
"Goyal",
"Nitesh",
""
]
] |
2505.22419 | Ezra Alexander | Ezra Alexander, Alexandra Alexiu, Matthias Kick, and Troy Van Voorhis | Structural Hole Traps in III-V Quantum Dots | 21 pages, 5 figures | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Non-toxic III-V quantum dots (QDs) are plagued with a higher density of performance-limiting trap states than II-VI and IV-VI QDs. Such trap states are generally understood to arise from under-coordinated atoms on the QD surface. Here, we present computational evidence for, and an exploration of, trap states in InP and GaP QDs that arise from fully-coordinated atoms with distorted geometries, denoted here as structural traps. In particular, we focus on the properties of anion-centered hole traps, which we show to be relatively insensitive to the choice of the (typically cation-coordinating) ligand. Through interpolation of trap center cutouts, we arrive at a simple molecular orbital (MO) argument for the existence of structural traps, finding two main modalities: bond stretches and angular distortion to a see-saw-like geometry. These structural trap states will be important for understanding the low performance of III-V QDs, as even core-shell passivation may not remove these defects unless they can rigidify the structure. Moreover, they may lead to interesting dynamical properties as distorted structures could form transiently. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:44:42 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Alexander",
"Ezra",
""
],
[
"Alexiu",
"Alexandra",
""
],
[
"Kick",
"Matthias",
""
],
[
"Van Voorhis",
"Troy",
""
]
] |
2505.22420 | Alexander Sattler | Alexander Sattler and Maria Daghofer | Robustness of topological edge states in open alternating spin chains | null | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | Both the Haldane spin-$1$ chain and dimerized chains of
spin-$1/2$ exhibit topologically protected edge states that
are robust against specific perturbations. Recently, such spin
chains have been specifically assembled on surfaces and we
investigate here the robustness of these edge states against
coupling to the surface. Since no physical system can be considered perfectly isolated, it
is crucial to examine whether topological robustness is maintained in the
presence of environmental coupling.
We apply exact diagonalization to a Lindblad master equation that couples an
alternating Heisenberg spin chain based on spins $1/2$ to a surface via
various jump operators. The robustness of topological states is assessed via the
time evolution of quantities such as the ground-state degeneracy, correlation
function, entropy, and magnetization of edge states.
We investigate chains built from dimers with antiferromagnetic and
ferromagnetic intra-dimer coupling, which resemble
Su-Schrieffer-Heeger and the Haldane models, resp., and assess the
impact of $z$-axis anisotropy and longer-ranged couplings. Generally,
we find that signatures of topological properties are
more robust in Su-Schrieffer-Heeger-like chains than in
Haldane-like chains. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:45:16 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Sattler",
"Alexander",
""
],
[
"Daghofer",
"Maria",
""
]
] |
2505.22422 | Vaclav Voracek | V\'aclav Vor\'a\v{c}ek and Francesco Orabona | STaR-Bets: Sequential Target-Recalculating Bets for Tighter Confidence Intervals | comments are welcome | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The construction of confidence intervals for the mean of a bounded random variable is a classical problem in statistics with numerous applications in machine learning and virtually all scientific fields. In particular, obtaining the tightest possible confidence intervals is vital every time the sampling of the random variables is expensive. The current state-of-the-art method to construct confidence intervals is by using betting algorithms. This is a very successful approach for deriving optimal confidence sequences, even matching the rate of law of iterated logarithms. However, in the fixed horizon setting, these approaches are either sub-optimal or based on heuristic solutions with strong empirical performance but without a finite-time guarantee. Hence, no betting-based algorithm guaranteeing the optimal $\mathcal{O}(\sqrt{\frac{\sigma^2\log\frac1\delta}{n}})$ width of the confidence intervals are known. This work bridges this gap. We propose a betting-based algorithm to compute confidence intervals that empirically outperforms the competitors. Our betting strategy uses the optimal strategy in every step (in a certain sense), whereas the standard betting methods choose a constant strategy in advance. Leveraging this fact results in strict improvements even for classical concentration inequalities, such as the ones of Hoeffding or Bernstein. Moreover, we also prove that the width of our confidence intervals is optimal up to an $1+o(1)$ factor diminishing with $n$. The code is available on~https://github.com/vvoracek/STaR-bets-confidence-interval. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:48:07 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Voráček",
"Václav",
""
],
[
"Orabona",
"Francesco",
""
]
] |
2505.22423 | Jonathan Hill | Jonathan B. Hill | Max-laws of large numbers for weakly dependent high dimensional arrays with applications | null | null | null | null | math.ST stat.TH | http://creativecommons.org/licenses/by/4.0/ | We derive so-called weak and strong \textit{max-laws of large numbers} for $% \max_{1\leq i\leq k_{n}}|1/n\sum_{t=1}^{n}x_{i,n,t}|$ for zero mean stochastic triangular arrays $\{x_{i,n,t}$ $:$ $1$ $\leq $ $t$ $\leq n\}_{n\geq 1}$, with dimension counter $i$ $=$ $1,...,k_{n}$ and dimension $% k_{n}$ $\rightarrow $ $\infty $. Rates of convergence are also analyzed based on feasible sequences $\{k_{n}\}$. We work in three dependence settings: independence, Dedecker and Prieur's (2004) $\tau $-mixing and Wu's (2005) physical dependence. We initially ignore cross-coordinate $i$ dependence as a benchmark. We then work with martingale, nearly martingale, and mixing coordinates to deliver improved bounds on $k_{n}$. Finally, we use the results in three applications, each representing a key novelty: we ($i$) bound $k_{n}$\ for a max-correlation statistic for regression residuals under $\alpha $-mixing or physical dependence; ($ii$) extend correlation screening, or marginal regressions, to physical dependent data with diverging dimension $k_{n}$ $\rightarrow $ $\infty $; and ($iii$) test a high dimensional parameter after partialling out a fixed dimensional nuisance parameter in a linear time series regression model under $\tau $% -mixing. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:50:36 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hill",
"Jonathan B.",
""
]
] |
2505.22424 | Jianxiong Guo | Jingxi Lu, Wenhao Li, Jianxiong Guo, Xingjian Ding, Zhiqing Tang, Tian Wang, Weijia Jia | Hybrid Learning for Cold-Start-Aware Microservice Scheduling in Dynamic Edge Environments | null | null | null | null | cs.NI eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid growth of IoT devices and their diverse workloads, container-based microservices deployed at edge nodes have become a lightweight and scalable solution. However, existing microservice scheduling algorithms often assume static resource availability, which is unrealistic when multiple containers are assigned to an edge node. Besides, containers suffer from cold-start inefficiencies during early-stage training in currently popular reinforcement learning (RL) algorithms. In this paper, we propose a hybrid learning framework that combines offline imitation learning (IL) with online Soft Actor-Critic (SAC) optimization to enable a cold-start-aware microservice scheduling with dynamic allocation for computing resources. We first formulate a delay-and-energy-aware scheduling problem and construct a rule-based expert to generate demonstration data for behavior cloning. Then, a GRU-enhanced policy network is designed in the policy network to extract the correlation among multiple decisions by separately encoding slow-evolving node states and fast-changing microservice features, and an action selection mechanism is given to speed up the convergence. Extensive experiments show that our method significantly accelerates convergence and achieves superior final performance. Compared with baselines, our algorithm improves the total objective by $50\%$ and convergence speed by $70\%$, and demonstrates the highest stability and robustness across various edge configurations. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:51:57 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Lu",
"Jingxi",
""
],
[
"Li",
"Wenhao",
""
],
[
"Guo",
"Jianxiong",
""
],
[
"Ding",
"Xingjian",
""
],
[
"Tang",
"Zhiqing",
""
],
[
"Wang",
"Tian",
""
],
[
"Jia",
"Weijia",
""
]
] |
2505.22425 | Xueliang Zhao | Xueliang Zhao, Wei Wu, Lingpeng Kong | Scaling Reasoning without Attention | preprint | null | null | null | cs.LG cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have made significant advances in complex reasoning tasks, yet they remain bottlenecked by two core challenges: architectural inefficiency due to reliance on Transformers, and a lack of structured fine-tuning for high-difficulty domains. We introduce \ourmodel, an attention-free language model that addresses both issues through architectural and data-centric innovations. Built on the state space dual (SSD) layers of Mamba-2, our model eliminates the need for self-attention and key-value caching, enabling fixed-memory, constant-time inference. To train it for complex reasoning, we propose a two-phase curriculum fine-tuning strategy based on the \textsc{PromptCoT} synthesis paradigm, which generates pedagogically structured problems via abstract concept selection and rationale-guided generation. On benchmark evaluations, \ourmodel-7B outperforms strong Transformer and hybrid models of comparable scale, and even surpasses the much larger Gemma3-27B by 2.6\% on AIME 24, 0.6\% on AIME 25, and 3.0\% on Livecodebench. These results highlight the potential of state space models as efficient and scalable alternatives to attention-based architectures for high-capacity reasoning. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:52:15 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhao",
"Xueliang",
""
],
[
"Wu",
"Wei",
""
],
[
"Kong",
"Lingpeng",
""
]
] |
2505.22426 | Sandra Zavatarelli | R.S.Sidhu, F.Casaburo, E.Masha, M.Aliotta, C.Ananna, L.Barbieri, F.Barile, C.Baron, D.Bemmerer, A.Best, R.Biassisi, A.Boeltzig, R.Bonnel, C.Broggini, C.G.Bruno, A.Caciolli, M.Campostrini, F.Cavanna, T.Chillery, G.F.Ciani, P.Colombetti, A.Compagnucci, P.Corvisiero, L.Csedreki, T.Davinson, R.Depalo, A.Di Leva, A.D'Ottavi, Z.Elekes, F.Ferraro, A.Formicola, Zs.Fulop, G.Gervino, R.Gesue, G.Gosta, A.Guglielmetti, C.Gustavino, Gy.Gyurky, G.Imbriani, J.Jones, J.Jose, M.Junker, M.Lugaro, P.Marigo, J.Marsh, R.Menegazzo, D.Mercogliano, D.Piatti, P.Prati, D.Rapagnani, V.Rigato, D.Robb, M.Rossi, R.Sariyal, J.Skowronski, O.Straniero, T.Szucs, S.Turkat, M.Vagnoni, S.Zavatarelli | Improved direct measurement of low-energy resonances in the $^{21}$Ne(p,$\gamma$)$^{22}$Na reaction | null | null | null | null | nucl-ex hep-ex | http://creativecommons.org/licenses/by/4.0/ | In the nova temperature range, 0.1 GK $< T <$ 0.4 GK, several low-energy resonances dominate the $^{21}$Ne(p,$\gamma$)$^{22}$Na reaction rate, which is currently affected by large uncertainties. We present a high-precision study of the resonances at $E^{\rm{lab}}_{\rm{r}}$ = 127.3, 271.4, 272.3, 291.5, and 352.6 keV, measured directly at the Laboratory for Underground Nuclear Astrophysics in Italy. The strengths of the 127.3, 271.4, and 291.5 keV resonances are consistent with previous measurements within 1$\sigma$. However, for the 272.3 keV and 352.6 keV resonances, we report strength values of (129.9 $\pm$ 5.8) meV and (14.9 $\pm$ 0.8) meV, respectively, more than a factor of 1.5 higher than literature values. In addition, we report on new branching ratios for the 127.3, 272.3, and 352.6 keV resonances, leading to updated decay schemes. Finally, we present a revised thermonuclear reaction rate and investigate its impact on the NeNa nucleosynthesis. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:52:27 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Sidhu",
"R. S.",
""
],
[
"Casaburo",
"F.",
""
],
[
"Masha",
"E.",
""
],
[
"Aliotta",
"M.",
""
],
[
"Ananna",
"C.",
""
],
[
"Barbieri",
"L.",
""
],
[
"Barile",
"F.",
""
],
[
"Baron",
"C.",
""
],
[
"Bemmerer",
"D.",
""
],
[
"Best",
"A.",
""
],
[
"Biassisi",
"R.",
""
],
[
"Boeltzig",
"A.",
""
],
[
"Bonnel",
"R.",
""
],
[
"Broggini",
"C.",
""
],
[
"Bruno",
"C. G.",
""
],
[
"Caciolli",
"A.",
""
],
[
"Campostrini",
"M.",
""
],
[
"Cavanna",
"F.",
""
],
[
"Chillery",
"T.",
""
],
[
"Ciani",
"G. F.",
""
],
[
"Colombetti",
"P.",
""
],
[
"Compagnucci",
"A.",
""
],
[
"Corvisiero",
"P.",
""
],
[
"Csedreki",
"L.",
""
],
[
"Davinson",
"T.",
""
],
[
"Depalo",
"R.",
""
],
[
"Di Leva",
"A.",
""
],
[
"D'Ottavi",
"A.",
""
],
[
"Elekes",
"Z.",
""
],
[
"Ferraro",
"F.",
""
],
[
"Formicola",
"A.",
""
],
[
"Fulop",
"Zs.",
""
],
[
"Gervino",
"G.",
""
],
[
"Gesue",
"R.",
""
],
[
"Gosta",
"G.",
""
],
[
"Guglielmetti",
"A.",
""
],
[
"Gustavino",
"C.",
""
],
[
"Gyurky",
"Gy.",
""
],
[
"Imbriani",
"G.",
""
],
[
"Jones",
"J.",
""
],
[
"Jose",
"J.",
""
],
[
"Junker",
"M.",
""
],
[
"Lugaro",
"M.",
""
],
[
"Marigo",
"P.",
""
],
[
"Marsh",
"J.",
""
],
[
"Menegazzo",
"R.",
""
],
[
"Mercogliano",
"D.",
""
],
[
"Piatti",
"D.",
""
],
[
"Prati",
"P.",
""
],
[
"Rapagnani",
"D.",
""
],
[
"Rigato",
"V.",
""
],
[
"Robb",
"D.",
""
],
[
"Rossi",
"M.",
""
],
[
"Sariyal",
"R.",
""
],
[
"Skowronski",
"J.",
""
],
[
"Straniero",
"O.",
""
],
[
"Szucs",
"T.",
""
],
[
"Turkat",
"S.",
""
],
[
"Vagnoni",
"M.",
""
],
[
"Zavatarelli",
"S.",
""
]
] |
2505.22427 | Van-Tin Luu | Van-Tin Luu, Yon-Lin Cai, Vu-Hoang Tran, Wei-Chen Chiu, Yi-Ting Chen, Ching-Chun Huang | RC-AutoCalib: An End-to-End Radar-Camera Automatic Calibration Network | Accepted to CVPR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a groundbreaking approach - the first online automatic geometric calibration method for radar and camera systems. Given the significant data sparsity and measurement uncertainty in radar height data, achieving automatic calibration during system operation has long been a challenge. To address the sparsity issue, we propose a Dual-Perspective representation that gathers features from both frontal and bird's-eye views. The frontal view contains rich but sensitive height information, whereas the bird's-eye view provides robust features against height uncertainty. We thereby propose a novel Selective Fusion Mechanism to identify and fuse reliable features from both perspectives, reducing the effect of height uncertainty. Moreover, for each view, we incorporate a Multi-Modal Cross-Attention Mechanism to explicitly find location correspondences through cross-modal matching. During the training phase, we also design a Noise-Resistant Matcher to provide better supervision and enhance the robustness of the matching mechanism against sparsity and height uncertainty. Our experimental results, tested on the nuScenes dataset, demonstrate that our method significantly outperforms previous radar-camera auto-calibration methods, as well as existing state-of-the-art LiDAR-camera calibration techniques, establishing a new benchmark for future research. The code is available at https://github.com/nycu-acm/RC-AutoCalib. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:52:31 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Luu",
"Van-Tin",
""
],
[
"Cai",
"Yon-Lin",
""
],
[
"Tran",
"Vu-Hoang",
""
],
[
"Chiu",
"Wei-Chen",
""
],
[
"Chen",
"Yi-Ting",
""
],
[
"Huang",
"Ching-Chun",
""
]
] |
2505.22428 | Jie Cai | Ya-Fang Lin, Xiaotian Li, Wan-Hsuan Huang, Charan Pushpanathan Prabavathi, Jie Cai, John M. Carroll | Parental Collaboration and Closeness: Envisioning with New Couple Parents | DIS 2025 | null | 10.1145/3715336.3735837 | null | cs.HC cs.CY | http://creativecommons.org/licenses/by/4.0/ | Couples often experience a decrease in closeness as they cope with the demands of parenthood. Existing technologies have supported parenting and parental collaboration. However, these technologies do not adequately support closeness in co-parenting. We use scenarios and design probes to brainstorm with 10 new parent couples to explore and envision possibilities for technologies to support closeness. We reported parents' current technology use for co-parenting and how participants considered and envisioned co-parenting technology for closeness, including information and task sharing, emotion awareness and disclosure, and fostering fun interaction. We discuss the potential technology has for fostering closeness in co-parenting by (1) fostering interdependence by supporting parental competence and (2) integrating positive emotions and experiences, such as validation and fun, in parenting. Based on our findings, we expand the design space of technology for closeness to include interdependence. We also expand the design space for co-parenting technology by integrating more positive emotions. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:53:08 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Lin",
