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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.21985 | Naoto Yoshida | Naoto Yoshida and Tadahiro Taniguchi | Reward-Independent Messaging for Decentralized Multi-Agent Reinforcement Learning | null | null | null | null | cs.MA cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | In multi-agent reinforcement learning (MARL), effective communication improves agent performance, particularly under partial observability. We propose MARL-CPC, a framework that enables communication among fully decentralized, independent agents without parameter sharing. MARL-CPC incorporates a message learning model based on collective predictive coding (CPC) from emergent communication research. Unlike conventional methods that treat messages as part of the action space and assume cooperation, MARL-CPC links messages to state inference, supporting communication in non-cooperative, reward-independent settings. We introduce two algorithms -Bandit-CPC and IPPO-CPC- and evaluate them in non-cooperative MARL tasks. Benchmarks show that both outperform standard message-as-action approaches, establishing effective communication even when messages offer no direct benefit to the sender. These results highlight MARL-CPC's potential for enabling coordination in complex, decentralized environments. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:23:47 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Yoshida",
"Naoto",
""
],
[
"Taniguchi",
"Tadahiro",
""
]
] |
2505.21986 | Jonas Sch\"opf | Kanta Takahata and Jonas Sch\"opf and Naoki Nishida and Takahito Aoto | Characterizing Equivalence of Logically Constrained Terms via Existentially Constrained Terms (Full Version) | null | null | null | null | cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Logically constrained term rewriting is a rewriting framework that supports built-in data structures such as integers and bit vectors. Recently, constrained terms play a key role in various analyses and applications of logically constrained term rewriting. A fundamental question on constrained terms arising there is how to characterize equivalence between them. However, in the current literature only limited progress has been made on this. In this paper, we provide several sound and complete solutions to tackle this problem. Our key idea is the introduction of a novel concept, namely existentially constrained terms, into which the original form of constrained terms can be embedded. We present several syntactic characterizations of equivalence between existentially constrained terms. In particular, we provide two different kinds of complete characterizations: one is designed to facilitate equivalence checking, while the other is intended for theoretical analysis. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:23:50 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Takahata",
"Kanta",
""
],
[
"Schöpf",
"Jonas",
""
],
[
"Nishida",
"Naoki",
""
],
[
"Aoto",
"Takahito",
""
]
] |
2505.21987 | Zhendong Mi | Zhendong Mi, Zhenglun Kong, Geng Yuan, Shaoyi Huang | ACE: Exploring Activation Cosine Similarity and Variance for Accurate and Calibration-Efficient LLM Pruning | 9 pages, 2 figures, 13 tables | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | With the rapid expansion of large language models (LLMs), the demand for memory and computational resources has grown significantly. Recent advances in LLM pruning aim to reduce the size and computational cost of these models. However, existing methods often suffer from either suboptimal pruning performance or low time efficiency during the pruning process. In this work, we propose an efficient and effective pruning method that simultaneously achieves high pruning performance and fast pruning speed with improved calibration efficiency. Our approach introduces two key innovations: (1) An activation cosine similarity loss-guided pruning metric, which considers the angular deviation of the output activation between the dense and pruned models. (2) An activation variance-guided pruning metric, which helps preserve semantic distinctions in output activations after pruning, enabling effective pruning with shorter input sequences. These two components can be readily combined to enhance LLM pruning in both accuracy and efficiency. Experimental results show that our method achieves up to an 18% reduction in perplexity and up to 63% decrease in pruning time on prevalent LLMs such as LLaMA, LLaMA-2, and OPT. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:25:16 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Mi",
"Zhendong",
""
],
[
"Kong",
"Zhenglun",
""
],
[
"Yuan",
"Geng",
""
],
[
"Huang",
"Shaoyi",
""
]
] |
2505.21988 | Ziyang Zheng | Ziyang Zheng, Kezhi Li, Zhengyuan Shi, Qiang Xu | Functional Matching of Logic Subgraphs: Beyond Structural Isomorphism | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Subgraph matching in logic circuits is foundational for numerous Electronic Design Automation (EDA) applications, including datapath optimization, arithmetic verification, and hardware trojan detection. However, existing techniques rely primarily on structural graph isomorphism and thus fail to identify function-related subgraphs when synthesis transformations substantially alter circuit topology. To overcome this critical limitation, we introduce the concept of functional subgraph matching, a novel approach that identifies whether a given logic function is implicitly present within a larger circuit, irrespective of structural variations induced by synthesis or technology mapping. Specifically, we propose a two-stage multi-modal framework: (1) learning robust functional embeddings across AIG and post-mapping netlists for functional subgraph detection, and (2) identifying fuzzy boundaries using a graph segmentation approach. Evaluations on standard benchmarks (ITC99, OpenABCD, ForgeEDA) demonstrate significant performance improvements over existing structural methods, with average $93.8\%$ accuracy in functional subgraph detection and a dice score of $91.3\%$ in fuzzy boundary identification. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:31:49 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zheng",
"Ziyang",
""
],
[
"Li",
"Kezhi",
""
],
[
"Shi",
"Zhengyuan",
""
],
[
"Xu",
"Qiang",
""
]
] |
2505.21989 | Haiyang Wang | Bishnu Paudel, James A. Sellers, Haiyang Wang | Extending Recent Congruence Results on $(\ell,\mu)$-Regular Overpartitions | null | null | null | null | math.NT math.CO | http://creativecommons.org/licenses/by/4.0/ | Recently, Alanazi, Munagi, and Saikia employed the theory of modular forms to investigate the arithmetic properties of the function $\overline{R_{\ell,\mu}}(n)$, which enumerates the overpartitions of $n$ where no part is divisible by either $\ell$ or $\mu$, for various integer pairs $(\ell, \mu)$. In this paper, we substantially extend several of their results and establish infinitely many families of new congruences. Our proofs are entirely elementary, relying solely on classical $q$-series manipulations and dissection formulas. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:33:54 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Paudel",
"Bishnu",
""
],
[
"Sellers",
"James A.",
""
],
[
"Wang",
"Haiyang",
""
]
] |
2505.21990 | Zijian Zhou | Zijian Zhou, Jingze Ding, and Rui Zhang | Polarforming Design with Phase Shifter Based Polarization Reconfigurable Antennas | 5 pages, 5 figures | null | null | null | eess.SP | http://creativecommons.org/licenses/by/4.0/ | In this paper, we propose a new form of polarization reconfigurable antennas (PRAs) that can form linear, circular, and general elliptical polarizations assisted by phase shifters (PSs). With PRAs, polarforming is achieved, which enables the antenna to shape its polarization into a desired state for aligning with that of the received electromagnetic (EM) wave or reconfiguring that of the transmit EM wave. To demonstrate the benefits of polarforming, we investigate a PRA-aided single-input single-output (SISO) communication system equipped with tunable PSs for polarization adaptation. We characterize the achievable signal-to-noise ratio (SNR) at the receiver as a function of the phase shifts of PS-based PRAs. Moreover, we develop an alternating optimization approach to maximize the SNR by optimizing the phase shifts at both the transmitter and receiver. Finally, comprehensive simulation results are presented, which not only validate the effectiveness of polarforming in mitigating the channel depolarization effects, but also demonstrate its substantial performance improvement over conventional systems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:36:05 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhou",
"Zijian",
""
],
[
"Ding",
"Jingze",
""
],
[
"Zhang",
"Rui",
""
]
] |
2505.21991 | Zhixing Huang | Zhixing Huang, Yi Mei, Fangfang Zhang, Mengjie Zhang, Wolfgang Banzhaf | Bridging Fitness With Search Spaces By Fitness Supremums: A Theoretical Study on LGP | null | null | null | null | cs.NE | http://creativecommons.org/licenses/by/4.0/ | Genetic programming has undergone rapid development in recent years. However, theoretical studies of genetic programming are far behind. One of the major obstacles to theoretical studies is the challenge of developing a model to describe the relationship between fitness values and program genotypes. In this paper, we take linear genetic programming (LGP) as an example to study the fitness-to-genotype relationship. We find that the fitness expectation increases with fitness supremum over instruction editing distance, considering 1) the fitness supremum linearly increases with the instruction editing distance in LGP, 2) the fitness infimum is fixed, and 3) the fitness probabilities over different instruction editing distances are similar. We then extend these findings to explain the bloat effect and the minimum hitting time of LGP based on instruction editing distance. The bloat effect happens because it is more likely to produce better offspring by adding instructions than by removing them, given an instruction editing distance from the optimal program. The analysis of the minimum hitting time suggests that for a basic LGP genetic operator (i.e., freemut), maintaining a necessarily small program size and mutating multiple instructions each time can improve LGP performance. The reported empirical results verify our hypothesis. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:39:33 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Huang",
"Zhixing",
""
],
[
"Mei",
"Yi",
""
],
[
"Zhang",
"Fangfang",
""
],
[
"Zhang",
"Mengjie",
""
],
[
"Banzhaf",
"Wolfgang",
""
]
] |
2505.21992 | Hanseong Jo | Hanseong Jo, Pavel Shafirin, Christopher Le, Caden Chan, Artur Davoyan | Soft Electrothermal Meta-Actuator for Robust Multifunctional Control | 23 pages, 5 figures | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Soft electrothermal actuators are of great interest in diverse application domains for their simplicity, compliance, and ease of control. However, the very nature of thermally induced mechanical actuation sets inherent operation constraints: unidirectional motion, environmental sensitivity, and slow response times limited by passive cooling. To overcome these constraints, we propose a meta-actuator architecture, which uses engineered heat transfer in thin films to achieve multifunctional operation. We demonstrate electrically selectable bidirectional motion with large deflection ($ \geq $28% of actuator length at 0.75 W), suppressed thermal sensitivity to ambient temperature changes when compared to conventional actuators (>100$ \times $ lower), and actively forced return to the rest state, which is 10 times faster than that with passive cooling. We further show that our meta-actuator approach enables extended ranges of motions for manipulating complex objects. Versatile soft gripper operations highlight the meta-actuator's potential for soft robotics and devices. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:45:10 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jo",
"Hanseong",
""
],
[
"Shafirin",
"Pavel",
""
],
[
"Le",
"Christopher",
""
],
[
"Chan",
"Caden",
""
],
[
"Davoyan",
"Artur",
""
]
] |
2505.21993 | Hemant Singh | Dimpi and Hemant Kumar Singh | Free Circle Actions on the Product of Three Spheres | arXiv admin note: substantial text overlap with arXiv:2303.16474 | null | null | null | math.AT | http://creativecommons.org/licenses/by/4.0/ | The orbit spaces of free S^0-actions on the mod 2 cohomology product of three spheres, S^n x S^m x S^l, 1 <= n <= m <= l have been determined in [6]. In this paper, we extend these findings to free S^1-actions on the rational cohomology product of three spheres. This extension also builds upon the work of Dotzel et al. [7], who studied free circle actions on the rational cohomology product of two spheres. Additionally, we establish Borsuk-Ulam type theorems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:47:27 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Dimpi",
"",
""
],
[
"Singh",
"Hemant Kumar",
""
]
] |
2505.21994 | Seungchan Ko | Josef Dick, Seungchan Ko, Kassem Mustapha and Sanghyeon Park | Locking-Free Training of Physics-Informed Neural Network for Solving Nearly Incompressible Elasticity Equations | null | null | null | null | math.NA cs.LG cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Due to divergence instability, the accuracy of low-order conforming finite element methods for nearly incompressible homogeneous elasticity equations deteriorates as the Lam\'e coefficient $\lambda\to\infty$, or equivalently as the Poisson ratio $\nu\to1/2$. This phenomenon, known as locking or non-robustness, remains not fully understood despite extensive investigation. In this paper, we propose a robust method based on a fundamentally different, machine-learning-driven approach. Leveraging recently developed Physics-Informed Neural Networks (PINNs), we address the numerical solution of linear elasticity equations governing nearly incompressible materials. The core idea of our method is to appropriately decompose the given equations to alleviate the extreme imbalance in the coefficients, while simultaneously solving both the forward and inverse problems to recover the solutions of the decomposed systems as well as the associated external conditions. Through various numerical experiments, including constant, variable and parametric Lam\'e coefficients, we illustrate the efficiency of the proposed methodology. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:52:03 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Dick",
"Josef",
""
],
[
"Ko",
"Seungchan",
""
],
[
"Mustapha",
"Kassem",
""
],
[
"Park",
"Sanghyeon",
""
]
] |
2505.21996 | Zihan Ding | Taiye Chen, Xun Hu, Zihan Ding, Chi Jin | Learning World Models for Interactive Video Generation | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Foundational world models must be both interactive and preserve spatiotemporal coherence for effective future planning with action choices. However, present models for long video generation have limited inherent world modeling capabilities due to two main challenges: compounding errors and insufficient memory mechanisms. We enhance image-to-video models with interactive capabilities through additional action conditioning and autoregressive framework, and reveal that compounding error is inherently irreducible in autoregressive video generation, while insufficient memory mechanism leads to incoherence of world models. We propose video retrieval augmented generation (VRAG) with explicit global state conditioning, which significantly reduces long-term compounding errors and increases spatiotemporal consistency of world models. In contrast, naive autoregressive generation with extended context windows and retrieval-augmented generation prove less effective for video generation, primarily due to the limited in-context learning capabilities of current video models. Our work illuminates the fundamental challenges in video world models and establishes a comprehensive benchmark for improving video generation models with internal world modeling capabilities. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:55:44 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Chen",
"Taiye",
""
],
[
"Hu",
"Xun",
""
],
[
"Ding",
"Zihan",
""
],
[
"Jin",
"Chi",
""
]
] |
2505.21997 | Jihong Zhang | Jihong Zhang, Xinya Liang, Anqi Deng, Nicole Bonge, Lin Tan, Ling Zhang and Nicole Zarrett | Leveraging Interview-Informed LLMs to Model Survey Responses: Comparative Insights from AI-Generated and Human Data | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Mixed methods research integrates quantitative and qualitative data but faces challenges in aligning their distinct structures, particularly in examining measurement characteristics and individual response patterns. Advances in large language models (LLMs) offer promising solutions by generating synthetic survey responses informed by qualitative data. This study investigates whether LLMs, guided by personal interviews, can reliably predict human survey responses, using the Behavioral Regulations in Exercise Questionnaire (BREQ) and interviews from after-school program staff as a case study. Results indicate that LLMs capture overall response patterns but exhibit lower variability than humans. Incorporating interview data improves response diversity for some models (e.g., Claude, GPT), while well-crafted prompts and low-temperature settings enhance alignment between LLM and human responses. Demographic information had less impact than interview content on alignment accuracy. These findings underscore the potential of interview-informed LLMs to bridge qualitative and quantitative methodologies while revealing limitations in response variability, emotional interpretation, and psychometric fidelity. Future research should refine prompt design, explore bias mitigation, and optimize model settings to enhance the validity of LLM-generated survey data in social science research. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:57:26 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Jihong",