"Ya-Fang",
""
],
[
"Li",
"Xiaotian",
""
],
[
"Huang",
"Wan-Hsuan",
""
],
[
"Prabavathi",
"Charan Pushpanathan",
""
],
[
"Cai",
"Jie",
""
],
[
"Carroll",
"John M.",
""
]
] |
2505.22429 | Lingdong Kong | Rong Li and Shijie Li and Lingdong Kong and Xulei Yang and Junwei Liang | Zero-Shot 3D Visual Grounding from Vision-Language Models | 3D-LLM/VLA @ CVPR 2025; Project Page at https://seeground.github.io/ | null | null | null | cs.CV cs.RO | http://creativecommons.org/licenses/by-sa/4.0/ | 3D Visual Grounding (3DVG) seeks to locate target objects in 3D scenes using natural language descriptions, enabling downstream applications such as augmented reality and robotics. Existing approaches typically rely on labeled 3D data and predefined categories, limiting scalability to open-world settings. We present SeeGround, a zero-shot 3DVG framework that leverages 2D Vision-Language Models (VLMs) to bypass the need for 3D-specific training. To bridge the modality gap, we introduce a hybrid input format that pairs query-aligned rendered views with spatially enriched textual descriptions. Our framework incorporates two core components: a Perspective Adaptation Module that dynamically selects optimal viewpoints based on the query, and a Fusion Alignment Module that integrates visual and spatial signals to enhance localization precision. Extensive evaluations on ScanRefer and Nr3D confirm that SeeGround achieves substantial improvements over existing zero-shot baselines -- outperforming them by 7.7% and 7.1%, respectively -- and even rivals fully supervised alternatives, demonstrating strong generalization under challenging conditions. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:53:53 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Rong",
""
],
[
"Li",
"Shijie",
""
],
[
"Kong",
"Lingdong",
""
],
[
"Yang",
"Xulei",
""
],
[
"Liang",
"Junwei",
""
]
] |
2505.22430 | Kun Li | Kun Li, Yunxiang Li, Tianhua Zhang, Hongyin Luo, Xixin Wu, James Glass, Helen Meng | RAG-Zeval: Towards Robust and Interpretable Evaluation on RAG Responses through End-to-End Rule-Guided Reasoning | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Robust evaluation is critical for deploying trustworthy retrieval-augmented generation (RAG) systems. However, current LLM-based evaluation frameworks predominantly rely on directly prompting resource-intensive models with complex multi-stage prompts, underutilizing models' reasoning capabilities and introducing significant computational cost. In this paper, we present RAG-Zeval (RAG-Zero Evaluator), a novel end-to-end framework that formulates faithfulness and correctness evaluation as a rule-guided reasoning task. Our approach trains evaluators with reinforcement learning, facilitating compact models to generate comprehensive and sound assessments with detailed explanation in one-pass. We introduce a ranking-based outcome reward mechanism, using preference judgments rather than absolute scores, to address the challenge of obtaining precise pointwise reward signals. To this end, we synthesize the ranking references by generating quality-controlled responses with zero human annotation. Experiments demonstrate RAG-Zeval's superior performance, achieving the strongest correlation with human judgments and outperforming baselines that rely on LLMs with 10-100 times more parameters. Our approach also exhibits superior interpretability in response evaluation. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:55:33 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Kun",
""
],
[
"Li",
"Yunxiang",
""
],
[
"Zhang",
"Tianhua",
""
],
[
"Luo",
"Hongyin",
""
],
[
"Wu",
"Xixin",
""
],
[
"Glass",
"James",
""
],
[
"Meng",
"Helen",
""
]
] |
2505.22431 | Mina Akhyani | Mina Akhyani, Luca Zanini, Henrik R{\o}nnow | Brightify: A tool for calculating brightness in neutron sources | null | null | null | null | physics.data-an physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | Brightness is a critical metric for optimizing the design of neutron sources and beamlines, yet there is no direct way to calculate brightness within most Monte Carlo packages used for neutron source simulation. In this paper, we present Brightify, an open-source Python-based tool designed to calculate brightness from Monte Carlo Particle List (MCPL) files, which can be extracted from several Monte Carlo simulation packages. Brightify provides an efficient computational approach to calculate brightness for any particle type and energy spectrum recorded in the MCPL file. It enables localized, directionally-resolved brightness evaluations by scanning across both spatial and angular domains, facilitating the identification of positions and directions corresponding to maximum brightness. This functionality is particularly valuable for identifying brightness hotspots and helping fine-tune the design of neutron sources for optimal performance. We validate Brightify against standard methods, such as surface current tally and point estimator tally, and demonstrate its accuracy and adaptability, particularly in high-resolution analyses. By overcoming the limitations of traditional methods, Brightify streamlines neutron source re-optimization, reduces computational burden, and accelerates source development workflows. The full code is available on the Brightify GitHub repository. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:55:36 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Akhyani",
"Mina",
""
],
[
"Zanini",
"Luca",
""
],
[
"Rønnow",
"Henrik",
""
]
] |
2505.22432 | Grigorios Zouros | Grigorios P. Zouros, Konstantinos Delimaris, Carsten Rockstuhl, Georgios D. Kolezas | Oblique Multiple Scattering by Gyrotropic Cylinders | This work has been submitted to the IEEE for possible publication | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | In this work, we develop a full-wave vectorial solution for the 2.5-dimensional (2.5-D), i.e., at oblique plane wave incidence, electromagnetic (EM) multiple scattering (MS) by a collection of gyrotropic cylinders. All cylinders are infinitely long and share a common $z$-axis. However, each cylinder can have a different cross-section with an arbitrary shape and different gyrotropic material properties, i.e., both gyroelectric and gyromagnetic anisotropies are considered. The solution to the problem combines the following three elements: (i) development of a superpotentials-based cylindrical vector wave function (CVWF) expansion to express the EM field in the gyrotropic region; (ii) utilization of the extended boundary condition method (EBCM) to account for non-circular cylinders; (iii) use of Graf's formulas, specifically adapted for the CVWFs, to apply the EBCM at each cylinder. The developed theory allows us to calculate various scattering characteristics, including the scattering and extinction cross-sections and the multipole decomposition, enabling the design and in-depth investigation of various contemporary engineering and physics applications. The method is exhaustively validated with analytical techniques and COMSOL Multiphysics. The computational performance is also discussed. Finally, we study a potential microwave application of the MS by ferrite configurations, and demonstrate broadband forward scattering by introducing oblique incidence and anisotropy. Our method may be used to analyze, design, and optimize contemporary microwave, optical, and photonic applications by beneficially tailoring the scattering properties via oblique incidence and anisotropy. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:56:04 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zouros",
"Grigorios P.",
""
],
[
"Delimaris",
"Konstantinos",
""
],
[
"Rockstuhl",
"Carsten",
""
],
[
"Kolezas",
"Georgios D.",
""
]
] |
2505.22434 | Zobia Batool | Zobia Batool, Huseyin Ozkan, Erchan Aptoula | Distance Transform Guided Mixup for Alzheimer's Detection | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Alzheimer's detection efforts aim to develop accurate models for early disease diagnosis. Significant advances have been achieved with convolutional neural networks and vision transformer based approaches. However, medical datasets suffer heavily from class imbalance, variations in imaging protocols, and limited dataset diversity, which hinder model generalization. To overcome these challenges, this study focuses on single-domain generalization by extending the well-known mixup method. The key idea is to compute the distance transform of MRI scans, separate them spatially into multiple layers and then combine layers stemming from distinct samples to produce augmented images. The proposed approach generates diverse data while preserving the brain's structure. Experimental results show generalization performance improvement across both ADNI and AIBL datasets. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:56:59 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Batool",
"Zobia",
""
],
[
"Ozkan",
"Huseyin",
""
],
[
"Aptoula",
"Erchan",
""
]
] |
2505.22435 | Victor J\"uttner | Victor J\"uttner and Erik Buchmann | Does Johnny Get the Message? Evaluating Cybersecurity Notifications for Everyday Users | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Due to the increasing presence of networked devices in everyday life, not only cybersecurity specialists but also end users benefit from security applications such as firewalls, vulnerability scanners, and intrusion detection systems. Recent approaches use large language models (LLMs) to rewrite brief, technical security alerts into intuitive language and suggest actionable measures, helping everyday users understand and respond appropriately to security risks. However, it remains an open question how well such alerts are explained to users. LLM outputs can also be hallucinated, inconsistent, or misleading. In this work, we introduce the Human-Centered Security Alert Evaluation Framework (HCSAEF). HCSAEF assesses LLM-generated cybersecurity notifications to support researchers who want to compare notifications generated for everyday users, improve them, or analyze the capabilities of different LLMs in explaining cybersecurity issues. We demonstrate HCSAEF through three use cases, which allow us to quantify the impact of prompt design, model selection, and output consistency. Our findings indicate that HCSAEF effectively differentiates generated notifications along dimensions such as intuitiveness, urgency, and correctness. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 14:58:29 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jüttner",
"Victor",
""
],
[
"Buchmann",
"Erik",
""
]
] |
2505.22436 | Floris van Breugel | Arunava Nag, Floris van Breugel | COSMOS: A Data-Driven Probabilistic Time Series simulator for Chemical Plumes across Spatial Scales | 16 pages, 4 primary figures | null | null | null | stat.AP cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The development of robust odor navigation strategies for automated environmental monitoring applications requires realistic simulations of odor time series for agents moving across large spatial scales. Traditional approaches that rely on computational fluid dynamics (CFD) methods can capture the spatiotemporal dynamics of odor plumes, but are impractical for large-scale simulations due to their computational expense. On the other hand, puff-based simulations, although computationally tractable for large scales and capable of capturing the stochastic nature of plumes, fail to reproduce naturalistic odor statistics. Here, we present COSMOS (Configurable Odor Simulation Model over Scalable Spaces), a data-driven probabilistic framework that synthesizes realistic odor time series from spatial and temporal features of real datasets. COSMOS generates similar distributions of key statistical features such as whiff frequency, duration, and concentration as observed in real data, while dramatically reducing computational overhead. By reproducing critical statistical properties across a variety of flow regimes and scales, COSMOS enables the development and evaluation of agent-based navigation strategies with naturalistic odor experiences. To demonstrate its utility, we compare odor-tracking agents exposed to CFD-generated plumes versus COSMOS simulations, showing that both their odor experiences and resulting behaviors are quite similar. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:00:22 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Nag",
"Arunava",
""
],
[
"van Breugel",
"Floris",
""
]
] |
2505.22437 | Lucas Butsch | Lucas Butsch and Vicky Fasen-Hartmann | Estimation of the number of principal components in high-dimensional multivariate extremes | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For multivariate regularly random vectors of dimension $d$, the dependence structure of the extremes is modeled by the so-called angular measure. When the dimension $d$ is high, estimating the angular measure is challenging because of its complexity. In this paper, we use Principal Component Analysis (PCA) as a method for dimension reduction and estimate the number of significant principal components of the empirical covariance matrix of the angular measure under the assumption of a spiked covariance structure. Therefore, we develop Akaike Information Criteria (AIC) and Bayesian Information Criteria (BIC) to estimate the location of the spiked eigenvalue of the covariance matrix, reflecting the number of significant components, and explore these information criteria on consistency. On the one hand, we investigate the case where the dimension $d$ is fixed, and on the other hand, where the dimension $d$ converges to $\infty$ under different high-dimensional scenarios. When the dimension $d$ is fixed, we establish that the AIC is not consistent, whereas the BIC is weakly consistent. In the high-dimensional setting, with techniques from random matrix theory, we derive sufficient conditions for the AIC and the BIC to be consistent. Finally, the performance of the different AIC and BIC versions is compared in a simulation study and applied to high-dimensional precipitation data. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:01:10 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Butsch",
"Lucas",
""
],
[
"Fasen-Hartmann",
"Vicky",
""
]
] |
2505.22438 | Zijian Liang | Zijian Liang, Kai Niu, Changshuo Wang, Jin Xu, Ping Zhang | Synonymous Variational Inference for Perceptual Image Compression | 31 pages, 20 figures. This paper is accepted by Proceedings of the 42nd International Conference on Machine Learning (ICML 2025) Poster | null | null | null | cs.IT cs.AI cs.CV cs.LG eess.IV math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent contributions of semantic information theory reveal the set-element relationship between semantic and syntactic information, represented as synonymous relationships. In this paper, we propose a synonymous variational inference (SVI) method based on this synonymity viewpoint to re-analyze the perceptual image compression problem. It takes perceptual similarity as a typical synonymous criterion to build an ideal synonymous set (Synset), and approximate the posterior of its latent synonymous representation with a parametric density by minimizing a partial semantic KL divergence. This analysis theoretically proves that the optimization direction of perception image compression follows a triple tradeoff that can cover the existing rate-distortion-perception schemes. Additionally, we introduce synonymous image compression (SIC), a new image compression scheme that corresponds to the analytical process of SVI, and implement a progressive SIC codec to fully leverage the model's capabilities. Experimental results demonstrate comparable rate-distortion-perception performance using a single progressive SIC codec, thus verifying the effectiveness of our proposed analysis method. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:03:27 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Liang",
"Zijian",
""
],
[
"Niu",
"Kai",
""
],
[
"Wang",
"Changshuo",
""
],
[
"Xu",
"Jin",
""
],
[
"Zhang",
"Ping",
""
]
] |
2505.22439 | Abra\~ao Mendes | M\'arcio Batista, Marcos P. Cavalcante, Abra\~ao Mendes, Ivaldo Nunes | Rigidity of surfaces with nonpositive Euler characteristic by the second eigenvalue of the Jacobi operator | 15 pages. Comments welcome | null | null | null | math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we investigate the spectral properties of the Jacobi operator for immersed surfaces with nonpositive Euler characteristic, extending previous results in the field. We first prove a sharp upper bound for the second eigenvalue of the Jacobi operator for compact surfaces with nonpositive Euler characteristic that are fully immersed in the Euclidean sphere, and then we classify all such surfaces attaining this upper bound. Furthermore, we demonstrate that totally geodesic tori maximize the second eigenvalue among all compact orientable surfaces with positive genus in the product space $\mathbb{S}^1(r) \times \mathbb{S}^2(s)$. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:04:30 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Batista",