""
],
[
"Liang",
"Xinya",
""
],
[
"Deng",
"Anqi",
""
],
[
"Bonge",
"Nicole",
""
],
[
"Tan",
"Lin",
""
],
[
"Zhang",
"Ling",
""
],
[
"Zarrett",
"Nicole",
""
]
] |
2505.21998 | Yuhao Hu | Yuhao Hu | Hyperbolic Monge-Amp\`ere systems with $S_1=0$ | 30 pages | null | null | null | math.DG math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For hyperbolic Monge-Amp\`ere systems, a well-known solution of the equivalence problem yields two invariant tensors, ${S}_1$ and ${S}_2$, defined on the underlying $5$-manifold, where ${S}_2=0$ characterizes systems that are Euler-Lagrange. In this article, we consider the `opposite' case, ${S}_1 = 0$, and show that the local generality of such systems is `$2$ arbitrary functions of $3$ variables'. In addition, we classify all $S_1=0$ systems with cohomogeneity at most one, which turn out to be linear up to contact transformations. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 05:58:04 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hu",
"Yuhao",
""
]
] |
2505.21999 | Ashim Gupta | Ashim Gupta, Maitrey Mehta, Zhichao Xu, Vivek Srikumar | Found in Translation: Measuring Multilingual LLM Consistency as Simple as Translate then Evaluate | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) provide detailed and impressive responses to queries in English. However, are they really consistent at responding to the same query in other languages? The popular way of evaluating for multilingual performance of LLMs requires expensive-to-collect annotated datasets. Further, evaluating for tasks like open-ended generation, where multiple correct answers may exist, is nontrivial. Instead, we propose to evaluate the predictability of model response across different languages. In this work, we propose a framework to evaluate LLM's cross-lingual consistency based on a simple Translate then Evaluate strategy. We instantiate this evaluation framework along two dimensions of consistency: information and empathy. Our results reveal pronounced inconsistencies in popular LLM responses across thirty languages, with severe performance deficits in certain language families and scripts, underscoring critical weaknesses in their multilingual capabilities. These findings necessitate cross-lingual evaluations that are consistent along multiple dimensions. We invite practitioners to use our framework for future multilingual LLM benchmarking. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:00:21 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Gupta",
"Ashim",
""
],
[
"Mehta",
"Maitrey",
""
],
[
"Xu",
"Zhichao",
""
],
[
"Srikumar",
"Vivek",
""
]
] |
2505.22000 | Xiaochen Wei | Xiaochen Wei, Weiwei Guo, Wenxian Yu | Collaborative Learning for Unsupervised Multimodal Remote Sensing Image Registration: Integrating Self-Supervision and MIM-Guided Diffusion-Based Image Translation | null | null | null | null | eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The substantial modality-induced variations in radiometric, texture, and structural characteristics pose significant challenges for the accurate registration of multimodal images. While supervised deep learning methods have demonstrated strong performance, they often rely on large-scale annotated datasets, limiting their practical application. Traditional unsupervised methods usually optimize registration by minimizing differences in feature representations, yet often fail to robustly capture geometric discrepancies, particularly under substantial spatial and radiometric variations, thus hindering convergence stability. To address these challenges, we propose a Collaborative Learning framework for Unsupervised Multimodal Image Registration, named CoLReg, which reformulates unsupervised registration learning into a collaborative training paradigm comprising three components: (1) a cross-modal image translation network, MIMGCD, which employs a learnable Maximum Index Map (MIM) guided conditional diffusion model to synthesize modality-consistent image pairs; (2) a self-supervised intermediate registration network which learns to estimate geometric transformations using accurate displacement labels derived from MIMGCD outputs; (3) a distilled cross-modal registration network trained with pseudo-label predicted by the intermediate network. The three networks are jointly optimized through an alternating training strategy wherein each network enhances the performance of the others. This mutual collaboration progressively reduces modality discrepancies, enhances the quality of pseudo-labels, and improves registration accuracy. Extensive experimental results on multiple datasets demonstrate that our ColReg achieves competitive or superior performance compared to state-of-the-art unsupervised approaches and even surpasses several supervised baselines. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:02:06 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wei",
"Xiaochen",
""
],
[
"Guo",
"Weiwei",
""
],
[
"Yu",
"Wenxian",
""
]
] |
2505.22001 | Koji Hasegawa | Misaki Mitsuno, Xiao Ma, and Koji Hasegawa | Evaporation-induced freezing dynamics of droplets levitated in acoustic field | null | null | null | null | physics.flu-dyn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents the evaporation-induced freezing dynamics of pure cyclohexane droplets levitated via acoustic levitation. Acoustic levitation has attracted considerable attention across various fields owing to its potential to create lab-in-a-drop systems. While droplet evaporation is a fundamental physicochemical process in such a platform, the freezing of droplets induced by evaporation has been sparsely explored experimentally. For pure cyclohexane, the rapid evaporation of levitated droplets initiated freezing at the droplet surface. To better understand this evaporation-induced freezing process, the evaporation behavior of the levitated cyclohexane droplets was visualized and quantified using a high-speed camera and an infrared camera. According to the obtained experimental data, the evaporative heat transfer characteristics of the droplets were identified with theoretical models. Using the derived heat transfer coefficient, a mathematical prediction method for the onset of freezing was proposed and validated with the experimental data. These experimental findings offer valuable insights into the phase transition process and its potential physicochemical applications in a containerless environment. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:02:11 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Mitsuno",
"Misaki",
""
],
[
"Ma",
"Xiao",
""
],
[
"Hasegawa",
"Koji",
""
]
] |
2505.22002 | Zijing Hu | Zijing Hu, Fengda Zhang, Kun Kuang | D-Fusion: Direct Preference Optimization for Aligning Diffusion Models with Visually Consistent Samples | Accepted to ICML 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The practical applications of diffusion models have been limited by the misalignment between generated images and corresponding text prompts. Recent studies have introduced direct preference optimization (DPO) to enhance the alignment of these models. However, the effectiveness of DPO is constrained by the issue of visual inconsistency, where the significant visual disparity between well-aligned and poorly-aligned images prevents diffusion models from identifying which factors contribute positively to alignment during fine-tuning. To address this issue, this paper introduces D-Fusion, a method to construct DPO-trainable visually consistent samples. On one hand, by performing mask-guided self-attention fusion, the resulting images are not only well-aligned, but also visually consistent with given poorly-aligned images. On the other hand, D-Fusion can retain the denoising trajectories of the resulting images, which are essential for DPO training. Extensive experiments demonstrate the effectiveness of D-Fusion in improving prompt-image alignment when applied to different reinforcement learning algorithms. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:03:41 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hu",
"Zijing",
""
],
[
"Zhang",
"Fengda",
""
],
[
"Kuang",
"Kun",
""
]
] |
2505.22003 | Jatin Gupta | Jatin Gupta, Akhil Sharma, Saransh Singhania, and Ali Imam Abidi | Legal Assist AI: Leveraging Transformer-Based Model for Effective Legal Assistance | 9 pages, 5 tables, 4 figures. This is a revised version of a preprint previously available at this URL: https://doi.org/10.21203/rs.3.rs-5351879/v1 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Pursuit of accessible legal assistance in India faces a critical gap, as many citizens struggle to leverage their legal rights due to limited awareness and access to relevant legal information. This paper introduces Legal Assist AI, a transformer-based model designed to bridge this gap by offering effective legal assistance through large language models (LLMs). The system retrieves relevant legal information from a curated database and generates accurate responses, enabling effective assistance for diverse users, including legal professionals, scholars, and the general public. The model was fine-tuned on extensive datasets from the Indian legal domain, including Indian Constitution, Bharatiya Nyaya Sanhita (BNS), Bharatiya Nagarik Suraksha Sanhita (BNSS) and so forth, providing a robust understanding of the complexities of Indian law. By incorporating domain-specific legal datasets, the proposed model demonstrated remarkable efficiency and specialization in legal Question-Answering. The model was evaluated against state-of-the-art models such as GPT-3.5 Turbo and Mistral 7B, achieving a 60.08% score on the AIBE, outperforming its competitors in legal reasoning and accuracy. Unlike other models, Legal Assist AI avoided common issues such as hallucinations, making it highly reliable for practical legal applications. It showcases the model's applicability in real-world legal scenarios, with future iterations aiming to enhance performance and expand its dataset to cover a broader range of multilingual and case-specific queries as well. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:06:53 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Gupta",
"Jatin",
""
],
[
"Sharma",
"Akhil",
""
],
[
"Singhania",
"Saransh",
""
],
[
"Abidi",
"Ali Imam",
""
]
] |
2505.22004 | Bruno Vallette | Eric Hoffbeck, Johan Leray, Bruno Vallette | Simplicial properadic homotopy | 39 pages, comments are welcome | null | null | null | math.AT math.CT math.QA | http://creativecommons.org/licenses/by/4.0/ | In this paper, we settle the homotopy properties of the infinity-morphisms of homotopy (bial)-gebras over properads, i.e. algebraic structures made up of operations with several inputs and outputs. We start by providing the literature with characterizations for the various types of infinity-morphisms, the most seminal one being the equivalence between infinity-quasi-isomorphisms and zig-zags of quasi-isomorphisms which plays a key role in the study the formality property. We establish a simplicial enrichment for the categories of gebras over some cofibrant properads together with their infinity-morphisms, whose homotopy category provides us with the localisation with respect to infinity-quasi-isomorphisms. These results extend to the properadic level known properties for operads, but the lack of the rectification procedure in this setting forces us to use different methods. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:11:03 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hoffbeck",
"Eric",
""
],
[
"Leray",
"Johan",
""
],
[
"Vallette",
"Bruno",
""
]
] |
2505.22005 | Shangkun Huang | Shangkun Huang, Jing Deng, Jintao Kang, Rong Zheng | Leveraging LLM for Stuttering Speech: A Unified Architecture Bridging Recognition and Event Detection | Accepted to Interspeech 2025 | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The performance bottleneck of Automatic Speech Recognition (ASR) in stuttering speech scenarios has limited its applicability in domains such as speech rehabilitation. This paper proposed an LLM-driven ASR-SED multi-task learning framework that jointly optimized the ASR and Stuttering Event Detection (SED) tasks. We proposed a dynamic interaction mechanism where the ASR branch leveraged CTC-generated soft prompts to assist LLM context modeling, while the SED branch output stutter embeddings to enhance LLM comprehension of stuttered speech. We incorporated contrastive learning to strengthen the discriminative power of stuttering acoustic features and applied Focal Loss to mitigate the long-tailed distribution in stuttering event categories. Evaluations on the AS-70 Mandarin stuttering dataset demonstrated that our framework reduced the ASR character error rate (CER) to 5.45% (-37.71% relative reduction) and achieved an average SED F1-score of 73.63% (+46.58% relative improvement). | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:12:19 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Huang",
"Shangkun",
""
],
[
"Deng",
"Jing",
""
],
[
"Kang",
"Jintao",
""
],
[
"Zheng",
"Rong",
""
]
] |
2505.22006 | Changze Qiao | Changze Qiao, Mingming Lu | Efficiently Enhancing General Agents With Hierarchical-categorical Memory | null | null | null | null | cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With large language models (LLMs) demonstrating remarkable capabilities, there has been a surge in research on leveraging LLMs to build general-purpose multi-modal agents. However, existing approaches either rely on computationally expensive end-to-end training using large-scale multi-modal data or adopt tool-use methods that lack the ability to continuously learn and adapt to new environments. In this paper, we introduce EHC, a general agent capable of learning without parameter updates. EHC consists of a Hierarchical Memory Retrieval (HMR) module and a Task-Category Oriented Experience Learning (TOEL) module. The HMR module facilitates rapid retrieval of relevant memories and continuously stores new information without being constrained by memory capacity. The TOEL module enhances the agent's comprehension of various task characteristics by classifying experiences and extracting patterns across different categories. Extensive experiments conducted on multiple standard datasets demonstrate that EHC outperforms existing methods, achieving state-of-the-art performance and underscoring its effectiveness as a general agent for handling complex multi-modal tasks. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:12:51 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Qiao",
"Changze",
""
],
[
"Lu",
"Mingming",
""
]
] |
2505.22007 | Wataru Ikeda | Wataru Ikeda, Masashi Hatano, Ryosei Hara, Mariko Isogawa | Event-based Egocentric Human Pose Estimation in Dynamic Environment | Accepted at ICIP 2025, Project Page: https://wataru823.github.io/D-EventEgo/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Estimating human pose using a front-facing egocentric camera is essential for applications such as sports motion analysis, VR/AR, and AI for wearable devices. However, many existing methods rely on RGB cameras and do not account for low-light environments or motion blur. Event-based cameras have the potential to address these challenges. In this work, we introduce a novel task of human pose estimation using a front-facing event-based camera mounted on the head and propose D-EventEgo, the first framework for this task. The proposed method first estimates the head poses, and then these are used as conditions to generate body poses. However, when estimating head poses, the presence of dynamic objects mixed with background events may reduce head pose estimation accuracy. Therefore, we introduce the Motion Segmentation Module to remove dynamic objects and extract background information. Extensive experiments on our synthetic event-based dataset derived from EgoBody, demonstrate that our approach outperforms our baseline in four out of five evaluation metrics in dynamic environments. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:13:01 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Ikeda",
"Wataru",
""
],
[
"Hatano",
"Masashi",
""
],
[
"Hara",
"Ryosei",
""
],
[
"Isogawa",
"Mariko",
""
]
] |
2505.22008 | Jing-An Sun | Jing-An Sun, Hang Fan, Junchao Gong, Ben Fei, Kun Chen, Fenghua Ling, Wenlong Zhang, Wanghan Xu, Li Yan, Pierre Gentine, Lei Bai | Align-DA: Align Score-based Atmospheric Data Assimilation with Multiple Preferences | null | null | null | null | physics.ao-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data assimilation (DA) aims to estimate the full state of a dynamical system by combining partial and noisy observations with a prior model forecast, commonly referred to as the background. In atmospheric applications, this problem is fundamentally ill-posed due to the sparsity of observations relative to the high-dimensional state space. Traditional methods address this challenge by simplifying background priors to regularize the solution, which are empirical and require continual tuning for application. Inspired by alignment techniques in text-to-image diffusion models, we propose Align-DA, which formulates DA as a generative process and uses reward signals to guide background priors, replacing manual tuning with data-driven alignment. Specifically, we train a score-based model in the latent space to approximate the background-conditioned prior, and align it using three complementary reward signals for DA: (1) assimilation accuracy, (2) forecast skill initialized from the assimilated state, and (3) physical adherence of the analysis fields. Experiments with multiple reward signals demonstrate consistent improvements in analysis quality across different evaluation metrics and observation-guidance strategies. These results show that preference alignment, implemented as a soft constraint, can automatically adapt complex background priors tailored to DA, offering a promising new direction for advancing the field. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:15:41 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Sun",
"Jing-An",
""
],
[
"Fan",
"Hang",
""
],
[
"Gong",
"Junchao",
""
],
[
"Fei",
"Ben",
""
],
[
"Chen",
"Kun",
""
],
[
"Ling",
"Fenghua",
""
],
[
"Zhang",
"Wenlong",
""
],
[
"Xu",
"Wanghan",
""
],
[
"Yan",
"Li",
""
],
[
"Gentine",
"Pierre",
""
],
[
"Bai",
"Lei",
""
]
] |
2505.22009 | Akihiko Fukui | Akihiko Fukui | Hidden Transits: TOI-2285 b is a Warmer sub-Neptune Likely with a Super-Earth Companion | 4 pages, 1 figure, published in RNAAS (the initial submission to arXiv was in April 2025) | Research Notes of the AAS, Volume 9, Issue 4, id.73 (2025) | 10.3847/2515-5172/adc7fb | null | astro-ph.EP astro-ph.IM | http://creativecommons.org/licenses/by/4.0/ | TOI-2285 b is a sub-Neptune-sized planet orbiting a nearby M dwarf, discovered through the TESS photometric survey and ground-based follow-up observations. The planet was initially reported to have an orbital period of 27.27 d, making it one of the lowest temperature sub-Neptunes transiting a bright M dwarf. However, additional TESS data reveal that its true orbital period is 13.64 d, half the original value, resulting in a warmer equilibrium temperature (358 K) than previously estimated (284 K). The misidentification likely resulted from the low signal-to-noise ratio of individual transit signals and the limited number of transits observed by TESS at that time. This case highlights the importance of carefully considering harmonic solutions for similar cases. The additional TESS data also reveal another planetary candidate with an orbital period of 9.67 d and a radius of 1.5 $R_\oplus$, requiring validation in future studies. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:17:04 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Fukui",
"Akihiko",
""
]
] |
2505.22010 | Nasir Hussain | Nasir Hussain, Haohan Chen, Chanh Tran, Philip Huang, Zhuohao Li, Pravir Chugh, William Chen, Ashish Kundu, Yuan Tian | VulBinLLM: LLM-powered Vulnerability Detection for Stripped Binaries | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | Recognizing vulnerabilities in stripped binary files presents a significant challenge in software security. Although some progress has been made in generating human-readable information from decompiled binary files with Large Language Models (LLMs), effectively and scalably detecting vulnerabilities within these binary files is still an open problem. This paper explores the novel application of LLMs to detect vulnerabilities within these binary files. We demonstrate the feasibility of identifying vulnerable programs through a combined approach of decompilation optimization to make the vulnerabilities more prominent and long-term memory for a larger context window, achieving state-of-the-art performance in binary vulnerability analysis. Our findings highlight the potential for LLMs to overcome the limitations of traditional analysis methods and advance the field of binary vulnerability detection, paving the way for more secure software systems. In this paper, we present Vul-BinLLM , an LLM-based framework for binary vulnerability detection that mirrors traditional binary analysis workflows with fine-grained optimizations in decompilation and vulnerability reasoning with an extended context. In the decompilation phase, Vul-BinLLM adds vulnerability and weakness comments without altering the code structure or functionality, providing more contextual information for vulnerability reasoning later. Then for vulnerability reasoning, Vul-BinLLM combines in-context learning and chain-of-thought prompting along with a memory management agent to enhance accuracy. Our evaluations encompass the commonly used synthetic dataset Juliet to evaluate the potential feasibility for analysis and vulnerability detection in C/C++ binaries. Our evaluations show that Vul-BinLLM is highly effective in detecting vulnerabilities on the compiled Juliet dataset. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:17:56 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hussain",
"Nasir",
""
],
[
"Chen",
"Haohan",
""
],
[
"Tran",
"Chanh",
""
],
[
"Huang",
"Philip",
""
],
[
"Li",
"Zhuohao",
""
],
[
"Chugh",
"Pravir",
""
],
[
"Chen",
"William",
""
],
[
"Kundu",
"Ashish",
""
],
[
"Tian",
"Yuan",
""
]
] |
2505.22011 | Jing Zhang | Menghui Zhang, Jing Zhang, Lin Chen and Li Zhuo | Prototype Embedding Optimization for Human-Object Interaction Detection in Livestreaming | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Livestreaming often involves interactions between streamers and objects, which is critical for understanding and regulating web content. While human-object interaction (HOI) detection has made some progress in general-purpose video downstream tasks, when applied to recognize the interaction behaviors between a streamer and different objects in livestreaming, it tends to focuses too much on the objects and neglects their interactions with the streamer, which leads to object bias. To solve this issue, we propose a prototype embedding optimization for human-object interaction detection (PeO-HOI). First, the livestreaming is preprocessed using object detection and tracking techniques to extract features of the human-object (HO) pairs. Then, prototype embedding optimization is adopted to mitigate the effect of object bias on HOI. Finally, after modelling the spatio-temporal context between HO pairs, the HOI detection results are obtained by the prediction head. The experimental results show that the detection accuracy of the proposed PeO-HOI method has detection accuracies of 37.19%@full, 51.42%@non-rare, 26.20%@rare on the publicly available dataset VidHOI, 45.13%@full, 62.78%@non-rare and 30.37%@rare on the self-built dataset BJUT-HOI, which effectively improves the HOI detection performance in livestreaming. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:19:37 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Menghui",
""
],
[
"Zhang",
"Jing",
""
],
[
"Chen",
"Lin",
""
],
[
"Zhuo",
"Li",
""
]
] |
2505.22012 | Martin Huang Mr | Martin Huang, Samuel Muller, Garth Tarr | Data-Adaptive Automatic Threshold Calibration for Stability Selection | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Stability selection has gained popularity as a method for enhancing the performance of variable selection algorithms while controlling false discovery rates. However, achieving these desirable properties depends on correctly specifying the stable threshold parameter, which can be challenging. An arbitrary choice of this parameter can substantially alter the set of selected variables, as the variables' selection probabilities are inherently data-dependent. To address this issue, we propose Exclusion Automatic Threshold Selection (EATS), a data-adaptive algorithm that streamlines stability selection by automating the threshold specification process. Additionally, we introduce Automatic Threshold Selection (ATS), the motivation behind EATS. We evaluate our approach through an extensive simulation study, benchmarking across commonly used variable selection algorithms and several static stable threshold values. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:21:45 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Huang",
"Martin",
""
],
[
"Muller",
"Samuel",
""
],
[
"Tarr",
"Garth",
""
]
] |
2505.22013 | Shangkun Huang | Shangkun Huang, Yuxuan Du, Jingwen Yang, Dejun Zhang, Xupeng Jia, Jing Deng, Jintao Kang, Rong Zheng | Overlap-Adaptive Hybrid Speaker Diarization and ASR-Aware Observation Addition for MISP 2025 Challenge | Accepted to Interspeech 2025 | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents the system developed to address the MISP 2025 Challenge. For the diarization system, we proposed a hybrid approach combining a WavLM end-to-end segmentation method with a traditional multi-module clustering technique to adaptively select the appropriate model for handling varying degrees of overlapping speech. For the automatic speech recognition (ASR) system, we proposed an ASR-aware observation addition method that compensates for the performance limitations of Guided Source Separation (GSS) under low signal-to-noise ratio conditions. Finally, we integrated the speaker diarization and ASR systems in a cascaded architecture to address Track 3. Our system achieved character error rates (CER) of 9.48% on Track 2 and concatenated minimum permutation character error rate (cpCER) of 11.56% on Track 3, ultimately securing first place in both tracks and thereby demonstrating the effectiveness of the proposed methods in real-world meeting scenarios. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:22:37 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Huang",
"Shangkun",
""
],
[
"Du",
"Yuxuan",
""
],
[
"Yang",
"Jingwen",
""
],
[
"Zhang",
"Dejun",
""
],
[
"Jia",
"Xupeng",
""
],
[
"Deng",
"Jing",
""
],
[
"Kang",
"Jintao",
""
],
[
"Zheng",
"Rong",
""
]
] |
2505.22014 | J\"org Franke | J\"org K.H. Franke, Urs Spiegelhalter, Marianna Nezhurina, Jenia Jitsev, Frank Hutter, Michael Hefenbrock | Learning in Compact Spaces with Approximately Normalized Transformers | Preprint | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | In deep learning, regularization and normalization are common solutions for challenges such as overfitting, numerical instabilities, and the increasing variance in the residual stream. An alternative approach is to force all parameters and representations to lie on a hypersphere. This removes the need for regularization and increases convergence speed, but comes with additional costs. In this work, we propose a more holistic but approximate normalization (anTransformer). Our approach constrains the norm of parameters and normalizes all representations via scalar multiplications motivated by the tight concentration of the norms of high-dimensional random vectors. When applied to GPT training, we observe a 40% faster convergence compared to models with QK normalization, with less than 3% additional runtime. Deriving scaling laws for anGPT, we found our method enables training with larger batch sizes and fewer hyperparameters, while matching the favorable scaling characteristics of classic GPT architectures. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:23:19 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Franke",
"Jörg K. H.",
""
],
[
"Spiegelhalter",
"Urs",
""
],
[
"Nezhurina",
"Marianna",
""
],
[
"Jitsev",
"Jenia",
""
],
[
"Hutter",
"Frank",
""
],
[
"Hefenbrock",
"Michael",
""
]
] |
2505.22015 | Zeng Li | Weiming Li, Zeng Li, Siyu Wang, Yanqing Yin, Junpeng Zhu | Debiased distributed PCA under high dimensional spiked model | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study distributed principal component analysis (PCA) in high-dimensional settings under the spiked model. In such regimes, sample eigenvectors can deviate significantly from population ones, introducing a persistent bias. Existing distributed PCA methods are sensitive to this bias, particularly when the number of machines is small. Their consistency typically relies on the number of machines tending to infinity. We propose a debiased distributed PCA algorithm that corrects the local bias before aggregation and incorporates a sparsity-detection step to adaptively handle sparse and non-sparse eigenvectors. Theoretically, we establish the consistency of our estimator under much weaker conditions compared to existing literature. In particular, our approach does not require symmetric innovations and only assumes a finite sixth moment. Furthermore, our method generally achieves smaller estimation error, especially when the number of machines is small. Empirically, extensive simulations and real data experiments demonstrate that our method consistently outperforms existing distributed PCA approaches. The advantage is especially prominent when the leading eigenvectors are sparse or the number of machines is limited. Our method and theoretical analysis are also applicable to the sample correlation matrix. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:23:33 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Weiming",
""
],
[
"Li",
"Zeng",
""
],
[
"Wang",
"Siyu",
""
],
[
"Yin",
"Yanqing",
""
],
[
"Zhu",
"Junpeng",
""
]
] |
2505.22016 | Yifei Xia | Yifei Xia, Shuchen Weng, Siqi Yang, Jingqi Liu, Chengxuan Zhu, Minggui Teng, Zijian Jia, Han Jiang, Boxin Shi | PanoWan: Lifting Diffusion Video Generation Models to 360{\deg} with Latitude/Longitude-aware Mechanisms | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Panoramic video generation enables immersive 360{\deg} content creation, valuable in applications that demand scene-consistent world exploration. However, existing panoramic video generation models struggle to leverage pre-trained generative priors from conventional text-to-video models for high-quality and diverse panoramic videos generation, due to limited dataset scale and the gap in spatial feature representations. In this paper, we introduce PanoWan to effectively lift pre-trained text-to-video models to the panoramic domain, equipped with minimal modules. PanoWan employs latitude-aware sampling to avoid latitudinal distortion, while its rotated semantic denoising and padded pixel-wise decoding ensure seamless transitions at longitude boundaries. To provide sufficient panoramic videos for learning these lifted representations, we contribute PanoVid, a high-quality panoramic video dataset with captions and diverse scenarios. Consequently, PanoWan achieves state-of-the-art performance in panoramic video generation and demonstrates robustness for zero-shot downstream tasks. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:24:21 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Xia",
"Yifei",
""
],
[
"Weng",
"Shuchen",
""
],
[
"Yang",
"Siqi",
""
],
[
"Liu",
"Jingqi",
""
],
[
"Zhu",
"Chengxuan",
""
],
[
"Teng",
"Minggui",
""
],
[
"Jia",
"Zijian",
""
],
[
"Jiang",
"Han",
""
],
[
"Shi",
"Boxin",
""
]
] |
2505.22017 | Siqi Fan | Siqi Fan, Peng Han, Shuo Shang, Yequan Wang, Aixin Sun | CoThink: Token-Efficient Reasoning via Instruct Models Guiding Reasoning Models | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) benefit from increased test-time compute, a phenomenon known as test-time scaling. However, reasoning-optimized models often overthink even simple problems, producing excessively verbose outputs and leading to low token efficiency. By comparing these models with equally sized instruct models, we identify two key causes of this verbosity: (1) reinforcement learning reduces the information density of forward reasoning, and (2) backward chain-of thought training encourages redundant and often unnecessary verification steps. Since LLMs cannot assess the difficulty of a given problem, they tend to apply the same cautious reasoning strategy across all tasks, resulting in inefficient overthinking. To address this, we propose CoThink, an embarrassingly simple pipeline: an instruct model first drafts a high-level solution outline; a reasoning model then works out the solution. We observe that CoThink enables dynamic adjustment of reasoning depth based on input difficulty. Evaluated with three reasoning models DAPO, DeepSeek-R1, and QwQ on three datasets GSM8K, MATH500, and AIME24, CoThink reduces total token generation by 22.3% while maintaining pass@1 accuracy within a 0.42% margin on average. With reference to the instruct model, we formally define reasoning efficiency and observe a potential reasoning efficiency scaling law in LLMs. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:24:45 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Fan",
"Siqi",
""
],
[
"Han",
"Peng",
""
],
[
"Shang",
"Shuo",
""
],
[
"Wang",
"Yequan",
""
],
[
"Sun",
"Aixin",
""
]
] |
2505.22019 | Qiuchen Wang | Qiuchen Wang, Ruixue Ding, Yu Zeng, Zehui Chen, Lin Chen, Shihang Wang, Pengjun Xie, Fei Huang, Feng Zhao | VRAG-RL: Empower Vision-Perception-Based RAG for Visually Rich Information Understanding via Iterative Reasoning with Reinforcement Learning | null | null | null | null | cs.CL cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Effectively retrieving, reasoning and understanding visually rich information remains a challenge for RAG methods. Traditional text-based methods cannot handle visual-related information. On the other hand, current vision-based RAG approaches are often limited by fixed pipelines and frequently struggle to reason effectively due to the insufficient activation of the fundamental capabilities of models. As RL has been proven to be beneficial for model reasoning, we introduce VRAG-RL, a novel RL framework tailored for complex reasoning across visually rich information. With this framework, VLMs interact with search engines, autonomously sampling single-turn or multi-turn reasoning trajectories with the help of visual perception tokens and undergoing continual optimization based on these samples. Our approach highlights key limitations of RL in RAG domains: (i) Prior Multi-modal RAG approaches tend to merely incorporate images into the context, leading to insufficient reasoning token allocation and neglecting visual-specific perception; and (ii) When models interact with search engines, their queries often fail to retrieve relevant information due to the inability to articulate requirements, thereby leading to suboptimal performance. To address these challenges, we define an action space tailored for visually rich inputs, with actions including cropping and scaling, allowing the model to gather information from a coarse-to-fine perspective. Furthermore, to bridge the gap between users' original inquiries and the retriever, we employ a simple yet effective reward that integrates query rewriting and retrieval performance with a model-based reward. Our VRAG-RL optimizes VLMs for RAG tasks using specially designed RL strategies, aligning the model with real-world applications. The code is available at \hyperlink{https://github.com/Alibaba-NLP/VRAG}{https://github.com/Alibaba-NLP/VRAG}. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:30:51 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wang",