"Márcio",
""
],
[
"Cavalcante",
"Marcos P.",
""
],
[
"Mendes",
"Abraão",
""
],
[
"Nunes",
"Ivaldo",
""
]
] |
2505.22440 | Ali Shiri Sichani | Khan Masood Parvez, Sk Md Abidar Rahaman, and Ali Shiri Sichani | Data-Driven Antenna Miniaturization: A Knowledge-Based System Integrating Quantum PSO and Predictive Machine Learning Models | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid evolution of wireless technologies necessitates automated design frameworks to address antenna miniaturization and performance optimization within constrained development cycles. This study demonstrates a machine learning enhanced workflow integrating Quantum-Behaved Dynamic Particle Swarm Optimization (QDPSO) with ANSYS HFSS simulations to accelerate antenna design. The QDPSO algorithm autonomously optimized loop dimensions in 11.53 seconds, achieving a resonance frequency of 1.4208 GHz a 12.7 percent reduction compared to conventional 1.60 GHz designs. Machine learning models (SVM, Random Forest, XGBoost, and Stacked ensembles) predicted resonance frequencies in 0.75 seconds using 936 simulation datasets, with stacked models showing superior training accuracy (R2=0.9825) and SVM demonstrating optimal validation performance (R2=0.7197). The complete design cycle, encompassing optimization, prediction, and ANSYS validation, required 12.42 minutes on standard desktop hardware (Intel i5-8500, 16GB RAM), contrasting sharply with the 50-hour benchmark of PSADEA-based approaches. This 240 times of acceleration eliminates traditional trial-and-error methods that often extend beyond seven expert-led days. The system enables precise specifications of performance targets with automated generation of fabrication-ready parameters, particularly benefiting compact consumer devices requiring rapid frequency tuning. By bridging AI-driven optimization with CAD validation, this framework reduces engineering workloads while ensuring production-ready designs, establishing a scalable paradigm for next-generation RF systems in 6G and IoT applications. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:04:36 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Parvez",
"Khan Masood",
""
],
[
"Rahaman",
"Sk Md Abidar",
""
],
[
"Sichani",
"Ali Shiri",
""
]
] |
2505.22441 | Chaitanya Amballa | Chaitanya Amballa, Sattwik Basu, Yu-Lin Wei, Zhijian Yang, Mehmet Ergezer, Romit Roy Choudhury | Can NeRFs See without Cameras? | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Neural Radiance Fields (NeRFs) have been remarkably successful at synthesizing novel views of 3D scenes by optimizing a volumetric scene function. This scene function models how optical rays bring color information from a 3D object to the camera pixels. Radio frequency (RF) or audio signals can also be viewed as a vehicle for delivering information about the environment to a sensor. However, unlike camera pixels, an RF/audio sensor receives a mixture of signals that contain many environmental reflections (also called "multipath"). Is it still possible to infer the environment using such multipath signals? We show that with redesign, NeRFs can be taught to learn from multipath signals, and thereby "see" the environment. As a grounding application, we aim to infer the indoor floorplan of a home from sparse WiFi measurements made at multiple locations inside the home. Although a difficult inverse problem, our implicitly learnt floorplans look promising, and enables forward applications, such as indoor signal prediction and basic ray tracing. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:04:46 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Amballa",
"Chaitanya",
""
],
[
"Basu",
"Sattwik",
""
],
[
"Wei",
"Yu-Lin",
""
],
[
"Yang",
"Zhijian",
""
],
[
"Ergezer",
"Mehmet",
""
],
[
"Choudhury",
"Romit Roy",
""
]
] |
2505.22442 | Mattie Fellows | Mattie Fellows, Clarisse Wibault, Uljad Berdica, Johannes Forkel, Jakob N. Foerster, Michael A. Osborne | SOReL and TOReL: Two Methods for Fully Offline Reinforcement Learning | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Sample efficiency remains a major obstacle for real world adoption of reinforcement learning (RL): success has been limited to settings where simulators provide access to essentially unlimited environment interactions, which in reality are typically costly or dangerous to obtain. Offline RL in principle offers a solution by exploiting offline data to learn a near-optimal policy before deployment. In practice, however, current offline RL methods rely on extensive online interactions for hyperparameter tuning, and have no reliable bound on their initial online performance. To address these two issues, we introduce two algorithms. Firstly, SOReL: an algorithm for safe offline reinforcement learning. Using only offline data, our Bayesian approach infers a posterior over environment dynamics to obtain a reliable estimate of the online performance via the posterior predictive uncertainty. Crucially, all hyperparameters are also tuned fully offline. Secondly, we introduce TOReL: a tuning for offline reinforcement learning algorithm that extends our information rate based offline hyperparameter tuning methods to general offline RL approaches. Our empirical evaluation confirms SOReL's ability to accurately estimate regret in the Bayesian setting whilst TOReL's offline hyperparameter tuning achieves competitive performance with the best online hyperparameter tuning methods using only offline data. Thus, SOReL and TOReL make a significant step towards safe and reliable offline RL, unlocking the potential for RL in the real world. Our implementations are publicly available: https://github.com/CWibault/sorel\_torel. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:07:24 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Fellows",
"Mattie",
""
],
[
"Wibault",
"Clarisse",
""
],
[
"Berdica",
"Uljad",
""
],
[
"Forkel",
"Johannes",
""
],
[
"Foerster",
"Jakob N.",
""
],
[
"Osborne",
"Michael A.",
""
]
] |
2505.22443 | Selina Cheggour | Selina Cheggour and Valeria Loscri | Frequency Resource Management in 6G User-Centric CFmMIMO: A Hybrid Reinforcement Learning and Metaheuristic Approach | null | null | null | null | cs.NI eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As sixth-generation (6G) networks continue to evolve, AI-driven solutions are playing a crucial role in enabling more efficient and adaptive resource management in wireless communication. One of the key innovations in 6G is user-centric cell-free massive Multiple-Input Multiple-Output (UC-CFmMIMO), a paradigm that eliminates traditional cell boundaries and enhances network performance by dynamically assigning access points (APs) to users. This approach is particularly well-suited for vehicular networks, offering seamless, homogeneous, ultra-reliable, and low-latency connectivity. However, in dense networks, a key challenge lies in efficiently allocating frequency resources within a limited shared subband spectrum while accounting for frequency selectivity and the dependency of signal propagation on bandwidth. These factors make resource allocation increasingly complex, especially in dynamic environments where maintaining Quality of Service (QoS) is critical. This paper tackles these challenges by proposing a hybrid multi-user allocation strategy that integrates reinforcement learning (RL) and metaheuristic optimization to enhance spectral efficiency (SE), ensure fairness, and mitigate interference within shared subbands. To assess its effectiveness, we compare this hybrid approach with two other methods: the bio-inspired Aquila Optimizer (AO) and Deep Deterministic Policy Gradient (DDPG)-based Actor-Critic Reinforcement Learning (AC-RL). Our evaluation is grounded in real-world patterns and channel characteristics, utilizing the 3GPP-3D channel modeling framework (QuaDRiGa) to capture realistic propagation conditions. The results demonstrate that the proposed hybrid strategy achieves a superior balance among competing objectives, underscoring the role of AI-driven resource allocation in advancing UC-CFmMIMO systems for next-generation wireless networks. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:07:56 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Cheggour",
"Selina",
""
],
[
"Loscri",
"Valeria",
""
]
] |
2505.22444 | Liyao Tang | Liyao Tang, Zhe Chen, Dacheng Tao | On Geometry-Enhanced Parameter-Efficient Fine-Tuning for 3D Scene Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The emergence of large-scale pre-trained point cloud models has significantly advanced 3D scene understanding, but adapting these models to specific downstream tasks typically demands full fine-tuning, incurring high computational and storage costs. Parameter-efficient fine-tuning (PEFT) techniques, successful in natural language processing and 2D vision tasks, would underperform when naively applied to 3D point cloud models due to significant geometric and spatial distribution shifts. Existing PEFT methods commonly treat points as orderless tokens, neglecting important local spatial structures and global geometric contexts in 3D modeling. To bridge this gap, we introduce the Geometric Encoding Mixer (GEM), a novel geometry-aware PEFT module specifically designed for 3D point cloud transformers. GEM explicitly integrates fine-grained local positional encodings with a lightweight latent attention mechanism to capture comprehensive global context, thereby effectively addressing the spatial and geometric distribution mismatch. Extensive experiments demonstrate that GEM achieves performance comparable to or sometimes even exceeding full fine-tuning, while only updating 1.6% of the model's parameters, fewer than other PEFT methods. With significantly reduced training time and memory requirements, our approach thus sets a new benchmark for efficient, scalable, and geometry-aware fine-tuning of large-scale 3D point cloud models. Code will be released. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:08:36 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Tang",
"Liyao",
""
],
[
"Chen",
"Zhe",
""
],
[
"Tao",
"Dacheng",
""
]
] |
2505.22445 | Zhangquan Chen | Puhua Jiang, Zhangquan Chen, Mingze Sun, Ruqi Huang | NFR: Neural Feature-Guided Non-Rigid Shape Registration | 20 pages, 9 figures. arXiv admin note: substantial text overlap with arXiv:2311.04494 | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | In this paper, we propose a novel learning-based framework for 3D shape registration, which overcomes the challenges of significant non-rigid deformation and partiality undergoing among input shapes, and, remarkably, requires no correspondence annotation during training. Our key insight is to incorporate neural features learned by deep learning-based shape matching networks into an iterative, geometric shape registration pipeline. The advantage of our approach is two-fold -- On one hand, neural features provide more accurate and semantically meaningful correspondence estimation than spatial features (e.g., coordinates), which is critical in the presence of large non-rigid deformations; On the other hand, the correspondences are dynamically updated according to the intermediate registrations and filtered by consistency prior, which prominently robustify the overall pipeline. Empirical results show that, with as few as dozens of training shapes of limited variability, our pipeline achieves state-of-the-art results on several benchmarks of non-rigid point cloud matching and partial shape matching across varying settings, but also delivers high-quality correspondences between unseen challenging shape pairs that undergo both significant extrinsic and intrinsic deformations, in which case neither traditional registration methods nor intrinsic methods work. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:08:49 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jiang",
"Puhua",
""
],
[
"Chen",
"Zhangquan",
""
],
[
"Sun",
"Mingze",
""
],
[
"Huang",
"Ruqi",
""
]
] |
2505.22446 | Qingwu Wang | Qing-Wu Wang, Xiao-Fu \L\text\"u, Hua-Zhong Guo | $ \rho\to \pi\pi $ Hadronic Decay in the Nambu-Jona-Lasinio Model: Mass-Width Interplay and Beyond-RPA Corrections | 6 pages,2 figures | null | null | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | We present a novel framework for analyzing unstable composite particles using Green's functions and dispersion relations. As an illustrative example, we explore the
$\rho$ vector meson decay process $\rho\to\pi\pi$
within the Nambu -- Jona - Lasinio (NJL) model. Our approach addresses a key limitation of the four-quark interaction description, which adequately describes two-quark bound states but fails to describe decay
processes. The Bethe-Salpeter(BS) wave function of the $\rho$ meson exhibits time evolution that leads to the physical mass
$M$ incorporating a correction $\Delta M $. This correction depends on the decay width $\Gamma(M)$. This work provides crucial insights into the dynamical relationship between resonance masses and their decay properties, addressing a long-standing challenge in hadron physics. The calculated mass and width are in good agreement with the experimental values, demonstrating the effectiveness of this approach for studying unstable hadronic systems beyond conventional bound-state approximations. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:09:40 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wang",
"Qing-Wu",
""
],
[
"Ł\\textü",
"Xiao-Fu",
""
],
[
"Guo",
"Hua-Zhong",
""
]
] |
2505.22447 | Sizai Hou | Sizai Hou, Songze Li, Baturalp Buyukates | Privacy-preserving Prompt Personalization in Federated Learning for Multimodal Large Language Models | Under Review | null | null | null | cs.CR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Prompt learning is a crucial technique for adapting pre-trained multimodal language models (MLLMs) to user tasks. Federated prompt personalization (FPP) is further developed to address data heterogeneity and local overfitting, however, it exposes personalized prompts - valuable intellectual assets - to privacy risks like prompt stealing or membership inference attacks. Widely-adopted techniques like differential privacy add noise to prompts, whereas degrading personalization performance. We propose SecFPP, a secure FPP protocol harmonizing generalization, personalization, and privacy guarantees. SecFPP employs hierarchical prompt adaptation with domain-level and class-level components to handle multi-granular data imbalance. For privacy, it uses a novel secret-sharing-based adaptive clustering algorithm for domain-level adaptation while keeping class-level components private. While theoretically and empirically secure, SecFPP achieves state-of-the-art accuracy under severe heterogeneity in data distribution. Extensive experiments show it significantly outperforms both non-private and privacy-preserving baselines, offering a superior privacy-performance trade-off. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:09:56 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hou",
"Sizai",
""
],
[
"Li",
"Songze",
""
],
[
"Buyukates",
"Baturalp",
""
]
] |
2505.22448 | Andrew Toms | Andrew S. Toms, Hao Wan | Self-adjoint operators in Z-stable C$^*$-algebras with prescribed spectral data | 18 pages, no figures | null | null | null | math.OA math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the variety of spectral measures that are induced by quasitraces on the spectrum of a self-adjoint operator in a simple separable unital and Z-stable C$^*$-algebra. This amounts to a continuous map from the simplex of quasitraces of the C$^*$-algebra into regular Borel probability measures on the spectrum of the operator under consideration. In the case of a connected spectrum this data determines the unitary equivalence class of the operator, and may be reduced to to the case of an operator with spectrum equal to the closed unit interval. We prove that any continuous map from the simplex of quasitraces with the topology of pointwise convergence into regular faithful Borel probability measures on $[0,1]$ with the Levy-Prokhorov metric is realized by some self-adjoint operator in the C$^*$-algebra. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:10:20 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Toms",
"Andrew S.",
""
],
[
"Wan",
"Hao",
""
]
] |
2505.22449 | Joel Daniel Andersson | Joel Daniel Andersson, Lukas Retschmeier, Boel Nelson, Rasmus Pagh | Private Lossless Multiple Release | null | null | null | null | cs.CR cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Koufogiannis et al. (2016) showed a $\textit{gradual release}$ result for Laplace noise-based differentially private mechanisms: given an $\varepsilon$-DP release, a new release with privacy parameter $\varepsilon' > \varepsilon$ can be computed such that the combined privacy loss of both releases is at most $\varepsilon'$ and the distribution of the latter is the same as a single release with parameter $\varepsilon'$. They also showed gradual release techniques for Gaussian noise, later also explored by Whitehouse et al. (2022).
In this paper, we consider a more general $\textit{multiple release}$ setting in which analysts hold private releases with different privacy parameters corresponding to different access/trust levels. These releases are determined one by one, with privacy parameters in arbitrary order. A multiple release is $\textit{lossless}$ if having access to a subset $S$ of the releases has the same privacy guarantee as the least private release in $S$, and each release has the same distribution as a single release with the same privacy parameter. Our main result is that lossless multiple release is possible for a large class of additive noise mechanisms. For the Gaussian mechanism we give a simple method for lossless multiple release with a short, self-contained analysis that does not require knowledge of the mathematics of Brownian motion. We also present lossless multiple release for the Laplace and Poisson mechanisms. Finally, we consider how to efficiently do gradual release of sparse histograms, and present a mechanism with running time independent of the number of dimensions. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:10:27 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Andersson",
"Joel Daniel",
""
],
[
"Retschmeier",
"Lukas",
""
],
[
"Nelson",
"Boel",
""
],
[
"Pagh",
"Rasmus",
""
]
] |
2505.22450 | Ossi R\"ais\"a | Ossi R\"ais\"a, Boris van Breugel, Mihaela van der Schaar | Position: All Current Generative Fidelity and Diversity Metrics are Flawed | ICML 2025 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Any method's development and practical application is limited by our ability to measure its reliability. The popularity of generative modeling emphasizes the importance of good synthetic data metrics. Unfortunately, previous works have found many failure cases in current metrics, for example lack of outlier robustness and unclear lower and upper bounds. We propose a list of desiderata for synthetic data metrics, and a suite of sanity checks: carefully chosen simple experiments that aim to detect specific and known generative modeling failure modes. Based on these desiderata and the results of our checks, we arrive at our position: all current generative fidelity and diversity metrics are flawed. This significantly hinders practical use of synthetic data. Our aim is to convince the research community to spend more effort in developing metrics, instead of models. Additionally, through analyzing how current metrics fail, we provide practitioners with guidelines on how these metrics should (not) be used. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:10:33 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Räisä",
"Ossi",
""
],
[
"van Breugel",
"Boris",
""
],
[
"van der Schaar",
"Mihaela",
""
]
] |
2505.22451 | Yuanhang Liu | Yuanhang Liu, Yanxing Huang, Yanqiao Wang, Peng Li, Yang Liu | AI Mathematician: Towards Fully Automated Frontier Mathematical Research | 95 pages, 1 figure | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Reasoning Models (LRMs) have made significant progress in mathematical capabilities in recent times. However, these successes have been primarily confined to competition-level problems. In this work, we propose AI Mathematician (AIM) framework, which harnesses the reasoning strength of LRMs to support frontier mathematical research. We have identified two critical challenges of mathematical research compared to competition, {\it the intrinsic complexity of research problems} and {\it the requirement of procedural rigor}. To address these challenges, AIM incorporates two core strategies: an exploration mechanism to foster longer solution paths, and the pessimistic reasonable verification method to ensure reliability.
This early version of AIM already exhibits strong capability in tackling research-level tasks. We conducted extensive experiments across several real-world mathematical topics and obtained promising results. AIM is able to autonomously construct substantial portions of proofs and uncover non-trivial insights within each research area. These findings highlight the potential of LRMs in mathematical discovery and suggest that LRM-based agent systems could significantly accelerate mathematical research in the future. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:10:37 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Liu",
"Yuanhang",
""
],
[
"Huang",
"Yanxing",
""
],
[
"Wang",
"Yanqiao",
""
],
[
"Li",
"Peng",
""
],
[
"Liu",
"Yang",
""
]
] |
2505.22452 | Luca Fresta | Luca Fresta, Giovanna Marcelli | Spin transport and lack of quantisation in the $A\mathrm{II}$ class on the honeycomb structure | 48 pages, 4 figures | null | null | null | math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate spin transport in a class of two-dimensional $A\mathrm{II}$ insulators on the honeycomb structure, the Kane-Mele model being an emblematic example in this class. We derive the spin conductivity by the linear response \`a la Kubo and show that it is well-defined and independent of the choice of the spin current. For models that do not conserve the spin, we demonstrate that the deviation of the spin conductivity from the quantised value is, at worst, quadratic in the spin-non-conserving terms, thus improving previous results. Additionally, we show that the leading-order corrections are actually quadratic for some models in the class, demonstrating that the spin conductivity is not universally quantised. Consequently, our results show that, in general, there is no direct connection between the spin conductivity and the Fu-Kane-Mele index. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:10:54 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Fresta",
"Luca",
""
],
[
"Marcelli",
"Giovanna",
""
]
] |
2505.22453 | Lai Wei | Lai Wei, Yuting Li, Chen Wang, Yue Wang, Linghe Kong, Weiran Huang, Lichao Sun | Unsupervised Post-Training for Multi-Modal LLM Reasoning via GRPO | null | null | null | null | cs.CL cs.AI cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Improving Multi-modal Large Language Models (MLLMs) in the post-training stage typically relies on supervised fine-tuning (SFT) or reinforcement learning (RL). However, these supervised methods require expensive and manually annotated multi-modal data--an ultimately unsustainable resource. While recent efforts have explored unsupervised post-training, their methods are complex and difficult to iterate. In this work, we are the first to investigate the use of GRPO, a stable and scalable online RL algorithm, for enabling continual self-improvement without any external supervision. We propose MM-UPT, a simple yet effective framework for unsupervised post-training of MLLMs. MM-UPT builds upon GRPO, replacing traditional reward signals with a self-rewarding mechanism based on majority voting over multiple sampled responses. Our experiments demonstrate that MM-UPT significantly improves the reasoning ability of Qwen2.5-VL-7B (e.g., 66.3 %$\rightarrow$72.9 % on MathVista, 62.9 %$\rightarrow$68.7 % on We-Math), using standard dataset without ground truth labels. MM-UPT also outperforms prior unsupervised baselines and even approaches the results of supervised GRPO. Furthermore, we show that incorporating synthetic questions, generated solely by MLLM itself, can boost performance as well, highlighting a promising approach for scalable self-improvement. Overall, MM-UPT offers a new paradigm for continual, autonomous enhancement of MLLMs in the absence of external supervision. Our code is available at https://github.com/waltonfuture/MM-UPT. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:11:16 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wei",
"Lai",
""
],
[
"Li",
"Yuting",
""
],
[
"Wang",
"Chen",
""
],
[
"Wang",
"Yue",
""
],
[
"Kong",
"Linghe",
""
],
[
"Huang",
"Weiran",
""
],
[
"Sun",
"Lichao",
""
]
] |
2505.22454 | Sonia Lopez Alarcon | Mark Danza and Sonia Lopez Alarcon and Cory Merkel | Depth-Based Matrix Classification for the HHL Quantum Algorithm | null | null | null | null | quant-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Under the nearing error-corrected era of quantum computing, it is necessary to understand the suitability of certain post-NISQ algorithms for practical problems. One of the most promising, applicable and yet difficult to implement in practical terms is the Harrow, Hassidim and Lloyd (HHL) algorithm for linear systems of equations. An enormous number of problems can be expressed as linear systems of equations, from Machine Learning to fluid dynamics. However, in most cases, HHL will not be able to provide a practical, reasonable solution to these problems. This paper's goal inquires about whether problems can be labeled using Machine Learning classifiers as suitable or unsuitable for HHL implementation when some numerical information about the problem is known beforehand. This work demonstrates that training on significantly representative data distributions is critical to achieve good classifications of the problems based on the numerical properties of the matrix representing the system of equations. Accurate classification is possible through Multi-Layer Perceptrons, although with careful design of the training data distribution and classifier parameters. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:11:53 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Danza",
"Mark",
""
],
[
"Alarcon",
"Sonia Lopez",
""
],
[
"Merkel",
"Cory",
""
]
] |
2505.22455 | Fr\'ed\'eric Berthommier | Fr\'ed\'eric Berthommier | Articulatory modeling of the S-shaped F2 trajectories observed in \"Ohman's spectrographic analysis of VCV syllables | 5 pages, 4 figures, submitted to Interspeech 2025 | null | null | null | eess.AS cs.SD | http://creativecommons.org/licenses/by/4.0/ | The synthesis of Ohman's VCV sequences with intervocalic plosive consonants was first achieved 30 years ago using the DRM model. However, this approach remains primarily acoustic and lacks articulatory constraints. In this study, the same 75 VCVs are analyzed, but generated with the Maeda model, using trajectory planning that differentiates vowel-to-vowel transitions from consonantal influences. Synthetic data exhibit similar characteristics to Ohman's sequences, including the presence of S-shaped F2 trajectories. Furthermore, locus equations (LEs) for F2 and F3 are computed from synthetic CV data to investigate their underlying determinism, leading to a reassessment of conventional interpretations. The findings indicate that, although articulatory planning is structured separately for vowel and consonant groups, S-shaped F2 trajectories emerge from a composite mechanism governed by the coordinated synergy of all articulators. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:12:53 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Berthommier",
"Frédéric",
""
]
] |
2505.22456 | Roni Blushtein | Roni Blushtein-Livnon, Tal Svoray, Itai Ficshhendler, Havatzelet Yanel, Emir Galilee, Michael Dorman | Beyond Leaders and Laggards: A Typology of Renewable Energy Adoption Trajectories with Evidence from Off-Grid Communities | null | null | null | null | econ.GN q-fin.EC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding the dynamics of renewable energy adoption is essential for designing strategies that accelerate its spread - an urgent priority for advancing climate goals and improving well-being, especially in off-grid regions facing energy poverty. This study introduces a time-series-based analytical framework that quantifies and classifies adoption behaviors of geographic entities within a region. A novel metric, the Adoption over Time Index (ATI), captures cumulative adoption intensity and identifies shifts in adoption trends, improving the ability to distinguish between fundamental adoption paths. By combining ATI with three key features found to be indicative of adoption dynamics, we define a typology of eight distinct paths, including two newly identified trajectories - the decelerating path and the declining moderate path. Applying this framework to a case study of an off-grid Bedouin population in southern Israel reveals that these retreating paths exist in substantial proportions. Identifying such trends is critical for addressing stagnation and preventing backsliding in the diffusion process. The leaping path, by contrast, was nearly absent. We also identify behavioral diversity within both front-runner and trailing groups. Differentiating among these groups can help tailor acceleration strategies. The analysis further reveals significant disparities in adoption levels across the region, with lagging clusters being widespread and overall adoption falling short of the region's potential. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:13:05 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Blushtein-Livnon",
"Roni",
""
],
[
"Svoray",
"Tal",
""
],
[
"Ficshhendler",
"Itai",
""
],
[
"Yanel",
"Havatzelet",
""
],
[
"Galilee",
"Emir",
""
],
[
"Dorman",
"Michael",
""
]
] |
2505.22457 | Haonan Wang | Haonan Wang, Hongfu Liu, Xiangyan Liu, Chao Du, Kenji Kawaguchi, Ye Wang, Tianyu Pang | Fostering Video Reasoning via Next-Event Prediction | null | null | null | null | cs.CV cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Next-token prediction serves as the foundational learning task enabling reasoning in LLMs. But what should the learning task be when aiming to equip MLLMs with temporal reasoning capabilities over video inputs? Existing tasks such as video question answering often rely on annotations from humans or much stronger MLLMs, while video captioning tends to entangle temporal reasoning with spatial information. To address this gap, we propose next-event prediction (NEP), a learning task that harnesses future video segments as a rich, self-supervised signal to foster temporal reasoning. We segment each video into past and future frames: the MLLM takes the past frames as input and predicts a summary of events derived from the future frames, thereby encouraging the model to reason temporally in order to complete the task. To support this task, we curate V1-33K, a dataset comprising 33,000 automatically extracted video segments spanning diverse real-world scenarios. We further explore a range of video instruction-tuning strategies to study their effects on temporal reasoning. To evaluate progress, we introduce FutureBench to assess coherence in predicting unseen future events. Experiments validate that NEP offers a scalable and effective training paradigm for fostering temporal reasoning in MLLMs. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:13:34 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wang",
"Haonan",
""
],
[
"Liu",
"Hongfu",
""
],
[
"Liu",
"Xiangyan",
""
],
[
"Du",
"Chao",
""
],
[
"Kawaguchi",
"Kenji",
""
],
[
"Wang",
"Ye",
""
],
[
"Pang",
"Tianyu",
""
]
] |
2505.22458 | SeunAn Choe | Seun-An Choe, Keon-Hee Park, Jinwoo Choi, Gyeong-Moon Park | Universal Domain Adaptation for Semantic Segmentation | Accepted by CVPR 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Unsupervised domain adaptation for semantic segmentation (UDA-SS) aims to transfer knowledge from labeled source data to unlabeled target data. However, traditional UDA-SS methods assume that category settings between source and target domains are known, which is unrealistic in real-world scenarios. This leads to performance degradation if private classes exist. To address this limitation, we propose Universal Domain Adaptation for Semantic Segmentation (UniDA-SS), achieving robust adaptation even without prior knowledge of category settings. We define the problem in the UniDA-SS scenario as low confidence scores of common classes in the target domain, which leads to confusion with private classes. To solve this problem, we propose UniMAP: UniDA-SS with Image Matching and Prototype-based Distinction, a novel framework composed of two key components. First, Domain-Specific Prototype-based Distinction (DSPD) divides each class into two domain-specific prototypes, enabling finer separation of domain-specific features and enhancing the identification of common classes across domains. Second, Target-based Image Matching (TIM) selects a source image containing the most common-class pixels based on the target pseudo-label and pairs it in a batch to promote effective learning of common classes. We also introduce a new UniDA-SS benchmark and demonstrate through various experiments that UniMAP significantly outperforms baselines. The code is available at \href{https://github.com/KU-VGI/UniMAP}{this https URL}. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:14:11 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Choe",
"Seun-An",
""
],
[
"Park",
"Keon-Hee",
""
],
[
"Choi",
"Jinwoo",
""
],
[
"Park",
"Gyeong-Moon",
""
]
] |