"Qiuchen",
""
],
[
"Ding",
"Ruixue",
""
],
[
"Zeng",
"Yu",
""
],
[
"Chen",
"Zehui",
""
],
[
"Chen",
"Lin",
""
],
[
"Wang",
"Shihang",
""
],
[
"Xie",
"Pengjun",
""
],
[
"Huang",
"Fei",
""
],
[
"Zhao",
"Feng",
""
]
] |
2505.22020 | Yidian Wu | Yidian Wu, Rui Liu, Runbin Luo, Wensi Wang | Falling Threads During Solar Filament Eruptions | null | null | null | null | astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mass drainage is frequently observed in solar filaments. During filament eruptions, falling material most likely flows along magnetic field lines, which may provide important clues for the magnetic structures of filaments. Here we study three filament eruptions exhibiting significant mass draining, often manifested as falling threads at a constant speed ranging between 30--300 km/s. We found that most of the falling material lands onto the hooked segments of flare ribbons, only a small fraction lands inside the hooks, and almost none lands onto the straight segments of ribbons. Based on these observations we surmise that before eruptions most of the filament mass is entrained by field lines threading the quasi-separatrix layers (QSLs), which wrap around the filament field and whose footpoints are mapped by the hooked ribbons, and that the magnetic reconnection involving these field lines is the major cause of the mass drainage during eruptions. Additionally, the light curves of the hooked ribbons suggest that during eruptions the earlier (later) QSL boundary of filaments is threaded by mass-loaded (depleted) field lines. By assuming that the constant-speed motion is due to a drag force balancing the gravity, we proposed a simplified analytical model to estimate the density contrast of the falling material. The estimated density contrast is then fed into a numerical model, in which filament threads fall along vertical magnetic field lines through a gravitationally stratified atmosphere. The resultant falling speeds, however, are short of observed values, which calls for further investigations. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:31:50 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wu",
"Yidian",
""
],
[
"Liu",
"Rui",
""
],
[
"Luo",
"Runbin",
""
],
[
"Wang",
"Wensi",
""
]
] |
2505.22021 | Zhihong Tang | Zhihong Tang, Yang Li | GL-PGENet: A Parameterized Generation Framework for Robust Document Image Enhancement | 12 pages, 7 figures | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Document Image Enhancement (DIE) serves as a critical component in Document AI systems, where its performance substantially determines the effectiveness of downstream tasks. To address the limitations of existing methods confined to single-degradation restoration or grayscale image processing, we present Global with Local Parametric Generation Enhancement Network (GL-PGENet), a novel architecture designed for multi-degraded color document images, ensuring both efficiency and robustness in real-world scenarios. Our solution incorporates three key innovations: First, a hierarchical enhancement framework that integrates global appearance correction with local refinement, enabling coarse-to-fine quality improvement. Second, a Dual-Branch Local-Refine Network with parametric generation mechanisms that replaces conventional direct prediction, producing enhanced outputs through learned intermediate parametric representations rather than pixel-wise mapping. This approach enhances local consistency while improving model generalization. Finally, a modified NestUNet architecture incorporating dense block to effectively fuse low-level pixel features and high-level semantic features, specifically adapted for document image characteristics. In addition, to enhance generalization performance, we adopt a two-stage training strategy: large-scale pretraining on a synthetic dataset of 500,000+ samples followed by task-specific fine-tuning. Extensive experiments demonstrate the superiority of GL-PGENet, achieving state-of-the-art SSIM scores of 0.7721 on DocUNet and 0.9480 on RealDAE. The model also exhibits remarkable cross-domain adaptability and maintains computational efficiency for high-resolution images without performance degradation, confirming its practical utility in real-world scenarios. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:37:06 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Tang",
"Zhihong",
""
],
[
"Li",
"Yang",
""
]
] |
2505.22022 | Farjana Siddiqua | Farjana Siddiqua and Catalin Trenchea | A High Accuracy Symplectic Scheme for Advection Diffusion Reaction Models in Bioseparation | 33 pages, 5 Figures, 2 tables | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We analyze an advection-diffusion-reaction problem with non-homogeneous boundary conditions that models the chromatography process, a vital stage in bioseparation. We prove stability and error estimates for both constant and affine adsorption, using the symplectic one-step implicit midpoint method for time discretization and finite elements for spatial discretization. In addition, we perform the stability analysis for the nonlinear, explicit adsorption in the continuous and semi-discrete cases. For the nonlinear, explicit adsorption, we also complete the error analysis for the semi-discrete case and prove the existence of a solution for the fully discrete case. The numerical tests validate our theoretical results. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:38:20 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Siddiqua",
"Farjana",
""
],
[
"Trenchea",
"Catalin",
""
]
] |
2505.22023 | Ritwik Murali | Ritwik Murali and Akash Ravi | Securing the Software Package Supply Chain for Critical Systems | 14 Pages,3 Figures. Published as a chapter in the book, "Cybersecurity and Data Science Innovations for Sustainable Development of HEICC". CRC Press, eBook ISBN: 9781032711300 | Cybersecurity and Data Science Innovations for Sustainable Development of HEICC. CRC Press 361-374.2025 | 10.1201/9781032711300-25 | null | cs.SE cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Software systems have grown as an indispensable commodity used across various industries, and almost all essential services depend on them for effective operation. The software is no longer an independent or stand-alone piece of code written by a developer but rather a collection of packages designed by multiple developers across the globe. Ensuring the reliability and resilience of these systems is crucial since emerging threats target software supply chains, as demonstrated by the widespread SolarWinds hack in late 2020. These supply chains extend beyond patches and updates, involving distribution networks throughout the software lifecycle. Industries like smart grids, manufacturing, healthcare, and finance rely on interconnected software systems and their dependencies for effective functioning. To secure software modules and add-ons, robust distribution architectures are essential. The proposed chapter enhances the existing delivery frameworks by including a permissioned ledger with Proof of Authority consensus and multi-party signatures. The proposed system aims to prevent attacks while permitting every stakeholder to verify the same. Critical systems can interface with the secure pipeline without disrupting existing functionalities, thus preventing the cascading effect of an attack at any point in the supply chain. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:42:37 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Murali",
"Ritwik",
""
],
[
"Ravi",
"Akash",
""
]
] |
2505.22024 | Long-Khanh Pham | Long-Khanh Pham, Thanh V. T. Tran, Minh-Tan Pham, Van Nguyen | RESOUND: Speech Reconstruction from Silent Videos via Acoustic-Semantic Decomposed Modeling | accepted in Interspeech 2025 | null | null | null | cs.SD cs.CV eess.AS | http://creativecommons.org/licenses/by/4.0/ | Lip-to-speech (L2S) synthesis, which reconstructs speech from visual cues, faces challenges in accuracy and naturalness due to limited supervision in capturing linguistic content, accents, and prosody. In this paper, we propose RESOUND, a novel L2S system that generates intelligible and expressive speech from silent talking face videos. Leveraging source-filter theory, our method involves two components: an acoustic path to predict prosody and a semantic path to extract linguistic features. This separation simplifies learning, allowing independent optimization of each representation. Additionally, we enhance performance by integrating speech units, a proven unsupervised speech representation technique, into waveform generation alongside mel-spectrograms. This allows RESOUND to synthesize prosodic speech while preserving content and speaker identity. Experiments conducted on two standard L2S benchmarks confirm the effectiveness of the proposed method across various metrics. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:46:13 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Pham",
"Long-Khanh",
""
],
[
"Tran",
"Thanh V. T.",
""
],
[
"Pham",
"Minh-Tan",
""
],
[
"Nguyen",
"Van",
""
]
] |
2505.22025 | Manchao Bao | Manchao Bao, Shengjiang Fang, Tao Yue, Xuemei Hu | Learnable Burst-Encodable Time-of-Flight Imaging for High-Fidelity Long-Distance Depth Sensing | null | null | null | null | cs.CV eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Long-distance depth imaging holds great promise for applications such as autonomous driving and robotics. Direct time-of-flight (dToF) imaging offers high-precision, long-distance depth sensing, yet demands ultra-short pulse light sources and high-resolution time-to-digital converters. In contrast, indirect time-of-flight (iToF) imaging often suffers from phase wrapping and low signal-to-noise ratio (SNR) as the sensing distance increases. In this paper, we introduce a novel ToF imaging paradigm, termed Burst-Encodable Time-of-Flight (BE-ToF), which facilitates high-fidelity, long-distance depth imaging. Specifically, the BE-ToF system emits light pulses in burst mode and estimates the phase delay of the reflected signal over the entire burst period, thereby effectively avoiding the phase wrapping inherent to conventional iToF systems. Moreover, to address the low SNR caused by light attenuation over increasing distances, we propose an end-to-end learnable framework that jointly optimizes the coding functions and the depth reconstruction network. A specialized double well function and first-order difference term are incorporated into the framework to ensure the hardware implementability of the coding functions. The proposed approach is rigorously validated through comprehensive simulations and real-world prototype experiments, demonstrating its effectiveness and practical applicability. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:46:43 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Bao",
"Manchao",
""
],
[
"Fang",
"Shengjiang",
""
],
[
"Yue",
"Tao",
""
],
[
"Hu",
"Xuemei",
""
]
] |
2505.22026 | Matteo Gorgone | M. Gorgone, F.Oliveri, A. Ricciardello, P. Rogolino | Two-dimensional equilibrium configurations in Korteweg fluids | 15 pages, 16 figures | Theoretical and Applied Mechanics 49, 111-122 (2022) | 10.2298/TAM220930008G | null | math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, after reviewing the form of the constitutive equations for a third grade Korteweg fluid, recently derived by means of an extended Liu procedure, an equilibrium problem is investigated. By considering a two--dimensional setting, it is derived a single nonlinear elliptic equation such that the equilibrium conditions are identically satisfied. Such an equation is discussed both analytically and numerically. Moreover, by considering a particular boundary value problem of Dirichlet type, some preliminary numerical solutions are presented. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:47:07 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Gorgone",
"M.",
""
],
[
"Oliveri",
"F.",
""
],
[
"Ricciardello",
"A.",
""
],
[
"Rogolino",
"P.",
""
]
] |
2505.22027 | June-Woo Kim | Miika Toikkanen, June-Woo Kim | Improving Respiratory Sound Classification with Architecture-Agnostic Knowledge Distillation from Ensembles | Accepted to Interspeech 2025 | null | null | null | cs.SD cs.AI eess.AS | http://creativecommons.org/licenses/by/4.0/ | Respiratory sound datasets are limited in size and quality, making high performance difficult to achieve. Ensemble models help but inevitably increase compute cost at inference time. Soft label training distills knowledge efficiently with extra cost only at training. In this study, we explore soft labels for respiratory sound classification as an architecture-agnostic approach to distill an ensemble of teacher models into a student model. We examine different variations of our approach and find that even a single teacher, identical to the student, considerably improves performance beyond its own capability, with optimal gains achieved using only a few teachers. We achieve the new state-of-the-art Score of 64.39 on ICHBI, surpassing the previous best by 0.85 and improving average Scores across architectures by more than 1.16. Our results highlight the effectiveness of knowledge distillation with soft labels for respiratory sound classification, regardless of size or architecture. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:49:18 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Toikkanen",
"Miika",
""
],
[
"Kim",
"June-Woo",
""
]
] |
2505.22028 | ZiHao Zhou | Zi-Hao Zhou, Jun-Jie Wang, Tong Wei and Min-Ling Zhang | Weakly-Supervised Contrastive Learning for Imprecise Class Labels | 38 pages, 2 figures, 11 tables | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Contrastive learning has achieved remarkable success in learning effective representations, with supervised contrastive learning often outperforming self-supervised approaches. However, in real-world scenarios, data annotations are often ambiguous or inaccurate, meaning that class labels may not reliably indicate whether two examples belong to the same class. This limitation restricts the applicability of supervised contrastive learning. To address this challenge, we introduce the concept of ``continuous semantic similarity'' to define positive and negative pairs. Instead of directly relying on imprecise class labels, we measure the semantic similarity between example pairs, which quantifies how closely they belong to the same category by iteratively refining weak supervisory signals. Based on this concept, we propose a graph-theoretic framework for weakly-supervised contrastive learning, where semantic similarity serves as the graph weights. Our framework is highly versatile and can be applied to many weakly-supervised learning scenarios. We demonstrate its effectiveness through experiments in two common settings, i.e., noisy label and partial label learning, where existing methods can be easily integrated to significantly improve performance. Theoretically, we establish an error bound for our approach, showing that it can approximate supervised contrastive learning under mild conditions. The implementation code is available at https://github.com/Speechless-10308/WSC. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:50:40 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhou",
"Zi-Hao",
""
],
[
"Wang",
"Jun-Jie",
""
],
[
"Wei",
"Tong",
""
],
[
"Zhang",
"Min-Ling",
""
]
] |
2505.22029 | Jinming Zhang | Jinming Zhang, Xuanru Zhou, Jiachen Lian, Shuhe Li, William Li, Zoe Ezzes, Rian Bogley, Lisa Wauters, Zachary Miller, Jet Vonk, Brittany Morin, Maria Gorno-Tempini, Gopala Anumanchipalli | Analysis and Evaluation of Synthetic Data Generation in Speech Dysfluency Detection | Submitted to Interspeech 2025 | null | null | null | eess.AS cs.AI cs.SD | http://creativecommons.org/licenses/by/4.0/ | Speech dysfluency detection is crucial for clinical diagnosis and language assessment, but existing methods are limited by the scarcity of high-quality annotated data. Although recent advances in TTS model have enabled synthetic dysfluency generation, existing synthetic datasets suffer from unnatural prosody and limited contextual diversity. To address these limitations, we propose LLM-Dys -- the most comprehensive dysfluent speech corpus with LLM-enhanced dysfluency simulation. This dataset captures 11 dysfluency categories spanning both word and phoneme levels. Building upon this resource, we improve an end-to-end dysfluency detection framework. Experimental validation demonstrates state-of-the-art performance. All data, models, and code are open-sourced at https://github.com/Berkeley-Speech-Group/LLM-Dys. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:52:10 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Jinming",
""
],
[
"Zhou",
"Xuanru",
""
],
[
"Lian",
"Jiachen",
""
],
[
"Li",
"Shuhe",
""
],
[
"Li",
"William",
""
],
[
"Ezzes",
"Zoe",
""
],
[
"Bogley",
"Rian",
""
],
[
"Wauters",
"Lisa",
""
],
[
"Miller",
"Zachary",
""
],
[
"Vonk",
"Jet",
""
],
[
"Morin",
"Brittany",
""
],
[
"Gorno-Tempini",
"Maria",
""
],
[
"Anumanchipalli",
"Gopala",
""
]
] |