2505.22459 | Srijan Sengupta | Subhankar Bhadra, Minh Tang, Srijan Sengupta | A Unified Framework for Community Detection and Model Selection in Blockmodels | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Blockmodels are a foundational tool for modeling community structure in networks, with the stochastic blockmodel (SBM), degree-corrected blockmodel (DCBM), and popularity-adjusted blockmodel (PABM) forming a natural hierarchy of increasing generality. While community detection under these models has been extensively studied, much less attention has been paid to the model selection problem, i.e., determining which model best fits a given network. Building on recent theoretical insights about the spectral geometry of these models, we propose a unified framework for simultaneous community detection and model selection across the full blockmodel hierarchy. A key innovation is the use of loss functions that serve a dual role: they act as objective functions for community detection and as test statistics for hypothesis testing. We develop a greedy algorithm to minimize these loss functions and establish theoretical guarantees for exact label recovery and model selection consistency under each model. Extensive simulation studies demonstrate that our method achieves high accuracy in both tasks, outperforming or matching state-of-the-art alternatives. Applications to five real-world networks further illustrate the interpretability and practical utility of our approach. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:15:36 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Bhadra",
"Subhankar",
""
],
[
"Tang",
"Minh",
""
],
[
"Sengupta",
"Srijan",
""
]
] |
2505.22460 | Lukas Vogelhuber | Lukas Vogelhuber, Ihor Korolov, Mate Vass, Katharina Noesges, Tim Bolles, Kevin K\"ohn, Maximilian Klich, Ralf Peter Brinkmann and Thomas Mussenbrock | Electronegativity effects on plasma dynamics in He/O$_2$ RF microplasma jets at atmospheric pressure | null | null | null | null | physics.plasm-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work investigates the transitions between ohmic mode and Penning-Gamma mode in a capacitively coupled radio frequency micro atmospheric pressure plasma jets (CCRF $\mu$APPJ) operated in He/O$_2$ mixtures by comparing phase-resolved optical emission spectroscopy (PROES) measurements of helium excitation with numerical simulations. The simulations employ a hybrid model that treats electrons kinetically via PIC/MCC, while ions and neutrals are modeled fluid dynamically. These results reveal that increasing electronegativity causes inhomogeneities in the bulk electric field, consequently modulating electron impact excitation dynamics. A good agreement was found between experiments and simulations. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:15:58 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Vogelhuber",
"Lukas",
""
],
[
"Korolov",
"Ihor",
""
],
[
"Vass",
"Mate",
""
],
[
"Noesges",
"Katharina",
""
],
[
"Bolles",
"Tim",
""
],
[
"Köhn",
"Kevin",
""
],
[
"Klich",
"Maximilian",
""
],
[
"Brinkmann",
"Ralf Peter",
""
],
[
"Mussenbrock",
"Thomas",
""
]
] |
2505.22462 | Somnath Jana Dr. | Somnath Jana, Ronny Knut, Dima Afanasiev, Niko Pontius, Christian Sch\"u{\ss}ler-Langeheine, Christian Tzschaschel, Daniel Schick, Alexey V. Kimel, Olof Karis, Clemens von Korff Schmising, and Stefan Eisebitt | X-ray View of Light-Induced Spin Reorientation in TmFeO$_{3}$: Direct Observation of a 90$^\circ$ N\'eel Vector Rotation | 6 pages, 4 figures | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Using time-resolved X-ray magnetic linear dichroism in reflection, we provide a direct probe of the N\'eel vector dynamics in TmFeO$_3$ on a ultrafast timescale. Our measurements reveal that, following optical excitation, the N\'eel vector undergoes a spin reorientation transition primarily within the a-c plane, completing a full 90{\deg} rotation within approximately 20 ps. This study highlights the ability to probe dynamics of antiferromagnets at its intrinsic timescale in reflection geometry, paving the way for investigations of a wide range of antiferromagnets grown on application relevant substrates. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:17:12 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jana",
"Somnath",
""
],
[
"Knut",
"Ronny",
""
],
[
"Afanasiev",
"Dima",
""
],
[
"Pontius",
"Niko",
""
],
[
"Schüßler-Langeheine",
"Christian",
""
],
[
"Tzschaschel",
"Christian",
""
],
[
"Schick",
"Daniel",
""
],
[
"Kimel",
"Alexey V.",
""
],
[
"Karis",
"Olof",
""
],
[
"Schmising",
"Clemens von Korff",
""
],
[
"Eisebitt",
"Stefan",
""
]
] |
2505.22463 | Aaroodd Ujjayini Ramachandran | P. S. Bhupal Dev, Bhaskar Dutta, Srubabati Goswami, Jianrong (Paul) Tang, Aaroodd Ujjayini Ramachandran | Opening up New Parameter Space for Sterile Neutrino Dark Matter | 10 pages, 6 figures | null | null | MI-HET-858 | hep-ph astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | Sterile neutrinos are compelling dark matter (DM) candidates, yet the minimal production mechanism solely based on active ($\nu_a$)-sterile ($\nu_s$) oscillations is excluded by astrophysical observations. Non-standard self-interactions in either active ($\nu_a-\nu_a$) or sterile ($\nu_s-\nu_s$) sector are known to alter the sterile neutrino DM production in the early Universe, which could alleviate the tension with astrophysical constraints to some extent. Here we propose a novel solution where scalar-mediated non-standard interactions between active and sterile neutrinos ($\nu_a-\nu_s$) generate new production channels for $\nu_s$, independent of the active-sterile mixing and without the need for any fine-tuned resonance or primordial lepton asymmetry. This framework enables efficient sterile neutrino DM production even at vanishingly small mixing angles and opens up new viable regions of parameter space that can be tested with future $X$-ray and gamma-ray observations. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:17:25 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Dev",
"P. S. Bhupal",
"",
"Paul"
],
[
"Dutta",
"Bhaskar",
"",
"Paul"
],
[
"Goswami",
"Srubabati",
"",
"Paul"
],
[
"Jianrong",
"",
"",
"Paul"
],
[
"Tang",
"",
""
],
[
"Ramachandran",
"Aaroodd Ujjayini",
""
]
] |
2505.22464 | Jonas Knoerr | Jonas Knoerr | A Paley-Wiener-Schwartz Theorem for smooth valuations on convex functions | 54 pages, comments welcome | null | null | null | math.FA math.MG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Continuous dually epi-translation invariant valuations on convex functions are characterized in terms of the Fourier-Laplace transform of the associated Goodey-Weil distributions. This description is used to obtain integral representations of the smooth vectors of the natural representation of the group of translations on the space of these valuations. As an application, a complete classification of all closed and affine invariant subspaces is established, yielding density results for valuations defined in terms of mixed Monge-Amp\`ere operators. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:18:04 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Knoerr",
"Jonas",
""
]
] |
2505.22465 | Zobia Batool | Zobia Batool, Huseyin Ozkan, Erchan Aptoula | Single Domain Generalization for Alzheimer's Detection from 3D MRIs with Pseudo-Morphological Augmentations and Contrastive Learning | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Although Alzheimer's disease detection via MRIs has advanced significantly thanks to contemporary deep learning models, challenges such as class imbalance, protocol variations, and limited dataset diversity often hinder their generalization capacity. To address this issue, this article focuses on the single domain generalization setting, where given the data of one domain, a model is designed and developed with maximal performance w.r.t. an unseen domain of distinct distribution. Since brain morphology is known to play a crucial role in Alzheimer's diagnosis, we propose the use of learnable pseudo-morphological modules aimed at producing shape-aware, anatomically meaningful class-specific augmentations in combination with a supervised contrastive learning module to extract robust class-specific representations. Experiments conducted across three datasets show improved performance and generalization capacity, especially under class imbalance and imaging protocol variations. The source code will be made available upon acceptance at https://github.com/zobia111/SDG-Alzheimer. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:18:16 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Batool",
"Zobia",
""
],
[
"Ozkan",
"Huseyin",
""
],
[
"Aptoula",
"Erchan",
""
]
] |
2505.22466 | Timothy Burke | Timothy J. Burke, Xiaoyang Shi, Jasmine Sinanan-Singh, Isaac L. Chuang, John Chiaverini | Spontaneous Raman scattering from metastable states of Ba$^+$ | 6 pages, 4 figures | null | null | null | quant-ph physics.atom-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum logic gates performed via two-photon stimulated-Raman transitions in ions and atoms are fundamentally limited by spontaneous scattering errors. Recent theoretical treatment of these scattering processes has predicted no lower bound on the error rate of such gates when implemented with far-detuned lasers, while also providing an extension to metastable qubits. To validate this theoretical model, we provide experimental measurements of Raman scattering rates due to near-, and far-detuned lasers for initial states in the metastable D$_{5/2}$ level of $^{137}$Ba$^+$. The measured spontaneous Raman scattering rate is consistent with the theoretical prediction and suggests that metastable-level two-qubit gates with an error rate $\approx10^{-4}$ are possible with laser excitation detuned by tens of terahertz or more. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:19:21 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Burke",
"Timothy J.",
""
],
[
"Shi",
"Xiaoyang",
""
],
[
"Sinanan-Singh",
"Jasmine",
""
],
[
"Chuang",
"Isaac L.",
""
],
[
"Chiaverini",
"John",
""
]
] |
2505.22468 | Lo\"ic Marchesini | Marianne Akian, St\'ephane Gaubert, Lo\"ic Marchesini, Ian Morris | Continuity and approximability of competitive spectral radii | 8 pages | null | null | null | math.OC cs.NA math.DS math.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The competitive spectral radius extends the notion of joint spectral radius to the two-player case: two players alternatively select matrices in prescribed compact sets, resulting in an infinite matrix product; one player wishes to maximize the growth rate of this product, whereas the other player wishes to minimize it. We show that when the matrices represent linear operators preserving a cone and satisfying a "strict positivity" assumption, the competitive spectral radius depends continuously - and even in a Lipschitz-continuous way - on the matrix sets. Moreover, we show that the competive spectral radius can be approximated up to any accuracy. This relies on the solution of a discretized infinite dimensional non-linear eigenproblem. We illustrate the approach with an example of age-structured population dynamics. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:20:25 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Akian",
"Marianne",
""
],
[
"Gaubert",
"Stéphane",
""
],
[
"Marchesini",
"Loïc",
""
],
[
"Morris",
"Ian",
""
]
] |
2505.22469 | Mohamed Elshamy | Mohamed R. Elshamy, Mehdi Elahi, Ahmad Patooghy, and Abdel-Hameed A. Badawy | CPINN-ABPI: Physics-Informed Neural Networks for Accurate Power Estimation in MPSoCs | null | null | null | null | cs.PF cs.LG | http://creativecommons.org/licenses/by/4.0/ | Efficient thermal and power management in modern multiprocessor systems-on-chip (MPSoCs) demands accurate power consumption estimation. One of the state-of-the-art approaches, Alternative Blind Power Identification (ABPI), theoretically eliminates the dependence on steady-state temperatures, addressing a major shortcoming of previous approaches. However, ABPI performance has remained unverified in actual hardware implementations. In this study, we conduct the first empirical validation of ABPI on commercial hardware using the NVIDIA Jetson Xavier AGX platform. Our findings reveal that, while ABPI provides computational efficiency and independence from steady-state temperature, it exhibits considerable accuracy deficiencies in real-world scenarios. To overcome these limitations, we introduce a novel approach that integrates Custom Physics-Informed Neural Networks (CPINNs) with the underlying thermal model of ABPI. Our approach employs a specialized loss function that harmonizes physical principles with data-driven learning, complemented by multi-objective genetic algorithm optimization to balance estimation accuracy and computational cost. In experimental validation, CPINN-ABPI achieves a reduction of 84.7\% CPU and 73.9\% GPU in the mean absolute error (MAE) relative to ABPI, with the weighted mean absolute percentage error (WMAPE) improving from 47\%--81\% to $\sim$12\%. The method maintains real-time performance with 195.3~$\mu$s of inference time, with similar 85\%--99\% accuracy gains across heterogeneous SoCs. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:22:15 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Elshamy",
"Mohamed R.",
""
],
[
"Elahi",
"Mehdi",
""
],
[
"Patooghy",
"Ahmad",
""
],
[
"Badawy",
"Abdel-Hameed A.",
""
]
] |
2505.22470 | Stevan Gajovi\'c | Stevan Gajovi\'c and Sun Woo Park | Infinitely many genus two and three curves of small fixed positive rank | 8 pages, comments welcome | null | null | MPIM-Bonn-2025 | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove that there exist infinitely many genus two curves over $\mathbb{Q}$ whose Jacobians have ranks between $0$ and $11$, and genus three curves over $\mathbb{Q}$ whose Jacobians have ranks $0$, $1$, and $2$. Furthermore, we consider a generalisation of these results over number fields satisfying some conditions. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:22:40 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Gajović",
"Stevan",
""
],
[
"Park",
"Sun Woo",
""
]
] |
2505.22471 | Christian M\"onch | Benedikt Jahnel, Lukas L\"uchtrath and Christian M\"onch | Phase transitions for contact processes on sparse random graphs via metastability and local limits | null | null | null | null | math.PR | http://creativecommons.org/licenses/by/4.0/ | We propose a new perspective on the asymptotic regimes of fast and slow extinction in the contact process on locally converging sequences of sparse finite graphs. We characterise the phase boundary by the existence of a metastable density, which makes the study of the phase transition particularly amenable to local-convergence techniques. We use this approach to derive general conditions for the coincidence of the critical threshold with the survival/extinction threshold in the local limit. We further argue that the correct time scale to separate fast extinction from slow extinction in sparse graphs is, in general, the exponential scale, by showing that fast extinction may occur on stretched exponential time scales in sparse scale-free spatial networks. Together with recent results by Nam, Nguyen and Sly (Trans. Am. Math. Soc. 375, 2022), our methods can be applied to deduce that the fast/slow threshold in sparse configuration models coincides with the survival/extinction threshold on the limiting Galton-Watson tree. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:23:13 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jahnel",
"Benedikt",
""
],
[
"Lüchtrath",
"Lukas",
""
],
[
"Mönch",
"Christian",
""
]
] |
2505.22472 | Matteo Acciai | Matteo Acciai, Liliana Arrachea, Janine Splettstoesser | Quantum transport phenomena induced by time-dependent fields | Review article, 82 pages plus references | null | null | null | cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present an overview of time-dependent transport phenomena in quantum systems, with a particular emphasis on steady-state regimes. We present the ideas after the main theoretical frameworks to study open-quantum systems out of equilibrium, that are useful to study quantum transport under time-dependent driving. We discuss the fundamentals of the key mechanisms such as dissipation, quantum pumping, noise, and energy conversion that are associated to the problem of quantum transport.
Our primary focus is on electronic systems, where decades of research have established a rich theoretical foundation and a wealth of experimental realizations. Topics of interest include quantum optics with electrons, high-precision electron spectroscopy, quantum electrical metrology, and the critical role of quantum fluctuations in transport and thermodynamics. We also extend the discussion to atomic, molecular, and optical systems, as well as nanomechanical platforms, which offer complementary perspectives and are currently experiencing rapid experimental development. Finally, we examine the intersection of time-dependent transport and topological matter, a domain of active investigation.