2505.22030 | Marco Limongi | Marco Limongi, Lorenzo Roberti, Agnese Falla, Alessandro Chieffi and Ken'ichi Nomoto | The chemical yields of stars in the range 9-15 Msun | 24 pages, 12 figures and 3 tables, accepted for publication on ApJS | null | null | null | astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In Limongi et al. (2024) we presented and discussed the main evolutionary properties and final fate of stars in the mass range 7-15 Msun. The evolutions of those models were computed by means of a medium size nuclear network that guaranteed a proper calculation of the nuclear energy generation and hence a good modeling of the physical evolution of these stars. In the present paper, we extend this study by computing the detailed chemical yields of stars in the mass range 9-15 Msun, i.e., those stars that explode as core collapse supernovae (CCSNe). The explosive nucleosynthesis is then computed in the framework of the thermal bomb induced explosion by means of the HYPERION code (Limongi and Chieffi 2020). We find that: (1) the yields of the intermediate mass elements (i.e., O to P) show a steep decrease as the inital mass decreases; (2) the yields of s-weak component, i.e., those produced by the slow neutron captures from Ga to to the first neutron closure shell, decrease almost linearly as a function of the initial mass with respect to the ones produced by the more massive stars; (3) the global contribution of the stars in the mass range 9.22-13 Msun to the yields of a generation of massive stars averaged over a standard initial mass function is negligible for essentially all the isotopes. In spite of this, however, the models of stars in this mass range can be fundamental to interpret the observations of specific supernovae. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:52:16 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Limongi",
"Marco",
""
],
[
"Roberti",
"Lorenzo",
""
],
[
"Falla",
"Agnese",
""
],
[
"Chieffi",
"Alessandro",
""
],
[
"Nomoto",
"Ken'ichi",
""
]
] |
2505.22031 | Hasan Y\"ucedag | Hasan Yucedag and Adam Jatowt | Guess the Age of Photos: An Interactive Web Platform for Historical Image Age Estimation | 4 Pages,4 figures and 1 system architect | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This paper introduces Guess the Age of Photos, a web platform engaging users in estimating the years of historical photographs through two gamified modes: Guess the Year (predicting a single image's year) and Timeline Challenge (comparing two images to identify the older). Built with Python, Flask, Bootstrap, and PostgreSQL, it uses a 10,150-image subset of the Date Estimation in the Wild dataset (1930-1999). Features like dynamic scoring and leaderboards boost engagement. Evaluated with 113 users and 15,473 gameplays, the platform earned a 4.25/5 satisfaction rating. Users excelled in relative comparisons (65.9% accuracy) over absolute year guesses (25.6% accuracy), with older decades easier to identify. The platform serves as an educational tool, fostering historical awareness and analytical skills via interactive exploration of visual heritage. Furthermore, the platform provides a valuable resource for studying human perception of temporal cues in images and could be used to generate annotated data for training and evaluating computer vision models. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:52:49 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Yucedag",
"Hasan",
""
],
[
"Jatowt",
"Adam",
""
]
] |
2505.22032 | Tawfiq Ammari | Tawfiq Ammari, Anna Gutowska, Jacob Ziff, Casey Randazzo, Harihan Subramonyam | Retweets, Receipts, and Resistance: Discourse, Sentiment, and Credibility in Public Health Crisis Twitter | arXiv admin note: substantial text overlap with arXiv:2503.20262 | null | null | null | cs.SI cs.HC | http://creativecommons.org/licenses/by/4.0/ | As the COVID-19 pandemic evolved, the Centers for Disease Control and Prevention (CDC) used Twitter to disseminate safety guidance and updates, reaching millions of users. This study analyzes two years of tweets from, to, and about the CDC using a mixed methods approach to examine discourse characteristics, credibility, and user engagement. We found that the CDCs communication remained largely one directional and did not foster reciprocal interaction, while discussions around COVID19 were deeply shaped by political and ideological polarization. Users frequently cited earlier CDC messages to critique new and sometimes contradictory guidance. Our findings highlight the role of sentiment, media richness, and source credibility in shaping the spread of public health messages. We propose design strategies to help the CDC tailor communications to diverse user groups and manage misinformation more effectively during high-stakes health crises. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:53:33 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Ammari",
"Tawfiq",
""
],
[
"Gutowska",
"Anna",
""
],
[
"Ziff",
"Jacob",
""
],
[
"Randazzo",
"Casey",
""
],
[
"Subramonyam",
"Harihan",
""
]
] |
2505.22033 | Zhuang ZeMing | Zeming Zhuang, Kun Meng, Hongsheng Zhang | Phase Transition and Critical Phenomena of Charged Einstein-Maxwell-Scalar Black Holes | 28 pages, 6 figures | null | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | We study the phase transition and critical phenomenon of charged black holes in Einstein-Maxwell-scalar (EMs) theory. Through comprehensive analysis of thermodynamic behaviors manifested in P-V diagrams, G(T,P) surfaces, and C_P curves, we establish that these black holes exhibit van der Waals-type phase transition behavior. The derived critical exponents governing the phase transition show precise correspondence with both van der Waals gas-liquid systems, reinforcing the connection between black hole thermodynamics and mean field theory statistics. The findings reveal a crucial dependence of phase transition properties on the scalar charge parameter. A critical threshold emerges where phase transitions become prohibited when scalar charge exceeds a specific magnitude. However, the transition persists asymptotically as scalar charge approaches zero. The analysis further demonstrates nonlinear relationships between scalar charge and critical parameters: while small scalar charges induce increasing critical volume with charge magnitude, larger values produce an inverse trend. Critical temperature displays complementary behavior, maintaining monotonic variation under certain conditions while exhibiting inverse correlation with critical volume in others. Significantly, the transition points governing critical volume and temperature trends occur at distinct scalar charge values for different black holes, indicating a non-trivial parameter dependence. These results highlight the scalar charge's dual role as both an enabler and suppressor of phase transitions in EMs black holes, providing new insights into the interplay between geometric configurations and thermodynamic properties in modified gravity theories. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:53:51 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhuang",
"Zeming",
""
],
[
"Meng",
"Kun",
""
],
[
"Zhang",
"Hongsheng",
""
]
] |
2505.22034 | Oskar Simensen | Oskar H{\o}gberg Simensen and Dennis Christensen and Nils Lid Hjort | Random irregular histograms | null | null | null | null | stat.ME math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a new method of histogram construction, providing the first fully Bayesian approach to irregular histograms. Our procedure applies Bayesian model selection to a piecewise constant model of the underlying distribution, resulting in a method that selects both the number of bins as well as their location based on the data in a fully automatic fashion. We show that the histogram estimate is consistent with respect to the Hellinger metric under mild regularity conditions, and that it attains a convergence rate equal to the minimax rate (up to a logarithmic factor) for H\"{o}lder continuous densities. Simulation studies indicate that the new method performs comparably to other histogram procedures, both for minimizing the estimation error and for identifying modes. A software implementation is included as supplementary material. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:55:58 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Simensen",
"Oskar Høgberg",
""
],
[
"Christensen",
"Dennis",
""
],
[
"Hjort",
"Nils Lid",
""
]
] |
2505.22035 | Xinyi Chen | Xinyi Chen, Chenxiang Ma, Yujie Wu, Kay Chen Tan, Jibin Wu | Neuromorphic Sequential Arena: A Benchmark for Neuromorphic Temporal Processing | Accepted at 34th International Joint Conference on Artificial Intelligence (IJCAI 2025) | null | null | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Temporal processing is vital for extracting meaningful information from time-varying signals. Recent advancements in Spiking Neural Networks (SNNs) have shown immense promise in efficiently processing these signals. However, progress in this field has been impeded by the lack of effective and standardized benchmarks, which complicates the consistent measurement of technological advancements and limits the practical applicability of SNNs. To bridge this gap, we introduce the Neuromorphic Sequential Arena (NSA), a comprehensive benchmark that offers an effective, versatile, and application-oriented evaluation framework for neuromorphic temporal processing. The NSA includes seven real-world temporal processing tasks from a diverse range of application scenarios, each capturing rich temporal dynamics across multiple timescales. Utilizing NSA, we conduct extensive comparisons of recently introduced spiking neuron models and neural architectures, presenting comprehensive baselines in terms of task performance, training speed, memory usage, and energy efficiency. Our findings emphasize an urgent need for efficient SNN designs that can consistently deliver high performance across tasks with varying temporal complexities while maintaining low computational costs. NSA enables systematic tracking of advancements in neuromorphic algorithm research and paves the way for developing effective and efficient neuromorphic temporal processing systems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:57:40 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Chen",
"Xinyi",
""
],
[
"Ma",
"Chenxiang",
""
],
[
"Wu",
"Yujie",
""
],
[
"Tan",
"Kay Chen",
""
],
[
"Wu",
"Jibin",
""
]
] |
2505.22036 | Takahiro Tsushima | Tetsushi Ito, Daichi Takeuchi, Takahiro Tsushima | The $L$-polynomials of van der Geer--van der Vlugt curves in characteristic $2$ | null | null | null | null | math.NT math.AG | http://creativecommons.org/licenses/by/4.0/ | The van der Geer--van der Vlugt curves form a class of Artin--Schreier coverings of the projective line over finite fields. We provide an explicit formula for their $L$-polynomials in characteristic $2$, expressed in terms of characters of maximal abelian subgroups of associated Heisenberg groups. For this purpose, we develop new methods specific to characteristic $2$ that exploit the structure of the Heisenberg groups and the geometry of Lang torsors for $W_2$. As an application, we construct examples of curves in this family attaining the Hasse--Weil bound. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:59:05 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Ito",
"Tetsushi",
""
],
[
"Takeuchi",
"Daichi",
""
],
[
"Tsushima",
"Takahiro",
""
]
] |
2505.22037 | Jingyu Zhang | Jingyu Zhang, Ahmed Elgohary, Xiawei Wang, A S M Iftekhar, Ahmed Magooda, Benjamin Van Durme, Daniel Khashabi, Kyle Jackson | Jailbreak Distillation: Renewable Safety Benchmarking | Project page: https://aka.ms/jailbreak-distillation | null | null | null | cs.CL cs.CR cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) are rapidly deployed in critical applications, raising urgent needs for robust safety benchmarking. We propose Jailbreak Distillation (JBDistill), a novel benchmark construction framework that "distills" jailbreak attacks into high-quality and easily-updatable safety benchmarks. JBDistill utilizes a small set of development models and existing jailbreak attack algorithms to create a candidate prompt pool, then employs prompt selection algorithms to identify an effective subset of prompts as safety benchmarks. JBDistill addresses challenges in existing safety evaluation: the use of consistent evaluation prompts across models ensures fair comparisons and reproducibility. It requires minimal human effort to rerun the JBDistill pipeline and produce updated benchmarks, alleviating concerns on saturation and contamination. Extensive experiments demonstrate our benchmarks generalize robustly to 13 diverse evaluation models held out from benchmark construction, including proprietary, specialized, and newer-generation LLMs, significantly outperforming existing safety benchmarks in effectiveness while maintaining high separability and diversity. Our framework thus provides an effective, sustainable, and adaptable solution for streamlining safety evaluation. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 06:59:46 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Jingyu",
""
],
[
"Elgohary",
"Ahmed",
""
],
[
"Wang",
"Xiawei",
""
],
[
"Iftekhar",
"A S M",
""
],
[
"Magooda",
"Ahmed",
""
],
[
"Van Durme",
"Benjamin",
""
],
[
"Khashabi",
"Daniel",
""
],
[
"Jackson",
"Kyle",
""
]
] |
2505.22038 | Kaiyuan Li | Kaiyuan Li and Xiaoyue Chen and Chen Gao and Yong Li and Xinlei Chen | Balanced Token Pruning: Accelerating Vision Language Models Beyond Local Optimization | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large Vision-Language Models (LVLMs) have shown impressive performance across multi-modal tasks by encoding images into thousands of tokens. However, the large number of image tokens results in significant computational overhead, and the use of dynamic high-resolution inputs further increases this burden. Previous approaches have attempted to reduce the number of image tokens through token pruning, typically by selecting tokens based on attention scores or image token diversity. Through empirical studies, we observe that existing methods often overlook the joint impact of pruning on both the current layer's output (local) and the outputs of subsequent layers (global), leading to suboptimal pruning decisions. To address this challenge, we propose Balanced Token Pruning (BTP), a plug-and-play method for pruning vision tokens. Specifically, our method utilizes a small calibration set to divide the pruning process into multiple stages. In the early stages, our method emphasizes the impact of pruning on subsequent layers, whereas in the deeper stages, the focus shifts toward preserving the consistency of local outputs. Extensive experiments across various LVLMs demonstrate the broad effectiveness of our approach on multiple benchmarks. Our method achieves a 78% compression rate while preserving 96.7% of the original models' performance on average. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:00:50 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Kaiyuan",
""
],
[
"Chen",
"Xiaoyue",
""
],
[
"Gao",
"Chen",
""
],
[
"Li",
"Yong",
""
],
[
"Chen",
"Xinlei",
""
]
] |
2505.22039 | Shifang Zhao | Shifang Zhao, Yiheng Lin, Lu Han, Yao Zhao, Yunchao Wei | OmniAD: Detect and Understand Industrial Anomaly via Multimodal Reasoning | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | While anomaly detection has made significant progress, generating detailed analyses that incorporate industrial knowledge remains a challenge. To address this gap, we introduce OmniAD, a novel framework that unifies anomaly detection and understanding for fine-grained analysis. OmniAD is a multimodal reasoner that combines visual and textual reasoning processes. The visual reasoning provides detailed inspection by leveraging Text-as-Mask Encoding to perform anomaly detection through text generation without manually selected thresholds. Following this, Visual Guided Textual Reasoning conducts comprehensive analysis by integrating visual perception. To enhance few-shot generalization, we employ an integrated training strategy that combines supervised fine-tuning (SFT) with reinforcement learning (GRPO), incorporating three sophisticated reward functions. Experimental results demonstrate that OmniAD achieves a performance of 79.1 on the MMAD benchmark, surpassing models such as Qwen2.5-VL-7B and GPT-4o. It also shows strong results across multiple anomaly detection benchmarks. These results highlight the importance of enhancing visual perception for effective reasoning in anomaly understanding. All codes and models will be publicly available. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:02:15 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhao",
"Shifang",
""
],
[
"Lin",
"Yiheng",
""
],
[
"Han",
"Lu",
""
],
[
"Zhao",
"Yao",
""
],
[
"Wei",
"Yunchao",
""
]
] |
2505.22040 | Nachuan Xiao | Nachuan Xiao, Xiaoyin Hu, Xin Liu, Kim-Chuan Toh | A Hybrid Subgradient Method for Nonsmooth Nonconvex Bilevel Optimization | 27 pages | null | null | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we focus on the nonconvex-nonconvex bilevel optimization problem (BLO), where both upper-level and lower-level objectives are nonconvex, with the upper-level problem potentially being nonsmooth. We develop a two-timescale momentum-accelerated subgradient method (TMG) that employs two-timescale stepsizes, and establish its local convergence when initialized within a sufficiently small neighborhood of the feasible region. To develop a globally convergent algorithm for (BLO), we introduce a feasibility restoration scheme (FRG) that drives iterates toward the feasible region. Both (TMG) and (FRG) only require the first-order derivatives of the upper-level and lower-level objective functions, ensuring efficient computations in practice. We then develop a novel hybrid method that alternates between (TMG) and (FRG) and adaptively estimates its hyperparameters. Under mild conditions, we establish the global convergence properties of our proposed algorithm. Preliminary numerical experiments demonstrate the high efficiency and promising potential of our proposed algorithm. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:02:26 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Xiao",
"Nachuan",
""
],
[
"Hu",
"Xiaoyin",
""
],
[
"Liu",
"Xin",
""
],
[
"Toh",
"Kim-Chuan",
""
]
] |