This review aims to gather the diverse approaches and emerging trends that define the current landscape of quantum transport research under time-dependent conditions, bridging theoretical insights with experimental advances across multiple physical platforms. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:23:15 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Acciai",
"Matteo",
""
],
[
"Arrachea",
"Liliana",
""
],
[
"Splettstoesser",
"Janine",
""
]
] |
2505.22473 | Riccardo Poiani | Riccardo Poiani, Martino Bernasconi, Andrea Celli | Pure Exploration with Infinite Answers | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We study pure exploration problems where the set of correct answers is possibly infinite, e.g., the regression of any continuous function of the means of the bandit. We derive an instance-dependent lower bound for these problems. By analyzing it, we discuss why existing methods (i.e., Sticky Track-and-Stop) for finite answer problems fail at being asymptotically optimal in this more general setting. Finally, we present a framework, Sticky-Sequence Track-and-Stop, which generalizes both Track-and-Stop and Sticky Track-and-Stop, and that enjoys asymptotic optimality. Due to its generality, our analysis also highlights special cases where existing methods enjoy optimality. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:23:36 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Poiani",
"Riccardo",
""
],
[
"Bernasconi",
"Martino",
""
],
[
"Celli",
"Andrea",
""
]
] |
2505.22474 | Amirhossein Sohrabbeig | Amirhossein Sohrabbeig, Omid Ardakanian, and Petr Musilek | Forecasting Multivariate Urban Data via Decomposition and Spatio-Temporal Graph Analysis | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The forecasting of multivariate urban data presents a complex challenge due to the intricate dependencies between various urban metrics such as weather, air pollution, carbon intensity, and energy demand. This paper introduces a novel multivariate time-series forecasting model that utilizes advanced Graph Neural Networks (GNNs) to capture spatial dependencies among different time-series variables. The proposed model incorporates a decomposition-based preprocessing step, isolating trend, seasonal, and residual components to enhance the accuracy and interpretability of forecasts. By leveraging the dynamic capabilities of GNNs, the model effectively captures interdependencies and improves the forecasting performance. Extensive experiments on real-world datasets, including electricity usage, weather metrics, carbon intensity, and air pollution data, demonstrate the effectiveness of the proposed approach across various forecasting scenarios. The results highlight the potential of the model to optimize smart infrastructure systems, contributing to energy-efficient urban development and enhanced public well-being. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:24:04 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Sohrabbeig",
"Amirhossein",
""
],
[
"Ardakanian",
"Omid",
""
],
[
"Musilek",
"Petr",
""
]
] |
2505.22475 | Riccardo Poiani | Riccardo Poiani, Martino Bernasconi, Andrea Celli | Non-Asymptotic Analysis of (Sticky) Track-and-Stop | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | In pure exploration problems, a statistician sequentially collects information to answer a question about some stochastic and unknown environment. The probability of returning a wrong answer should not exceed a maximum risk parameter $\delta$ and good algorithms make as few queries to the environment as possible. The Track-and-Stop algorithm is a pioneering method to solve these problems. Specifically, it is well-known that it enjoys asymptotic optimality sample complexity guarantees for $\delta\to 0$ whenever the map from the environment to its correct answers is single-valued (e.g., best-arm identification with a unique optimal arm). The Sticky Track-and-Stop algorithm extends these results to settings where, for each environment, there might exist multiple correct answers (e.g., $\epsilon$-optimal arm identification). Although both methods are optimal in the asymptotic regime, their non-asymptotic guarantees remain unknown. In this work, we fill this gap and provide non-asymptotic guarantees for both algorithms. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:26:55 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Poiani",
"Riccardo",
""
],
[
"Bernasconi",
"Martino",
""
],
[
"Celli",
"Andrea",
""
]
] |
2505.22476 | Hannes Jung | J. Altmann, U. Amaldi, M. Barone, A. Bassalat, M. Bona, J. Beullens, H. Brand, S. Brentjes, D. Britzger, J. Ellis, S. Franchoo, A. Giammanco, A. Glazov, C. Heck, H. Jung, S. Kraml, L. L\"onnblad, M. Mangano, M. Renneberg, Th. Riebe, A. Sabio-Vera, R. Sanders, J. Scheffran, M. Schmelling, T. Schucker, T. Suzuki, A. Tanasijczuk, I. Tsakov, D. Valls-Gabaud, M. Walker | Science for Peace and the need for Civil Clauses at universities and civilian research institutions | null | null | null | null | physics.soc-ph hep-ex hep-ph | http://creativecommons.org/licenses/by/4.0/ | After the end of World War II, the commitment to confine scientific activities in universities and research institutions to peaceful and civilian purposes has entered, in the form of {\it Civil Clauses}, the charters of many research institutions and universities. In the wake of recent world events, the relevance and scope of such Civil Clauses has been questioned in reports issued by some governments and by the EU Commission, a development that opens the door to a possible blurring of the distinction between peaceful and military research.
This paper documents the reflections stimulated by a panel discussion on this issue recently organized by the Science4Peace Forum. We review the adoptions of Civil Clauses in research organizations and institutions in various countries, present evidence of the challenges that are emerging to such Civil Clauses, and collect arguments in favour of maintaining the purely civilian and peaceful focus of public (non-military) research. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:27:22 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Altmann",
"J.",
""
],
[
"Amaldi",
"U.",
""
],
[
"Barone",
"M.",
""
],
[
"Bassalat",
"A.",
""
],
[
"Bona",
"M.",
""
],
[
"Beullens",
"J.",
""
],
[
"Brand",
"H.",
""
],
[
"Brentjes",
"S.",
""
],
[
"Britzger",
"D.",
""
],
[
"Ellis",
"J.",
""
],
[
"Franchoo",
"S.",
""
],
[
"Giammanco",
"A.",
""
],
[
"Glazov",
"A.",
""
],
[
"Heck",
"C.",
""
],
[
"Jung",
"H.",
""
],
[
"Kraml",
"S.",
""
],
[
"Lönnblad",
"L.",
""
],
[
"Mangano",
"M.",
""
],
[
"Renneberg",
"M.",
""
],
[
"Riebe",
"Th.",
""
],
[
"Sabio-Vera",
"A.",
""
],
[
"Sanders",
"R.",
""
],
[
"Scheffran",
"J.",
""
],
[
"Schmelling",
"M.",
""
],
[
"Schucker",
"T.",
""
],
[
"Suzuki",
"T.",
""
],
[
"Tanasijczuk",
"A.",
""
],
[
"Tsakov",
"I.",
""
],
[
"Valls-Gabaud",
"D.",
""
],
[
"Walker",
"M.",
""
]
] |
2505.22477 | Qi Gao | Qi Gao, Wei Xu, Hanxi Pan, Mowei Shen, Zaifeng Gao | Human-Centered Human-AI Collaboration (HCHAC) | This article is a chapter from the upcoming book Handbook of Human-Centered Artificial Intelligence | null | null | null | cs.HC cs.AI cs.CY | http://creativecommons.org/licenses/by/4.0/ | In the intelligent era, the interaction between humans and intelligent systems fundamentally involves collaboration with autonomous intelligent agents. Human-AI Collaboration (HAC) represents a novel type of human-machine relationship facilitated by autonomous intelligent machines equipped with AI technologies. In this paradigm, AI agents serve not only as auxiliary tools but also as active teammates, partnering with humans to accomplish tasks collaboratively. Human-centered AI (HCAI) emphasizes that humans play critical leadership roles in the collaboration. This human-led collaboration imparts new dimensions to the human-machine relationship, necessitating innovative research perspectives, paradigms, and agenda to address the unique challenges posed by HAC. This chapter delves into the essence of HAC from the human-centered perspective, outlining its core concepts and distinguishing features. It reviews the current research methodologies and research agenda within the HAC field from the HCAI perspective, highlighting advancements and ongoing studies. Furthermore, a framework for human-centered HAC (HCHAC) is proposed by integrating these reviews and analyses. A case study of HAC in the context of autonomous vehicles is provided, illustrating practical applications and the synergistic interactions between humans and AI agents. Finally, it identifies potential future research directions aimed at enhancing the effectiveness, reliability, and ethical integration of human-centered HAC systems in diverse domains. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:27:52 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Gao",
"Qi",
""
],
[
"Xu",
"Wei",
""
],
[
"Pan",
"Hanxi",
""
],
[
"Shen",
"Mowei",
""
],
[
"Gao",
"Zaifeng",
""
]
] |
2505.22478 | Bjoern Bringmann | Bjoern Bringmann and Gigliola Staffilani | Invariant Gibbs measures for the one-dimensional quintic nonlinear Schr\"odinger equation in infinite volume | 40 pages | null | null | null | math.AP math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove the invariance of the Gibbs measure for the defocusing quintic nonlinear Schr\"odinger equation on the real line. This builds on earlier work by Bourgain, who treated the cubic nonlinearity. The key new ingredient is a growth estimate for the infinite-volume $\Phi^{p+1}_1$-measures, which is proven via the stochastic quantization method. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:28:32 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Bringmann",
"Bjoern",
""
],
[
"Staffilani",
"Gigliola",
""
]
] |
2505.22479 | Chamila Malagoda Gamage | Chamila Malagoda Gamage | Impact of Pre-Assessment and Post-Assessment in an Introductory Real Analysis Course | null | null | null | null | math.HO | http://creativecommons.org/licenses/by/4.0/ | This study explores how pre- and post-assessments shape learning outcomes in an Introductory Real Analysis course. Pre-assessments act as learning roadmaps, highlighting prior knowledge and guiding student focus, while post-assessments measure growth and conceptual mastery. By analyzing student performance and feedback, we assess their impact on engagement, self-efficacy, and deeper mathematical understanding. The findings offer valuable insights for enhancing instructional strategies and fostering a more effective, student-centered learning experience in advanced mathematics. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:28:52 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Gamage",
"Chamila Malagoda",
""
]
] |
2505.22480 | Francesco Carlo Massa | Lorenzo Botti, Daniele Antonio Di Pietro and Francesco Carlo Massa | Hybrid High-Order formulations with turbulence modelling capabilities for incompressible flow problems | null | null | null | null | physics.flu-dyn cs.NA math.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a Hybrid High-Order (HHO) formulation of the incompressible Navier--Stokes equations, that is well suited to be employed for the simulation of turbulent flows. The spatial discretization relies on hybrid velocity and pressure spaces and the temporal discretization is based on Explicit Singly Diagonal Implicit Runge-Kutta (ESDIRK) methods. The formulation possesses some attractive features that can be fruitfully exploited when high-fidelity computations are required, namely: pressure-robustness, conservation of mass enforced cell-by-cell up to machine precision, robustness in the inviscid limit, implicit high-order accurate time stepping with local time step adaptation, reduced memory footprint thanks to static condensation of both velocity and pressure, possibility to exploit inherited $p$-multilevel solution strategies to improve performance of iterative solvers. After demonstrating the relevant properties of the scheme in practice, performing challenging 2D and 3D test cases, we consider the simulation of the Taylor--Green Vortex flow problem at Reynolds 1600. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:29:00 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Botti",
"Lorenzo",
""
],
[
"Di Pietro",
"Daniele Antonio",
""
],
[
"Massa",
"Francesco Carlo",
""
]
] |
2505.22481 | Yiming Xi | Yiming Xi, Konstantinos Zygalakis, Marcelo Pereyra | Hypothesis Testing in Imaging Inverse Problems | null | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper proposes a framework for semantic hypothesis testing tailored to imaging inverse problems. Modern imaging methods struggle to support hypothesis testing, a core component of the scientific method that is essential for the rigorous interpretation of experiments and robust interfacing with decision-making processes. There are three main reasons why image-based hypothesis testing is challenging. First, the difficulty of using a single observation to simultaneously reconstruct an image, formulate hypotheses, and quantify their statistical significance. Second, the hypotheses encountered in imaging are mostly of semantic nature, rather than quantitative statements about pixel values. Third, it is challenging to control test error probabilities because the null and alternative distributions are often unknown. Our proposed approach addresses these difficulties by leveraging concepts from self-supervised computational imaging, vision-language models, and non-parametric hypothesis testing with e-values. We demonstrate our proposed framework through numerical experiments related to image-based phenotyping, where we achieve excellent power while robustly controlling Type I errors. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:29:43 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Xi",
"Yiming",
""
],
[
"Zygalakis",
"Konstantinos",
""
],
[
"Pereyra",
"Marcelo",
""
]
] |
2505.22482 | Dani\"el Cox | Dani\"el W.S. Cox, Harish Sasikumar and Ivo M. Vellekoop | Inline calibration of spatial light modulators in nonlinear microscopy | 10 pages, 5 figures | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | We present a method for calibrating the response of a phase-only spatial light modulator in nonlinear microscopy. Our method uses the microscope image itself as calibration measurement and requires no additional hardware components. Our method is adapted to the nonlinear signals encountered in multi-photon excitation fluorescence microscopes, and works well even under low light conditions and with strong photobleaching. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:31:30 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Cox",
"Daniël W. S.",
""
],
[
"Sasikumar",
"Harish",
""
],
[
"Vellekoop",
"Ivo M.",
""
]
] |
2505.22483 | Abhra Chaudhuri | Abhra Chaudhuri and Anjan Dutta and Tu Bui and Serban Georgescu | A Closer Look at Multimodal Representation Collapse | International Conference on Machine Learning (ICML) 2025 (Spotlight) | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | We aim to develop a fundamental understanding of modality collapse, a recently observed empirical phenomenon wherein models trained for multimodal fusion tend to rely only on a subset of the modalities, ignoring the rest. We show that modality collapse happens when noisy features from one modality are entangled, via a shared set of neurons in the fusion head, with predictive features from another, effectively masking out positive contributions from the predictive features of the former modality and leading to its collapse. We further prove that cross-modal knowledge distillation implicitly disentangles such representations by freeing up rank bottlenecks in the student encoder, denoising the fusion-head outputs without negatively impacting the predictive features from either modality. Based on the above findings, we propose an algorithm that prevents modality collapse through explicit basis reallocation, with applications in dealing with missing modalities. Extensive experiments on multiple multimodal benchmarks validate our theoretical claims. Project page: https://abhrac.github.io/mmcollapse/. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:31:53 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Chaudhuri",
"Abhra",
""
],
[
"Dutta",
"Anjan",
""
],
[
"Bui",
"Tu",
""
],
[
"Georgescu",
"Serban",
""
]
] |
2505.22484 | Fabiano Andrade | Eduardo K. Soares and G. D. de Moraes Neto and Fabiano M. Andrade | Comparative analysis of robust entanglement generation in engineered XX spin chains | 9 pages, 8 figures | null | null | null | quant-ph math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a numerical investigation comparing two entanglement generation protocols in finite XX spin chains with varying spin magnitudes ($s = 1/2, 1, 3/2 $). Protocol 1 (P1) relies on staggered couplings to steer correlations toward the ends of the chain. At the same time, Protocol 2 (P2) adopts a dual-port architecture that uses optimized boundary fields to mediate virtual excitations between terminal spins. Our results show that P2 consistently outperforms P1 in all spin values, generating higher-fidelity entanglement in shorter timescales when evaluated under the same system parameters. Furthermore, P2 exhibits superior robustness under realistic imperfections, including diagonal and off-diagonal disorder, as well as dephasing noise. These advantages stem from its ability to suppress the bulk population and minimize susceptibility to decoherence. Together, the scalability, efficiency, and noise resilience of the dual-port approach position it as a promising framework for entanglement distribution in solid-state quantum information platforms. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:35:41 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Soares",
"Eduardo K.",
""
],
[
"Neto",
"G. D. de Moraes",
""
],
[
"Andrade",
"Fabiano M.",
""
]
] |
2505.22485 | Adam Arras | Adam Arras | Random Schr\"odinger operators and convolution on wreath products | null | null | null | null | math.PR math.GR math.SP | http://creativecommons.org/licenses/by/4.0/ | We establish a spectral correspondence between random Schr\"odinger operators and deterministic convolution operators on wreath products, generalizing previous results that relate Lamplighter groups to Schr\"odinger operators with Bernoulli potentials. Using this correspondence in both directions, we obtain an elementary criterion for the absolute continuity of convolutions on wreath products, Lifschitz tail estimates for Schr\"odinger operators on Cayley graphs of polynomial growth, and an exact formula for the second moment of the Green function, expressed in terms of the wreath product with an Abelian group of lamps. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:35:45 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Arras",
"Adam",
""
]
] |
2505.22486 | Mujtaba Hussain Mirza | Mujtaba Hussain Mirza, Maria Rosaria Briglia, Filippo Bartolucci, Senad Beadini, Giuseppe Lisanti, Iacopo Masi | Understanding Adversarial Training with Energy-based Models | Under review for TPAMI | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We aim at using Energy-based Model (EBM) framework to better understand adversarial training (AT) in classifiers, and additionally to analyze the intrinsic generative capabilities of robust classifiers. By viewing standard classifiers through an energy lens, we begin by analyzing how the energies of adversarial examples, generated by various attacks, differ from those of the natural samples. The central focus of our work is to understand the critical phenomena of Catastrophic Overfitting (CO) and Robust Overfitting (RO) in AT from an energy perspective. We analyze the impact of existing AT approaches on the energy of samples during training and observe that the behavior of the ``delta energy' -- change in energy between original sample and its adversarial counterpart -- diverges significantly when CO or RO occurs. After a thorough analysis of these energy dynamics and their relationship with overfitting, we propose a novel regularizer, the Delta Energy Regularizer (DER), designed to smoothen the energy landscape during training. We demonstrate that DER is effective in mitigating both CO and RO across multiple benchmarks. We further show that robust classifiers, when being used as generative models, have limits in handling trade-off between image quality and variability. We propose an improved technique based on a local class-wise principal component analysis (PCA) and energy-based guidance for better class-specific initialization and adaptive stopping, enhancing sample diversity and generation quality. Considering that we do not explicitly train for generative modeling, we achieve a competitive Inception Score (IS) and Fr\'echet inception distance (FID) compared to hybrid discriminative-generative models. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:36:02 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Mirza",
"Mujtaba Hussain",
""
],
[
"Briglia",
"Maria Rosaria",
""
],
[
"Bartolucci",
"Filippo",
""
],
[
"Beadini",
"Senad",
""
],
[
"Lisanti",
"Giuseppe",
""
],
[
"Masi",
"Iacopo",
""
]
] |
2505.22487 | Yen Meng | Yen Meng, Sharon Goldwater, Hao Tang | Effective Context in Neural Speech Models | Accepted to Interspeech 2025 | null | null | null | cs.SD cs.CL eess.AS | http://creativecommons.org/licenses/by/4.0/ | Modern neural speech models benefit from having longer context, and many approaches have been proposed to increase the maximum context a model can use. However, few have attempted to measure how much context these models actually use, i.e., the effective context. Here, we propose two approaches to measuring the effective context, and use them to analyze different speech Transformers. For supervised models, we find that the effective context correlates well with the nature of the task, with fundamental frequency tracking, phone classification, and word classification requiring increasing amounts of effective context. For self-supervised models, we find that effective context increases mainly in the early layers, and remains relatively short -- similar to the supervised phone model. Given that these models do not use a long context during prediction, we show that HuBERT can be run in streaming mode without modification to the architecture and without further fine-tuning. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:36:44 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Meng",
"Yen",
""
],
[
"Goldwater",
"Sharon",
""
],
[
"Tang",
"Hao",
""
]
] |
2505.22488 | Gakuto Kusuno | Gakuto Kusuno, Takeshi Hayashida, Takayuki Nagai, Hikaru Watanabe, Rikuto Oiwa, Tsuyoshi Kimura, and Takuya Satoh | Raman Optical Activity Induced by Ferroaxial Order in $\textrm{NiTiO}_3$ | 16 pages and 4 figures | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Raman optical activity (ROA) -- the dependence of Raman scattered light intensity on the circular polarization of incident and scattered light -- has traditionally been associated with chiral molecules and magnetic materials. In this study, we demonstrate that ROA can also arise in ferroaxial materials that possess spatial inversion and time-reversal symmetries. Using circularly polarized Raman spectroscopy on single-crystalline $\textrm{NiTiO}_3$, we observed a pronounced ROA signal in the cross-circular polarization configuration, which correlates with the ferroaxial domain structure. Our symmetry analysis and tight-binding model calculations reveal that the natural ROA (NROA) originates from the ferroaxial order and persists even within the electric dipole approximation. These results establish ROA as a powerful probe of ferroaxial order in centrosymmetric systems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:36:50 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Kusuno",
"Gakuto",
""
],
[
"Hayashida",
"Takeshi",
""
],
[
"Nagai",
"Takayuki",
""
],
[
"Watanabe",
"Hikaru",
""
],
[
"Oiwa",
"Rikuto",
""
],
[
"Kimura",
"Tsuyoshi",
""
],
[
"Satoh",
"Takuya",
""
]
] |
2505.22489 | Siyeop Yoon | Siyeop Yoon, Sifan Song, Pengfei Jin, Matthew Tivnan, Yujin Oh, Sekeun Kim, Dufan Wu, Xiang Li, Quanzheng Li | Cascaded 3D Diffusion Models for Whole-body 3D 18-F FDG PET/CT synthesis from Demographics | MICCAI2025 Submitted version | null | null | null | eess.IV cs.CV cs.GR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We propose a cascaded 3D diffusion model framework to synthesize high-fidelity 3D PET/CT volumes directly from demographic variables, addressing the growing need for realistic digital twins in oncologic imaging, virtual trials, and AI-driven data augmentation. Unlike deterministic phantoms, which rely on predefined anatomical and metabolic templates, our method employs a two-stage generative process. An initial score-based diffusion model synthesizes low-resolution PET/CT volumes from demographic variables alone, providing global anatomical structures and approximate metabolic activity. This is followed by a super-resolution residual diffusion model that refines spatial resolution. Our framework was trained on 18-F FDG PET/CT scans from the AutoPET dataset and evaluated using organ-wise volume and standardized uptake value (SUV) distributions, comparing synthetic and real data between demographic subgroups. The organ-wise comparison demonstrated strong concordance between synthetic and real images. In particular, most deviations in metabolic uptake values remained within 3-5% of the ground truth in subgroup analysis. These findings highlight the potential of cascaded 3D diffusion models to generate anatomically and metabolically accurate PET/CT images, offering a robust alternative to traditional phantoms and enabling scalable, population-informed synthetic imaging for clinical and research applications. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:38:33 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Yoon",
"Siyeop",
""
],
[
"Song",
"Sifan",
""
],
[
"Jin",
"Pengfei",
""
],
[
"Tivnan",
"Matthew",
""
],
[
"Oh",
"Yujin",
""
],
[
"Kim",
"Sekeun",
""
],
[
"Wu",
"Dufan",
""
],
[
"Li",
"Xiang",
""
],
[
"Li",
"Quanzheng",
""
]
] |
2505.22490 | Ke Zhang | Ke Zhang, Tianyu Ding, Jiachen Jiang, Tianyi Chen, Ilya Zharkov, Vishal M. Patel, Luming Liang | ProCrop: Learning Aesthetic Image Cropping from Professional Compositions | 16 pages, 15 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Image cropping is crucial for enhancing the visual appeal and narrative impact of photographs, yet existing rule-based and data-driven approaches often lack diversity or require annotated training data. We introduce ProCrop, a retrieval-based method that leverages professional photography to guide cropping decisions. By fusing features from professional photographs with those of the query image, ProCrop learns from professional compositions, significantly boosting performance. Additionally, we present a large-scale dataset of 242K weakly-annotated images, generated by out-painting professional images and iteratively refining diverse crop proposals. This composition-aware dataset generation offers diverse high-quality crop proposals guided by aesthetic principles and becomes the largest publicly available dataset for image cropping. Extensive experiments show that ProCrop significantly outperforms existing methods in both supervised and weakly-supervised settings. Notably, when trained on the new dataset, our ProCrop surpasses previous weakly-supervised methods and even matches fully supervised approaches. Both the code and dataset will be made publicly available to advance research in image aesthetics and composition analysis. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:38:44 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Ke",
""
],
[
"Ding",
"Tianyu",
""
],
[
"Jiang",
"Jiachen",
""
],
[
"Chen",
"Tianyi",
""
],
[
"Zharkov",
"Ilya",
""
],
[
"Patel",
"Vishal M.",
""
],
[
"Liang",
"Luming",
""
]
] |
2505.22491 | Moritz Haas | Moritz Haas, Sebastian Bordt, Ulrike von Luxburg, Leena Chennuru Vankadara | On the Surprising Effectiveness of Large Learning Rates under Standard Width Scaling | null | null | null | null | cs.LG cs.AI stat.ML | http://creativecommons.org/licenses/by-sa/4.0/ | The dominant paradigm for training large-scale vision and language models is He initialization and a single global learning rate (\textit{standard parameterization}, SP). Despite its practical success, standard parametrization remains poorly understood from a theoretical perspective: Existing infinite-width theory would predict instability under large learning rates and vanishing feature learning under stable learning rates. However, empirically optimal learning rates consistently decay much slower than theoretically predicted. By carefully studying neural network training dynamics, we demonstrate that this discrepancy is not fully explained by finite-width phenomena such as catapult effects or a lack of alignment between weights and incoming activations. We instead show that the apparent contradiction can be fundamentally resolved by taking the loss function into account: In contrast to Mean Squared Error (MSE) loss, we prove that under cross-entropy (CE) loss, an intermediate \textit{controlled divergence} regime emerges, where logits diverge but loss, gradients, and activations remain stable. Stable training under large learning rates enables persistent feature evolution at scale in all hidden layers, which is crucial for the practical success of SP. In experiments across optimizers (SGD, Adam), architectures (MLPs, GPT) and data modalities (vision, language), we validate that neural networks operate in this controlled divergence regime under CE loss but not under MSE loss. Our empirical evidence suggests that width-scaling considerations are surprisingly useful for predicting empirically optimal learning rate exponents. Finally, our analysis clarifies the effectiveness and limitations of recently proposed layerwise learning rate scalings for standard initialization. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:40:48 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Haas",
"Moritz",
""
],
[
"Bordt",
"Sebastian",
""
],
[
"von Luxburg",
"Ulrike",
""
],
[
"Vankadara",
"Leena Chennuru",
""
]
] |
2505.22492 | Jin Zhu | Hongyi Zhou, Josiah P. Hanna, Jin Zhu, Ying Yang, Chengchun Shi | Demystifying the Paradox of Importance Sampling with an Estimated History-Dependent Behavior Policy in Off-Policy Evaluation | Accepted by ICML 2025 | null | null | null | cs.LG cs.AI stat.ML | http://creativecommons.org/licenses/by/4.0/ | This paper studies off-policy evaluation (OPE) in reinforcement learning with a focus on behavior policy estimation for importance sampling. Prior work has shown empirically that estimating a history-dependent behavior policy can lead to lower mean squared error (MSE) even when the true behavior policy is Markovian. However, the question of why the use of history should lower MSE remains open. In this paper, we theoretically demystify this paradox by deriving a bias-variance decomposition of the MSE of ordinary importance sampling (IS) estimators, demonstrating that history-dependent behavior policy estimation decreases their asymptotic variances while increasing their finite-sample biases. Additionally, as the estimated behavior policy conditions on a longer history, we show a consistent decrease in variance. We extend these findings to a range of other OPE estimators, including the sequential IS estimator, the doubly robust estimator and the marginalized IS estimator, with the behavior policy estimated either parametrically or non-parametrically. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:42:20 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhou",
"Hongyi",
""
],
[
"Hanna",
"Josiah P.",
""
],
[
"Zhu",
"Jin",
""
],
[
"Yang",
"Ying",
""
],
[
"Shi",
"Chengchun",
""
]
] |
2505.22493 | Salvador Ortiz-Latorre | Maria Jolis, Salvador Ortiz-Latorre, Llu\'is Quer-Sardanyons | Convergence in law for quasi-linear SPDEs | null | null | null | null | math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the quasi-linear stochastic wave and heat equations in $\mathbb{R}^d$ with $d\in \{1,2,3\}$ and $d\geq 1$, respectively, and perturbed by an additive Gaussian noise which is white in time and has a homogeneous spatial correlation with spectral measure $\mu_n$. We allow the Fourier transform of $\mu_n$ to be a genuine distribution. Let $u^n$ be the mild solution to these equations. We provide sufficient conditions on the measures $\mu_n$ and the initial data to ensure that $u^n$ converges in law, in the space of continuous functions, to the solution of our equations driven by a noise with spectral measure $\mu$, where $\mu_n\to\mu$ in some sense. We apply our main result to various types of noises, such as the anisotropic fractional noise. We also show that we cover existing results in the literature, such as the case of Riesz kernels and the fractional noise with $d=1$. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:45:23 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jolis",
"Maria",
""
],
[
"Ortiz-Latorre",
"Salvador",
""
],
[
"Quer-Sardanyons",
"Lluís",
""
]
] |
2505.22494 | Michal Kmicikiewicz | Michal Kmicikiewicz, Vincent Fortuin, Ewa Szczurek | ProSpero: Active Learning for Robust Protein Design Beyond Wild-Type Neighborhoods | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Designing protein sequences of both high fitness and novelty is a challenging task in data-efficient protein engineering. Exploration beyond wild-type neighborhoods often leads to biologically implausible sequences or relies on surrogate models that lose fidelity in novel regions. Here, we propose ProSpero, an active learning framework in which a frozen pre-trained generative model is guided by a surrogate updated from oracle feedback. By integrating fitness-relevant residue selection with biologically-constrained Sequential Monte Carlo sampling, our approach enables exploration beyond wild-type neighborhoods while preserving biological plausibility. We show that our framework remains effective even when the surrogate is misspecified. ProSpero consistently outperforms or matches existing methods across diverse protein engineering tasks, retrieving sequences of both high fitness and novelty. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:45:43 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Kmicikiewicz",
"Michal",
""
],
[
"Fortuin",
"Vincent",
""
],
[
"Szczurek",
"Ewa",
""
]
] |