2505.22041 | Michael Grohs | Michael Grohs, Adrian Rebmann, Jana-Rebecca Rehse | Detecting Undesired Process Behavior by Means of Retrieval Augmented Generation | Accepted at the BPM Forum, located at the International Conference on Business Process Management (BPM) 2025 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Conformance checking techniques detect undesired process behavior by comparing process executions that are recorded in event logs to desired behavior that is captured in a dedicated process model. If such models are not available, conformance checking techniques are not applicable, but organizations might still be interested in detecting undesired behavior in their processes. To enable this, existing approaches use Large Language Models (LLMs), assuming that they can learn to distinguish desired from undesired behavior through fine-tuning. However, fine-tuning is highly resource-intensive and the fine-tuned LLMs often do not generalize well. To address these limitations, we propose an approach that requires neither a dedicated process model nor resource-intensive fine-tuning to detect undesired process behavior. Instead, we use Retrieval Augmented Generation (RAG) to provide an LLM with direct access to a knowledge base that contains both desired and undesired process behavior from other processes, assuming that the LLM can transfer this knowledge to the process at hand. Our evaluation shows that our approach outperforms fine-tuned LLMs in detecting undesired behavior, demonstrating that RAG is a viable alternative to resource-intensive fine-tuning, particularly when enriched with relevant context from the event log, such as frequent traces and activities. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:03:46 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Grohs",
"Michael",
""
],
[
"Rebmann",
"Adrian",
""
],
[
"Rehse",
"Jana-Rebecca",
""
]
] |
2505.22042 | Hao Yang | Hao Yang, Haoxuan Li, Mengyue Yang, Xu Chen, Mingming Gong | Estimating the Effects of Sample Training Orders for Large Language Models without Retraining | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The order of training samples plays a crucial role in large language models (LLMs), significantly impacting both their external performance and internal learning dynamics. Traditional methods for investigating this effect generally require retraining the model with various sample orders, which is computationally infeasible for LLMs. In this work, we improve traditional methods by designing a retraining-free framework. By approximating Adam optimizer updates with first- and second-order Taylor expansions and utilizing random projection methods to store intermediate checkpoints, our framework can efficiently estimate model parameters for arbitrary training sample orders. Next, we apply our framework to two downstream research problems: (1) Training curriculum design for LLMs -- we base our retraining-free framework to propose a novel curriculum learning strategy that augments curriculum proposals with estimated model performances, enabling more informed sample scheduling. (2) LLMs' memorization and generalization effect analysis -- we use our retraining-free framework to estimate how the positions of training samples influence LLMs' capacity for memorization and generalization. We conduct extensive experiments to validate the effectiveness of our retraining-free framework in reproducing the true model performances, and further demonstrate its potential in optimizing LLM training curricula and analyzing the memorization and generalization effects of LLMs. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:07:02 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Yang",
"Hao",
""
],
[
"Li",
"Haoxuan",
""
],
[
"Yang",
"Mengyue",
""
],
[
"Chen",
"Xu",
""
],
[
"Gong",
"Mingming",
""
]
] |
2505.22043 | Timofei Snegirev | Timofei Snegirev | Perfect fluid equations with N=1,2 Schrodinger supersymmetry | 1+15 pages | null | null | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Superconformal extensions of the perfect fluid equations, which realize $N=1,2$ Schrodinger superalgebra, are constructed within the Hamiltonian formalism. They are built by introducing real (for $N=1$) or complex (for $N=2$) anticommuting field variables as superpartners for the density and velocity of a fluid. The full set of conserved charges associated with the $N=1,2$ Schrodinger superalgebra is constructed. Within the Lagrangian formalism, when the Clebsch decomposition for the velocity vector field is used, the anticommuting variables can be interpreted as potentials parameterizing fluid's vorticity. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:07:06 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Snegirev",
"Timofei",
""
]
] |
2505.22044 | Vladimir Ladygin | L.L. Kurchaninov (3), E.A. Ladygin (1), V.P. Ladygin (1), A.A. Semak (2) ((1) Joint Institute for Nuclear Research, (2) Institute for High Energy Physics, National Research Center Kurchatov Institute, (3) TRIUMF) | Design and implementation of the constant fraction discriminator for glass MRPC timing | 19 pages, 16 figures | null | null | null | physics.ins-det | http://creativecommons.org/publicdomain/zero/1.0/ | The analog front-end electronics based on the constant fraction discrimination method is designed and optimized for the Multigap Resistive Plate Chamber (MRPC) timing measurements. The total time resolution of 40 ps has been obtained for 10 and 12 gaps MRPCs using cosmic setup and a muon beam at the IHEP U-70 accelerator in Protvino, which complies with the conditions of the SPD experiment at NICA. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:07:34 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Kurchaninov",
"L. L.",
"",
"TRIUMF"
],
[
"Ladygin",
"E. A.",
"",
"Joint Institute for Nuclear Research"
],
[
"Ladygin",
"V. P.",
"",
"Joint Institute for Nuclear Research"
],
[
"Semak",
"A. A.",
"",
"Institute for High Energy Physics, National Research Center Kurchatov Institute"
]
] |
2505.22045 | Le Xu | Le Xu, Chenxing Li, Yong Ren, Yujie Chen, Yu Gu, Ruibo Fu, Shan Yang, Dong Yu | Mitigating Audiovisual Mismatch in Visual-Guide Audio Captioning | Accepted by INTERSPEECH 2025 | null | null | null | cs.MM cs.CV cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current vision-guided audio captioning systems frequently fail to address audiovisual misalignment in real-world scenarios, such as dubbed content or off-screen sounds. To bridge this critical gap, we present an entropy-aware gated fusion framework that dynamically modulates visual information flow through cross-modal uncertainty quantification. Our novel approach employs attention entropy analysis in cross-attention layers to automatically identify and suppress misleading visual cues during modal fusion. Complementing this architecture, we develop a batch-wise audiovisual shuffling technique that generates synthetic mismatched training pairs, greatly enhancing model resilience against alignment noise. Evaluations on the AudioCaps benchmark demonstrate our system's superior performance over existing baselines, especially in mismatched modality scenarios. Furthermore, our solution demonstrates an approximately 6x improvement in inference speed compared to the baseline. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:08:17 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Xu",
"Le",
""
],
[
"Li",
"Chenxing",
""
],
[
"Ren",
"Yong",
""
],
[
"Chen",
"Yujie",
""
],
[
"Gu",
"Yu",
""
],
[
"Fu",
"Ruibo",
""
],
[
"Yang",
"Shan",
""
],
[
"Yu",
"Dong",
""
]
] |
2505.22046 | Morteza Ghahremani | Ashkan Taghipour and Morteza Ghahremani and Mohammed Bennamoun and Farid Boussaid and Aref Miri Rekavandi and Zinuo Li and Qiuhong Ke and Hamid Laga | LatentMove: Towards Complex Human Movement Video Generation | 12 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image-to-video (I2V) generation seeks to produce realistic motion sequences from a single reference image. Although recent methods exhibit strong temporal consistency, they often struggle when dealing with complex, non-repetitive human movements, leading to unnatural deformations. To tackle this issue, we present LatentMove, a DiT-based framework specifically tailored for highly dynamic human animation. Our architecture incorporates a conditional control branch and learnable face/body tokens to preserve consistency as well as fine-grained details across frames. We introduce Complex-Human-Videos (CHV), a dataset featuring diverse, challenging human motions designed to benchmark the robustness of I2V systems. We also introduce two metrics to assess the flow and silhouette consistency of generated videos with their ground truth. Experimental results indicate that LatentMove substantially improves human animation quality--particularly when handling rapid, intricate movements--thereby pushing the boundaries of I2V generation. The code, the CHV dataset, and the evaluation metrics will be available at https://github.com/ --. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:10:49 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Taghipour",
"Ashkan",
""
],
[
"Ghahremani",
"Morteza",
""
],
[
"Bennamoun",
"Mohammed",
""
],
[
"Boussaid",
"Farid",
""
],
[
"Rekavandi",
"Aref Miri",
""
],
[
"Li",
"Zinuo",
""
],
[
"Ke",
"Qiuhong",
""
],
[
"Laga",
"Hamid",
""
]
] |
2505.22047 | Claus Metzner | Claus Metzner, Achim Schilling, Andreas Maier, Patrick Krauss | Organizational Regularities in Recurrent Neural Networks | null | null | null | null | q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Previous work has shown that the dynamical regime of Recurrent Neural Networks (RNNs) - ranging from oscillatory to chaotic and fixpoint behavior - can be controlled by the global distribution of weights in connection matrices with statistically independent elements. However, it remains unclear how network dynamics respond to organizational regularities in the weight matrix, as often observed in biological neural networks. Here, we investigate three such regularities: (1) monopolar output weights per neuron, in accordance with Dale's principle, (2) reciprocal symmetry between neuron pairs, as in Hopfield networks, and (3) modular structure, where strongly connected blocks are embedded in a background of weaker connectivity. We construct weight matrices in which the strength of each regularity can be continuously tuned via control parameters, and analyze how key dynamical signatures of the RNN evolve as a function of these parameters. Moreover, using the RNN for actual information processing in a reservoir computing framework, we study how each regularity affects performance. We find that Dale monopolarity and modularity significantly enhance task accuracy, while Hopfield reciprocity tends to reduce it by promoting early saturation, limiting reservoir flexibility. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:13:43 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Metzner",
"Claus",
""
],
[
"Schilling",
"Achim",
""
],
[
"Maier",
"Andreas",
""
],
[
"Krauss",
"Patrick",
""
]
] |
2505.22048 | Haihan Zhang | Haihan Zhang, Weicheng Lin, Yuanshi Liu, Cong Fang | Learning Curves of Stochastic Gradient Descent in Kernel Regression | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper considers a canonical problem in kernel regression: how good are the model performances when it is trained by the popular online first-order algorithms, compared to the offline ones, such as ridge and ridgeless regression? In this paper, we analyze the foundational single-pass Stochastic Gradient Descent (SGD) in kernel regression under source condition where the optimal predictor can even not belong to the RKHS, i.e. the model is misspecified. Specifically, we focus on the inner product kernel over the sphere and characterize the exact orders of the excess risk curves under different scales of sample sizes $n$ concerning the input dimension $d$. Surprisingly, we show that SGD achieves min-max optimal rates up to constants among all the scales, without suffering the saturation, a prevalent phenomenon observed in (ridge) regression, except when the model is highly misspecified and the learning is in a final stage where $n\gg d^{\gamma}$ with any constant $\gamma >0$. The main reason for SGD to overcome the curse of saturation is the exponentially decaying step size schedule, a common practice in deep neural network training. As a byproduct, we provide the \emph{first} provable advantage of the scheme over the iterative averaging method in the common setting. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:16:11 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Haihan",
""
],
[
"Lin",
"Weicheng",
""
],
[
"Liu",
"Yuanshi",
""
],
[
"Fang",
"Cong",
""
]
] |
2505.22049 | Laetitia Chapel | Laetitia Chapel and Romain Tavenard and Samuel Vaiter | Differentiable Generalized Sliced Wasserstein Plans | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Optimal Transport (OT) has attracted significant interest in the machine learning community, not only for its ability to define meaningful distances between probability distributions -- such as the Wasserstein distance -- but also for its formulation of OT plans. Its computational complexity remains a bottleneck, though, and slicing techniques have been developed to scale OT to large datasets. Recently, a novel slicing scheme, dubbed min-SWGG, lifts a single one-dimensional plan back to the original multidimensional space, finally selecting the slice that yields the lowest Wasserstein distance as an approximation of the full OT plan. Despite its computational and theoretical advantages, min-SWGG inherits typical limitations of slicing methods: (i) the number of required slices grows exponentially with the data dimension, and (ii) it is constrained to linear projections. Here, we reformulate min-SWGG as a bilevel optimization problem and propose a differentiable approximation scheme to efficiently identify the optimal slice, even in high-dimensional settings. We furthermore define its generalized extension for accommodating to data living on manifolds. Finally, we demonstrate the practical value of our approach in various applications, including gradient flows on manifolds and high-dimensional spaces, as well as a novel sliced OT-based conditional flow matching for image generation -- where fast computation of transport plans is essential. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:18:08 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Chapel",
"Laetitia",
""
],
[
"Tavenard",
"Romain",
""
],
[
"Vaiter",
"Samuel",
""
]
] |
2505.22050 | Di Wu | Di Wu, Jiaxin Fan, Junzhe Zang, Guanbo Wang, Wei Yin, Wenhao Li, Bo Jin | Reinforced Reasoning for Embodied Planning | null | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Embodied planning requires agents to make coherent multi-step decisions based on dynamic visual observations and natural language goals. While recent vision-language models (VLMs) excel at static perception tasks, they struggle with the temporal reasoning, spatial understanding, and commonsense grounding needed for planning in interactive environments. In this work, we introduce a reinforcement fine-tuning framework that brings R1-style reasoning enhancement into embodied planning. We first distill a high-quality dataset from a powerful closed-source model and perform supervised fine-tuning (SFT) to equip the model with structured decision-making priors. We then design a rule-based reward function tailored to multi-step action quality and optimize the policy via Generalized Reinforced Preference Optimization (GRPO). Our approach is evaluated on Embench, a recent benchmark for interactive embodied tasks, covering both in-domain and out-of-domain scenarios. Experimental results show that our method significantly outperforms models of similar or larger scale, including GPT-4o-mini and 70B+ open-source baselines, and exhibits strong generalization to unseen environments. This work highlights the potential of reinforcement-driven reasoning to advance long-horizon planning in embodied AI. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:21:37 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Wu",
"Di",
""
],
[
"Fan",
"Jiaxin",
""
],
[
"Zang",
"Junzhe",
""
],
[
"Wang",
"Guanbo",
""
],
[
"Yin",
"Wei",
""
],
[
"Li",
"Wenhao",
""
],
[
"Jin",
"Bo",
""
]
] |
2505.22051 | Pengjie Shen | Pengjie Shen, Xueliang Zhang, Zhong-Qiu Wang | ARiSE: Auto-Regressive Multi-Channel Speech Enhancement | null | null | null | null | eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose ARiSE, an auto-regressive algorithm for multi-channel speech enhancement. ARiSE improves existing deep neural network (DNN) based frame-online multi-channel speech enhancement models by introducing auto-regressive connections, where the estimated target speech at previous frames is leveraged as extra input features to help the DNN estimate the target speech at the current frame. The extra input features can be derived from (a) the estimated target speech in previous frames; and (b) a beamformed mixture with the beamformer computed based on the previous estimated target speech. On the other hand, naively training the DNN in an auto-regressive manner is very slow. To deal with this, we propose a parallel training mechanism to speed up the training. Evaluation results in noisy-reverberant conditions show the effectiveness and potential of the proposed algorithms. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:22:28 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Shen",
"Pengjie",
""
],
[
"Zhang",
"Xueliang",
""
],
[
"Wang",
"Zhong-Qiu",
""
]
] |
2505.22052 | Philipp G\"orz | Keno Hassler, Philipp G\"orz, Stephan Lipp, Thorsten Holz, Marcel B\"ohme | A Comparative Study of Fuzzers and Static Analysis Tools for Finding Memory Unsafety in C and C++ | 34 pages, 10 figures | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | Even today, over 70% of security vulnerabilities in critical software systems result from memory safety violations. To address this challenge, fuzzing and static analysis are widely used automated methods to discover such vulnerabilities. Fuzzing generates random program inputs to identify faults, while static analysis examines source code to detect potential vulnerabilities. Although these techniques share a common goal, they take fundamentally different approaches and have evolved largely independently.
In this paper, we present an empirical analysis of five static analyzers and 13 fuzzers, applied to over 100 known security vulnerabilities in C/C++ programs. We measure the number of bug reports generated for each vulnerability to evaluate how the approaches differ and complement each other. Moreover, we randomly sample eight bug-containing functions, manually analyze all bug reports therein, and quantify false-positive rates. We also assess limits to bug discovery, ease of use, resource requirements, and integration into the development process. We find that both techniques discover different types of bugs, but there are clear winners for each. Developers should consider these tools depending on their specific workflow and usability requirements. Based on our findings, we propose future directions to foster collaboration between these research domains. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:22:29 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hassler",
"Keno",
""
],
[
"Görz",
"Philipp",
""
],
[
"Lipp",
"Stephan",
""
],
[
"Holz",
"Thorsten",
""
],
[
"Böhme",
"Marcel",
""
]
] |
2505.22053 | Yan Rong | Yan Rong, Jinting Wang, Shan Yang, Guangzhi Lei, Li Liu | AudioGenie: A Training-Free Multi-Agent Framework for Diverse Multimodality-to-Multiaudio Generation | null | null | null | null | cs.SD cs.MA cs.MM eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodality-to-Multiaudio (MM2MA) generation faces significant challenges in synthesizing diverse and contextually aligned audio types (e.g., sound effects, speech, music, and songs) from multimodal inputs (e.g., video, text, images), owing to the scarcity of high-quality paired datasets and the lack of robust multi-task learning frameworks. Recently, multi-agent system shows great potential in tackling the above issues. However, directly applying it to MM2MA task presents three critical challenges: (1) inadequate fine-grained understanding of multimodal inputs (especially for video), (2) the inability of single models to handle diverse audio events, and (3) the absence of self-correction mechanisms for reliable outputs. To this end, we propose AudioGenie, a novel training-free multi-agent system featuring a dual-layer architecture with a generation team and a supervisor team. For the generation team, a fine-grained task decomposition and an adaptive Mixture-of-Experts (MoE) collaborative entity are designed for dynamic model selection, and a trial-and-error iterative refinement module is designed for self-correction. The supervisor team ensures temporal-spatial consistency and verifies outputs through feedback loops. Moreover, we build MA-Bench, the first benchmark for MM2MA tasks, comprising 198 annotated videos with multi-type audios. Experiments demonstrate that our AudioGenie outperforms state-of-the-art (SOTA) methods across 9 metrics in 8 tasks. User study further validate the effectiveness of the proposed method in terms of quality, accuracy, alignment, and aesthetic. The anonymous project website with samples can be found at https://audiogenie.github.io/. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:23:53 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Rong",
"Yan",
""
],
[
"Wang",
"Jinting",
""
],
[
"Yang",
"Shan",
""
],
[
"Lei",
"Guangzhi",
""
],
[
"Liu",
"Li",
""
]
] |
2505.22054 | Jan Deriu | Samuel Stucki, Jan Deriu, Mark Cieliebak | Voice Adaptation for Swiss German | Submitted to Interspeech | null | null | null | cs.CL cs.SD eess.AS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This work investigates the performance of Voice Adaptation models for Swiss German dialects, i.e., translating Standard German text to Swiss German dialect speech. For this, we preprocess a large dataset of Swiss podcasts, which we automatically transcribe and annotate with dialect classes, yielding approximately 5000 hours of weakly labeled training material. We fine-tune the XTTSv2 model on this dataset and show that it achieves good scores in human and automated evaluations and can correctly render the desired dialect. Our work shows a step towards adapting Voice Cloning technology to underrepresented languages. The resulting model achieves CMOS scores of up to -0.28 and SMOS scores of 3.8. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:24:40 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Stucki",
"Samuel",
""
],
[
"Deriu",
"Jan",
""
],
[
"Cieliebak",
"Mark",
""
]
] |
2505.22055 | Vladislav Taranchuk | Jozefien D'haeseleer, Vladislav Taranchuk | On the Chromatic Number of Grassmann Graphs | null | null | null | null | math.CO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this paper we study the chromatic number of the Grassmann graphs $J_q(n, m)$. We show that $\binom{n-m+1}{1}_q \leq \chi(J_q(n, m)) \leq \binom{n}{1}_q$, which is analogous to the best-known bounds for the chromatic number of the Johnson graphs $J(n, m)$. When $m = 2$, determining $\chi(J_q(n, 2))$ is equivalent to determining the smallest number of partial line parallelisms that one can partition the lines of PG$(n-1, q)$ into. We survey known results about line parallelisms and their implications for $\chi(J_q(n, 2))$. Finally, we prove that when $q$ is any power of two, and $n$ is any even integer, then $\chi(J_q(n, 2)) < 2\binom{n-1}{1}_q$. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:24:56 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"D'haeseleer",
"Jozefien",
""
],
[
"Taranchuk",
"Vladislav",
""
]
] |
2505.22056 | Pablo Garcia-Fernandez | Nayara Carral-Sainz and Toraya Fern\'andez-Ruiz and Jorge \'I\~niguez and Javier Junquera and Pablo Garcia-Fernandez | Systematic generation of electron models for Second-Principles Density Functional Theory Methods | 23 pages, 13 figures | null | null | null | cond-mat.mtrl-sci cond-mat.other | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a systematic, quasi-automated methodology for generating electronic models in the framework of second-principles density functional theory (SPDFT). This approach enables the construction of accurate and computationally efficient models by deriving all necessary parameters from first-principles calculations on a carefully designed training set. A key feature of our method is the enforcement of space group symmetries, which reduces both the number of independent parameters and the required computational effort. The formalism includes improved treatments of one-electron Hamiltonians, electron-lattice coupling-through both linear and quadratic terms-and electron-electron interactions, enabling accurate modeling of structural and electronic responses. We apply the methodology to SrTiO$_{3}$ and LiF, materials representative of transition-metal perovskites and wide-band-gap insulators, respectively. In both cases, the resulting models reproduce DFT reference data with high fidelity across various atomic configurations and charge states. Our results validate the robustness of the approach and highlight its potential for simulating complex phenomena such as polarons and excitons. This work lays the foundation for extending SPDFT to real-time simulations of optoelectronic properties and further integration with machine-learning methods. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:27:04 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Carral-Sainz",
"Nayara",
""
],
[
"Fernández-Ruiz",
"Toraya",
""
],
[
"Íñiguez",
"Jorge",
""
],
[
"Junquera",
"Javier",
""
],
[
"Garcia-Fernandez",
"Pablo",
""
]
] |
2505.22057 | Yansen Zhang | Yansen Zhang, Xiaokun Zhang, Ziqiang Cui, and Chen Ma | Shapley Value-driven Data Pruning for Recommender Systems | In SIGKDD (2025), 10 pages | null | 10.1145/3711896.3737127 | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | Recommender systems often suffer from noisy interactions like accidental clicks or popularity bias. Existing denoising methods typically identify users' intent in their interactions, and filter out noisy interactions that deviate from the assumed intent. However, they ignore that interactions deemed noisy could still aid model training, while some ``clean'' interactions offer little learning value. To bridge this gap, we propose Shapley Value-driven Valuation (SVV), a framework that evaluates interactions based on their objective impact on model training rather than subjective intent assumptions. In SVV, a real-time Shapley value estimation method is devised to quantify each interaction's value based on its contribution to reducing training loss. Afterward, SVV highlights the interactions with high values while downplaying low ones to achieve effective data pruning for recommender systems. In addition, we develop a simulated noise protocol to examine the performance of various denoising approaches systematically. Experiments on four real-world datasets show that SVV outperforms existing denoising methods in both accuracy and robustness. Further analysis also demonstrates that our SVV can preserve training-critical interactions and offer interpretable noise assessment. This work shifts denoising from heuristic filtering to principled, model-driven interaction valuation. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:27:59 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Zhang",
"Yansen",
""
],
[
"Zhang",
"Xiaokun",
""
],
[
"Cui",
"Ziqiang",
""
],
[
"Ma",
"Chen",
""
]
] |
2505.22058 | Ye-Ping Jiang | Dezhi Song, Fuyang Hang, Gang Yao, Jun Zhang, Ye-Ping Jiang, Jin-Feng Jia | The experimental determination of exchange mass terms in surface states on both terminations of MnBi4Te7 | 19 pages,9 figures, including supporting materials | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The intrinsic antiferromagnetic topological insulators in the Mn-Bi-Te family, composed of superlattice-like MnBi2Te4/(Bi2Te3)n (n = 0, 1, 2, 3...) layered structure, present intriguing states of matter such as quantum anomalous Hall effect and the axion insulator. However, the surface state gap, which is the prerequisite for the observation of these states, remains elusive. Here by molecular beam epitaxy, we obtain two types of MnBi4Te7 films with the exclusive Bi2Te3 (BT) or MnBi2Te4 (MBT) terminations. By scanning tunneling spectroscopy, the mass terms in the surface states are identified on both surface terminations. Experimental results reveal the existence of a hybridization gap of approximately 23 meV in surface states on the BT termination. This gap comes from the hybridization between the surface states and the spin-split states in the adjacent MBT layer. On the MBT termination, an exchange mass term of about 30 meV in surface states is identified by taking magnetic-field-dependent Landau level spectra as well as theoretical simulations. In addition, the mass term varies with the field in the film with a heavy BiMn doping level in the Mn layers. These findings demonstrate the existence of mass terms in surface states on both types of terminations in our epitaxial MnBi4Te7 films investigated by local probes. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:28:58 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Song",
"Dezhi",
""
],
[
"Hang",
"Fuyang",
""
],
[
"Yao",
"Gang",
""
],
[
"Zhang",
"Jun",
""
],
[
"Jiang",
"Ye-Ping",
""
],
[
"Jia",
"Jin-Feng",
""
]
] |
2505.22059 | Emmanuel Kowalski | Emmanuel Kowalski and Th\'eo Untrau | Wasserstein metrics and quantitative equidistribution of exponential sums over finite fields | 36 pages | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Wasserstein distance between probability measures on compact spaces provides a natural invariant quantitative measure of equidistribution, which is partly similar to the classical discrepancy appearing in Erd\"os-Tur\'an type inequalities in the case of tori, but is a more intrinsic quantity. We recall the basic properties of Wasserstein distances and present applications to quantitative forms of equidistribution of exponential sums in two examples, one related to our previous work on the equidistribution of ultra-short exponential sums, and the second a quantitative form of the equidistribution theorems of Deligne and Katz. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:29:40 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Kowalski",
"Emmanuel",
""
],
[
"Untrau",
"Théo",
""
]
] |
2505.22060 | Lukas Schmidbauer | Lukas Schmidbauer and Wolfgang Mauerer | SAT Strikes Back: Parameter and Path Relations in Quantum Toolchains | null | null | null | null | quant-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In the foreseeable future, toolchains for quantum computing should offer automatic means of transforming a high level problem formulation down to a hardware executable form. Thereby, it is crucial to find (multiple) transformation paths that are optimised for (hardware specific) metrics. We zoom into this pictured tree of transformations by focussing on k-SAT instances as input and their transformation to QUBO, while considering structure and characteristic metrics of input, intermediate and output representations. Our results can be used to rate valid paths of transformation in advance -- also in automated (quantum) toolchains. We support the automation aspect by considering stability and therefore predictability of free parameters and transformation paths. Moreover, our findings can be used in the manifesting era of error correction (since considering structure in a high abstraction layer can benefit error correcting codes in layers below). We also show that current research is closely linked to quadratisation techniques and their mathematical foundation. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:32:37 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Schmidbauer",
"Lukas",
""
],
[
"Mauerer",
"Wolfgang",
""
]
] |
2505.22061 | Yujin Choi | Yujin Choi, Youngjoo Park, Junyoung Byun, Jaewook Lee, and Jinseong Park | Safeguarding Privacy of Retrieval Data against Membership Inference Attacks: Is This Query Too Close to Home? | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Retrieval-augmented generation (RAG) mitigates the hallucination problem in large language models (LLMs) and has proven effective for specific, personalized applications. However, passing private retrieved documents directly to LLMs introduces vulnerability to membership inference attacks (MIAs), which try to determine whether the target datum exists in the private external database or not. Based on the insight that MIA queries typically exhibit high similarity to only one target document, we introduce Mirabel, a similarity-based MIA detection framework designed for the RAG system. With the proposed Mirabel, we show that simple detect-and-hide strategies can successfully obfuscate attackers, maintain data utility, and remain system-agnostic. We experimentally prove its detection and defense against various state-of-the-art MIA methods and its adaptability to existing private RAG systems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:35:07 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Choi",