2505.22495 | Osama Maklad | Osama M. Maklad and Muting Hao | Reduced order modelling of air puff test for corneal material characterisation | null | null | null | null | physics.flu-dyn physics.bio-ph physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | Models of the fluid-structure interaction (FSI) model for the air puff test were analysed. Using Abaqus, the air puff test is applied to eyes with varying biomechanical parameters, such as material properties, corneal thickness, and radius. A reduced order model of the air puff (a turbulent impinging jet) has been acquired to decrease simulation time from 48 hours for the FSI model to approximately 12 minutes for the finite element analysis (FEA) model alone. To further accelerate simulations and improve model accuracy, Physics-Informed Neural Networks (PINNs) will be integrated with the reduced-order model. This hybrid approach will help expand the model to a larger dataset, enhancing intraocular pressure (IOP) estimation accuracy and the corneal material properties algorithm through inverse FEA. Additionally, a neural network (NN) framework with embedded Gaussian-modulated waveforms is proposed to model the pressure and deformation distributions on the corneal surface as functions of spatial and temporal parameters. By learning the relationship between corneal biomechanical inputs such as Corneal Central Thickness (CCT), Intraocular Pressure (IOP), and baseline properties (Mu), and the governing coefficients of pressure and deformation, the network accurately reconstructs the result that matches well with the high-fidelity CFD data. This approach can quickly capture the distribution of pressure and deformation. It can also provide insights into the distinct spatial and temporal dynamics of pressure and deformation, giving a more comprehensive understanding of fluid-structure interaction phenomena in the air puff test. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:46:21 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Maklad",
"Osama M.",
""
],
[
"Hao",
"Muting",
""
]
] |
2505.22496 | Long Hui | Long Hui | Risk-Sensitive Conformal Prediction for Catheter Placement Detection in Chest X-rays | null | null | null | null | eess.IV cs.CV stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a novel approach to catheter and line position detection in chest X-rays, combining multi-task learning with risk-sensitive conformal prediction to address critical clinical requirements. Our model simultaneously performs classification, segmentation, and landmark detection, leveraging the synergistic relationship between these tasks to improve overall performance. We further enhance clinical reliability through risk-sensitive conformal prediction, which provides statistically guaranteed prediction sets with higher reliability for clinically critical findings. Experimental results demonstrate excellent performance with 90.68\% overall empirical coverage and 99.29\% coverage for critical conditions, while maintaining remarkable precision in prediction sets. Most importantly, our risk-sensitive approach achieves zero high-risk mispredictions (cases where the system dangerously declares problematic tubes as confidently normal), making the system particularly suitable for clinical deployment. This work offers both accurate predictions and reliably quantified uncertainty -- essential features for life-critical medical applications. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:47:10 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hui",
"Long",
""
]
] |
2505.22497 | Tzvika Geft | Tzvika Geft, Kostas Bekris, Jingjin Yu | Fully Packed and Ready to Go: High-Density, Rearrangement-Free, Grid-Based Storage and Retrieval | null | null | null | null | cs.RO cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Grid-based storage systems with uniformly shaped loads (e.g., containers, pallets, totes) are commonplace in logistics, industrial, and transportation domains. A key performance metric for such systems is the maximization of space utilization, which requires some loads to be placed behind or below others, preventing direct access to them. Consequently, dense storage settings bring up the challenge of determining how to place loads while minimizing costly rearrangement efforts necessary during retrieval. This paper considers the setting involving an inbound phase, during which loads arrive, followed by an outbound phase, during which loads depart. The setting is prevalent in distribution centers, automated parking garages, and container ports. In both phases, minimizing the number of rearrangement actions results in more optimal (e.g., fast, energy-efficient, etc.) operations. In contrast to previous work focusing on stack-based systems, this effort examines the case where loads can be freely moved along the grid, e.g., by a mobile robot, expanding the range of possible motions. We establish that for a range of scenarios, such as having limited prior knowledge of the loads' arrival sequences or grids with a narrow opening, a (best possible) rearrangement-free solution always exists, including when the loads fill the grid to its capacity. In particular, when the sequences are fully known, we establish an intriguing characterization showing that rearrangement can always be avoided if and only if the open side of the grid (used to access the storage) is at least 3 cells wide. We further discuss useful practical implications of our solutions. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:47:43 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Geft",
"Tzvika",
""
],
[
"Bekris",
"Kostas",
""
],
[
"Yu",
"Jingjin",
""
]
] |
2505.22498 | Francesco Hrobat | Angelo A. Casulli, Francesco Hrobat, Daniel Kressner | Lanczos with compression for symmetric matrix Lyapunov equations | 23 pages, 2 figures, 3 tables | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work considers large-scale Lyapunov matrix equations of the form $AX + XA = \boldsymbol{c}\boldsymbol{c}^T$, where $A$ is a symmetric positive definite matrix and $\boldsymbol{c}$ is a vector. Motivated by the need to solve such equations in a wide range of applications, various numerical methods have been developed to compute low-rank approximations of the solution matrix $X$. In this work, we focus on the Lanczos method, which has the distinct advantage of requiring only matrix-vector products with $A$, making it broadly applicable. However, the Lanczos method may suffer from slow convergence when $A$ is ill-conditioned, leading to excessive memory requirements for storing the Krylov subspace basis generated by the algorithm. To address this issue, we propose a novel compression strategy for the Krylov subspace basis that significantly reduces memory usage without hindering convergence. This is supported by both numerical experiments and a convergence analysis. Our analysis also accounts for the loss of orthogonality due to round-off errors in the Lanczos process. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:49:35 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Casulli",
"Angelo A.",
""
],
[
"Hrobat",
"Francesco",
""
],
[
"Kressner",
"Daniel",
""
]
] |
2505.22500 | Ronald Orozco | Ronald Orozco L\'opez | Deformed Bivariate $q$-Appell Polynomials | null | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | In this paper, we introduce bivariate polynomial sets of deformed $q$-Appell type, and we study the algebraic properties of these sets. We show the relation between deformed bivariate $q$-Appell polynomials and deformed homogeneous polynomials. Next, we give some of their characterizations and algebraic structure. Then, we introduce the deformed $q$-Appell operators and obtain Mehler's and Rogers-type formulas of quasi-$q$-Appell polynomials. Finally, some examples of polynomial sequences of deformed $q$-Appell type are given: Bernoulli, Euler, and Genocchi types. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:50:34 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"López",
"Ronald Orozco",
""
]
] |
2505.22501 | Dingchu Zhang | Dingchu Zhang, Yida Zhao, Jialong Wu, Baixuan Li, Wenbiao Yin, Liwen Zhang, Yong Jiang, Yufeng Li, Kewei Tu, Pengjun Xie, Fei Huang | EvolveSearch: An Iterative Self-Evolving Search Agent | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid advancement of large language models (LLMs) has transformed the landscape of agentic information seeking capabilities through the integration of tools such as search engines and web browsers. However, current mainstream approaches for enabling LLM web search proficiency face significant challenges: supervised fine-tuning struggles with data production in open-search domains, while RL converges quickly, limiting their data utilization efficiency. To address these issues, we propose EvolveSearch, a novel iterative self-evolution framework that combines SFT and RL to enhance agentic web search capabilities without any external human-annotated reasoning data. Extensive experiments on seven multi-hop question-answering (MHQA) benchmarks demonstrate that EvolveSearch consistently improves performance across iterations, ultimately achieving an average improvement of 4.7\% over the current state-of-the-art across seven benchmarks, opening the door to self-evolution agentic capabilities in open web search domains. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:50:48 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Dingchu",
""
],
[
"Zhao",
"Yida",
""
],
[
"Wu",
"Jialong",
""
],
[
"Li",
"Baixuan",
""
],
[
"Yin",
"Wenbiao",
""
],
[
"Zhang",
"Liwen",
""
],
[
"Jiang",
"Yong",
""
],
[
"Li",
"Yufeng",
""
],
[
"Tu",
"Kewei",
""
],
[
"Xie",
"Pengjun",
""
],
[
"Huang",
"Fei",
""
]
] |
2505.22502 | Dominic Lowe | Dominic Lowe, M.S. Kim, Roberto Bondesan | Assessing Quantum Advantage for Gaussian Process Regression | 18 pages, 2 figures | null | null | null | quant-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gaussian Process Regression is a well-known machine learning technique for which several quantum algorithms have been proposed. We show here that in a wide range of scenarios these algorithms show no exponential speedup. We achieve this by rigorously proving that the condition number of a kernel matrix scales at least linearly with the matrix size under general assumptions on the data and kernel. We additionally prove that the sparsity and Frobenius norm of a kernel matrix scale linearly under similar assumptions. The implications for the quantum algorithms runtime are independent of the complexity of loading classical data on a quantum computer and also apply to dequantised algorithms. We supplement our theoretical analysis with numerical verification for popular kernels in machine learning. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:50:56 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Lowe",
"Dominic",
""
],
[
"Kim",
"M. S.",
""
],
[
"Bondesan",
"Roberto",
""
]
] |
2505.22503 | Yuanfei Wang | Yuanfei Wang, Xinju Huang, Fangwei Zhong, Yaodong Yang, Yizhou Wang, Yuanpei Chen, Hao Dong | From Strangers to Assistants: Fast Desire Alignment for Embodied Agent-User Adaptation | null | null | null | null | cs.RO cs.AI cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While embodied agents have made significant progress in performing complex physical tasks, real-world applications demand more than pure task execution. The agents must collaborate with unfamiliar agents and human users, whose goals are often vague and implicit. In such settings, interpreting ambiguous instructions and uncovering underlying desires is essential for effective assistance. Therefore, fast and accurate desire alignment becomes a critical capability for embodied agents. In this work, we first develop a home assistance simulation environment HA-Desire that integrates an LLM-driven human user agent exhibiting realistic value-driven goal selection and communication. The ego agent must interact with this proxy user to infer and adapt to the user's latent desires. To achieve this, we present a novel framework FAMER for fast desire alignment, which introduces a desire-based mental reasoning mechanism to identify user intent and filter desire-irrelevant actions. We further design a reflection-based communication module that reduces redundant inquiries, and incorporate goal-relevant information extraction with memory persistence to improve information reuse and reduce unnecessary exploration. Extensive experiments demonstrate that our framework significantly enhances both task execution and communication efficiency, enabling embodied agents to quickly adapt to user-specific desires in complex embodied environments. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:51:13 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wang",
"Yuanfei",
""
],
[
"Huang",
"Xinju",
""
],
[
"Zhong",
"Fangwei",
""
],
[
"Yang",
"Yaodong",
""
],
[
"Wang",
"Yizhou",
""
],
[
"Chen",
"Yuanpei",
""
],
[
"Dong",
"Hao",
""
]
] |
2505.22504 | Ahmed Hossam Mohammed | Ahmed Hossam Mohammed, Kishansingh Rajput, Simon Taylor, Denis Furletov, Sergey Furletov, Malachi Schram | Geometric GNNs for Charged Particle Tracking at GlueX | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nuclear physics experiments are aimed at uncovering the fundamental building blocks of matter. The experiments involve high-energy collisions that produce complex events with many particle trajectories. Tracking charged particles resulting from collisions in the presence of a strong magnetic field is critical to enable the reconstruction of particle trajectories and precise determination of interactions. It is traditionally achieved through combinatorial approaches that scale worse than linearly as the number of hits grows. Since particle hit data naturally form a 3-dimensional point cloud and can be structured as graphs, Graph Neural Networks (GNNs) emerge as an intuitive and effective choice for this task. In this study, we evaluate the GNN model for track finding on the data from the GlueX experiment at Jefferson Lab. We use simulation data to train the model and test on both simulation and real GlueX measurements. We demonstrate that GNN-based track finding outperforms the currently used traditional method at GlueX in terms of segment-based efficiency at a fixed purity while providing faster inferences. We show that the GNN model can achieve significant speedup by processing multiple events in batches, which exploits the parallel computation capability of Graphical Processing Units (GPUs). Finally, we compare the GNN implementation on GPU and FPGA and describe the trade-off. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:52:22 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Mohammed",
"Ahmed Hossam",
""
],
[
"Rajput",
"Kishansingh",
""
],
[
"Taylor",
"Simon",
""
],
[
"Furletov",
"Denis",
""
],
[
"Furletov",
"Sergey",
""
],
[
"Schram",
"Malachi",
""
]
] |
2505.22505 | Alessandro Bosso | Alessandro Bosso, Marco Borghesi, Andrea Iannelli, Giuseppe Notarstefano, Andrew R. Teel | Data-Driven Control of Continuous-Time LTI Systems via Non-Minimal Realizations | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | This article proposes an approach to design output-feedback controllers for unknown continuous-time linear time-invariant systems using only input-output data from a single experiment. To address the lack of state and derivative measurements, we introduce non-minimal realizations whose states can be observed by filtering the available data. We first apply this concept to the disturbance-free case, formulating linear matrix inequalities (LMIs) from batches of sampled signals to design a dynamic, filter-based stabilizing controller. The framework is then extended to the problem of asymptotic tracking and disturbance rejection - in short, output regulation - by incorporating an internal model based on prior knowledge of the disturbance/reference frequencies. Finally, we discuss tuning strategies for a class of multi-input multi-output systems and illustrate the method via numerical examples. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:53:55 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Bosso",
"Alessandro",
""
],
[
"Borghesi",
"Marco",
""
],
[
"Iannelli",
"Andrea",
""
],
[
"Notarstefano",
"Giuseppe",
""
],
[
"Teel",
"Andrew R.",
""
]
] |
2505.22506 | Wenjie Sun | Wenjie Sun, Bingzhe Wu, Zhile Yang, Chengke Wu | Sparsification and Reconstruction from the Perspective of Representation Geometry | 24 pages, 5 figures | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Sparse Autoencoders (SAEs) have emerged as a predominant tool in mechanistic interpretability, aiming to identify interpretable monosemantic features. However, how does sparse encoding organize the representations of activation vector from language models? What is the relationship between this organizational paradigm and feature disentanglement as well as reconstruction performance? To address these questions, we propose the SAEMA, which validates the stratified structure of the representation by observing the variability of the rank of the symmetric semipositive definite (SSPD) matrix corresponding to the modal tensor unfolded along the latent tensor with the level of noise added to the residual stream. To systematically investigate how sparse encoding alters representational structures, we define local and global representations, demonstrating that they amplify inter-feature distinctions by merging similar semantic features and introducing additional dimensionality. Furthermore, we intervene the global representation from an optimization perspective, proving a significant causal relationship between their separability and the reconstruction performance. This study explains the principles of sparsity from the perspective of representational geometry and demonstrates the impact of changes in representational structure on reconstruction performance. Particularly emphasizes the necessity of understanding representations and incorporating representational constraints, providing empirical references for developing new interpretable tools and improving SAEs. The code is available at \hyperlink{https://github.com/wenjie1835/SAERepGeo}{https://github.com/wenjie1835/SAERepGeo}. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 15:54:33 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Sun",
"Wenjie",
""
],
[
"Wu",
"Bingzhe",
""
],
[
"Yang",
"Zhile",
""
],
[
"Wu",
"Chengke",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.