"Yujin",
""
],
[
"Park",
"Youngjoo",
""
],
[
"Byun",
"Junyoung",
""
],
[
"Lee",
"Jaewook",
""
],
[
"Park",
"Jinseong",
""
]
] |
2505.22062 | Jiahao Hu | Jiahao Hu, Ruiyang Zhang, Zhiliang Chen, Yunpeng Lu, Qun Ouyang, Lailin Xu | Characterization of the JadePix-3 CMOS Pixel Sensor for the CEPC Vertex Detector | 34 pages, 31 figures, preprint submitted to NIM-A | null | null | null | physics.ins-det hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Circular Electron-Positron Collider (CEPC), a proposed next-generation $e^+e^-$ collider to enable high-precision studies of the Higgs boson and potential new physics, imposes rigorous demands on detector technologies, particularly the vertex detector. JadePix-3, a prototype Monolithic Active Pixel Sensor (MAPS) designed for the CEPC vertex detector, targets a spatial resolution of 3 $\rm \mu m$ to meet the key requirement. This paper presents a detailed laboratory-based characterization of the JadePix-3 sensor, focusing on the previously under-explored effects of substrate reverse bias voltage on key performance metrics: charge collection efficiency, average cluster size, and detection efficiency. Systematic testing demonstrated that JadePix-3 operates reliably under reverse bias, exhibiting a reduced input capacitance, an expanded depletion region, enhanced charge collection efficiency, and a lower fake-hit rate. Simultaneously, an extensive methodology for MAPS characterization is established, which determined the optimal operational working conditions for the JadePix-3 sensor. Furthermore, these findings confirm the sensor's potential for high-precision particle tracking and vertexing at the CEPC while offering valuable references for future iterational R&D of the JadePix series. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:37:35 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Hu",
"Jiahao",
""
],
[
"Zhang",
"Ruiyang",
""
],
[
"Chen",
"Zhiliang",
""
],
[
"Lu",
"Yunpeng",
""
],
[
"Ouyang",
"Qun",
""
],
[
"Xu",
"Lailin",
""
]
] |
2505.22063 | Mingchen Shao | Mingchen Shao, Xinfa Zhu, Chengyou Wang, Bingshen Mu, Hai Li, Ying Yan, Junhui Liu, Danming Xie, Lei Xie | Weakly Supervised Data Refinement and Flexible Sequence Compression for Efficient Thai LLM-based ASR | Accepted by INTERSPEECH 2025 | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite remarkable achievements, automatic speech recognition (ASR) in low-resource scenarios still faces two challenges: high-quality data scarcity and high computational demands. This paper proposes EThai-ASR, the first to apply large language models (LLMs) to Thai ASR and create an efficient LLM-based ASR system. EThai-ASR comprises a speech encoder, a connection module and a Thai LLM decoder. To address the data scarcity and obtain a powerful speech encoder, EThai-ASR introduces a self-evolving data refinement strategy to refine weak labels, yielding an enhanced speech encoder. Moreover, we propose a pluggable sequence compression module used in the connection module with three modes designed to reduce the sequence length, thus decreasing computational demands while maintaining decent performance. Extensive experiments demonstrate that EThai-ASR has achieved state-of-the-art accuracy in multiple datasets. We release our refined text transcripts to promote further research. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:39:25 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Shao",
"Mingchen",
""
],
[
"Zhu",
"Xinfa",
""
],
[
"Wang",
"Chengyou",
""
],
[
"Mu",
"Bingshen",
""
],
[
"Li",
"Hai",
""
],
[
"Yan",
"Ying",
""
],
[
"Liu",
"Junhui",
""
],
[
"Xie",
"Danming",
""
],
[
"Xie",
"Lei",
""
]
] |
2505.22064 | Gunter Malle | Zhicheng Feng, Gunter Malle, Jiping Zhang | Generic weights for finite reductive groups | null | null | null | null | math.RT math.GR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper is motivated by the study of Alperin's weight conjecture in the representation theory of finite groups. We generalize the notion of $e$-cuspidality in the $e$-Harish-Chandra theory of finite reductive groups, and define generic weights in non-defining characteristic. We show that the generic weights play an analogous role as the weights defined by Alperin in the investigation of the inductive Alperin weight condition for simple groups of Lie type at most good primes. We hope that our approach will constitute a step towards an eventual proof of Alperin's weight conjecture. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:44:16 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Feng",
"Zhicheng",
""
],
[
"Malle",
"Gunter",
""
],
[
"Zhang",
"Jiping",
""
]
] |
2505.22065 | Mikko Impi\"o | Mikko Impi\"o, Philipp M. Rehsen, Tiina Laamanen, Arne J. Beermann, Florian Leese, Jenni Raitoharju | AquaMonitor: A multimodal multi-view image sequence dataset for real-life aquatic invertebrate biodiversity monitoring | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This paper presents the AquaMonitor dataset, the first large computer vision dataset of aquatic invertebrates collected during routine environmental monitoring. While several large species identification datasets exist, they are rarely collected using standardized collection protocols, and none focus on aquatic invertebrates, which are particularly laborious to collect. For AquaMonitor, we imaged all specimens from two years of monitoring whenever imaging was possible given practical limitations. The dataset enables the evaluation of automated identification methods for real-life monitoring purposes using a realistically challenging and unbiased setup. The dataset has 2.7M images from 43,189 specimens, DNA sequences for 1358 specimens, and dry mass and size measurements for 1494 specimens, making it also one of the largest biological multi-view and multimodal datasets to date. We define three benchmark tasks and provide strong baselines for these: 1) Monitoring benchmark, reflecting real-life deployment challenges such as open-set recognition, distribution shift, and extreme class imbalance, 2) Classification benchmark, which follows a standard fine-grained visual categorization setup, and 3) Few-shot benchmark, which targets classes with only few training examples from very fine-grained categories. Advancements on the Monitoring benchmark can directly translate to improvement of aquatic biodiversity monitoring, which is an important component of regular legislative water quality assessment in many countries. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:45:20 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Impiö",
"Mikko",
""
],
[
"Rehsen",
"Philipp M.",
""
],
[
"Laamanen",
"Tiina",
""
],
[
"Beermann",
"Arne J.",
""
],
[
"Leese",
"Florian",
""
],
[
"Raitoharju",
"Jenni",
""
]
] |
2505.22066 | Hanyu Cheng | Hanyu Cheng, Eleonora Di Valentino, Luca Visinelli | Cosmic Strings as Dynamical Dark Energy: Novel Constraints | 21 pages including references, 5 tables, and 6 figures | null | null | null | astro-ph.CO hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cosmic strings, topological defects predicted by high-energy theories, may contribute to the late-time expansion of the Universe, effectively mimicking dynamical dark energy. We investigate four phenomenological extensions of the $\Lambda$CDM model involving a residual string network: (i) a non-relativistic component with positive energy density (Model~1), (ii) a velocity-dependent extension (Model~2), (iii) a non-relativistic string network with energy density allowed to take both positive and negative values (Model~3), and (iv) a general scenario with free energy and velocity parameters (Model~4). These models are constrained using \textit{Planck} CMB data, SDSS or DESI baryon acoustic oscillations, and Type Ia supernovae observations. Models~1 and~2 yield strong upper bounds on the string density, for example, $\Omega_{\mathrm{s}} < 0.00901$ at 95\% CL from the CMB+DESI+DESY5 combination for Model~2, and mildly shift the inferred value of $H_0$ upward, though they are not favored by Bayesian evidence. For the same combination, the bulk velocity is bound as $v_{\mathrm{s}} < 0.569$. Models~3 and~4 exhibit a consistent preference for slightly negative values of $\Omega_{\mathrm{s}}$, with CMB-only data yielding $\Omega_{\mathrm{s}} = -0.038^{+0.029}_{-0.022}$ and $v_{\mathrm{s}}< 0.574$ in Model~4, and a best-fit improvement of $\Delta \chi^2 = -6.07$. However, these improvements are not sufficient to overcome the Occam penalty, and the Bayesian evidence continues to favor $\Lambda$CDM. These findings demonstrate the power of current data to constrain exotic energy components and encourage further exploration of string-inspired extensions to $\Lambda$CDM, particularly those involving negative-tension networks. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:45:53 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Cheng",
"Hanyu",
""
],
[
"Di Valentino",
"Eleonora",
""
],
[
"Visinelli",
"Luca",
""
]
] |
2505.22067 | Xinyu Xia | Xinyu Xia, Xingjun Ma, Yunfeng Hu, Ting Qu, Hong Chen, Xun Gong | From Failures to Fixes: LLM-Driven Scenario Repair for Self-Evolving Autonomous Driving | null | null | null | null | cs.CV cs.AI cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ensuring robust and generalizable autonomous driving requires not only broad scenario coverage but also efficient repair of failure cases, particularly those related to challenging and safety-critical scenarios. However, existing scenario generation and selection methods often lack adaptivity and semantic relevance, limiting their impact on performance improvement. In this paper, we propose \textbf{SERA}, an LLM-powered framework that enables autonomous driving systems to self-evolve by repairing failure cases through targeted scenario recommendation. By analyzing performance logs, SERA identifies failure patterns and dynamically retrieves semantically aligned scenarios from a structured bank. An LLM-based reflection mechanism further refines these recommendations to maximize relevance and diversity. The selected scenarios are used for few-shot fine-tuning, enabling targeted adaptation with minimal data. Experiments on the benchmark show that SERA consistently improves key metrics across multiple autonomous driving baselines, demonstrating its effectiveness and generalizability under safety-critical conditions. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:46:19 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Xia",
"Xinyu",
""
],
[
"Ma",
"Xingjun",
""
],
[
"Hu",
"Yunfeng",
""
],
[
"Qu",
"Ting",
""
],
[
"Chen",
"Hong",
""
],
[
"Gong",
"Xun",
""
]
] |
2505.22068 | Ran Li | Ran Li, Shimin Di, Yuchen Liu, Chen Jing, Yu Qiu, Lei Chen | Beyond path selection: Better LLMs for Scientific Information Extraction with MimicSFT and Relevance and Rule-induced(R$^2$)GRPO | null | null | null | null | cs.CL cs.AI cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Previous study suggest that powerful Large Language Models (LLMs) trained with Reinforcement Learning with Verifiable Rewards (RLVR) only refines reasoning path without improving the reasoning capacity in math tasks while supervised-finetuning(SFT) with distillation can. We study this from the view of Scientific information extraction (SciIE) where LLMs and reasoning LLMs underperforms small Bert-based models. SciIE require both the reasoning and memorization. We argue that both SFT and RLVR can refine the reasoning path and improve reasoning capacity in a simple way based on SciIE. We propose two-stage training with 1. MimicSFT, using structured reasoning templates without needing high-quality chain-of-thought data, 2. R$^2$GRPO with relevance and rule-induced rewards. Experiments on scientific IE benchmarks show that both methods can improve the reasoning capacity. R$^2$GRPO with mimicSFT surpasses baseline LLMs and specialized supervised models in relation extraction. Our code is available at https://github.com/ranlislz/R2GRPO. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:47:46 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Ran",
""
],
[
"Di",
"Shimin",
""
],
[
"Liu",
"Yuchen",
""
],
[
"Jing",
"Chen",
""
],
[
"Qiu",
"Yu",
""
],
[
"Chen",
"Lei",
""
]
] |
2505.22069 | Longhao Li | Longhao Li, Yangze Li, Hongfei Xue, Jie Liu, Shuai Fang, Kai Wang, Lei Xie | Delayed-KD: Delayed Knowledge Distillation based CTC for Low-Latency Streaming ASR | Accepted by Interspeech2025 | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | CTC-based streaming ASR has gained significant attention in real-world applications but faces two main challenges: accuracy degradation in small chunks and token emission latency. To mitigate these challenges, we propose Delayed-KD, which applies delayed knowledge distillation on CTC posterior probabilities from a non-streaming to a streaming model. Specifically, with a tiny chunk size, we introduce a Temporal Alignment Buffer (TAB) that defines a relative delay range compared to the non-streaming teacher model to align CTC outputs and mitigate non-blank token mismatches. Additionally, TAB enables fine-grained control over token emission delay. Experiments on 178-hour AISHELL-1 and 10,000-hour WenetSpeech Mandarin datasets show consistent superiority of Delayed-KD. Impressively, Delayed-KD at 40 ms latency achieves a lower character error rate (CER) of 5.42% on AISHELL-1, comparable to the competitive U2++ model running at 320 ms latency. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:51:21 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Longhao",
""
],
[
"Li",
"Yangze",
""
],
[
"Xue",
"Hongfei",
""
],
[
"Liu",
"Jie",
""
],
[
"Fang",
"Shuai",
""
],
[
"Wang",
"Kai",
""
],
[
"Xie",
"Lei",
""
]
] |
2505.22070 | Hendra Nurdin | Hendra I. Nurdin | Physical Reduced Stochastic Equations for Continuously Monitored Non-Markovian Quantum Systems with a Markovian Embedding | 7 pages, no figures. Accepted for publication in IEEE Control Systems Letters (http://ieee-cssletters.dei.unipd.it/) | null | null | null | quant-ph cs.SY eess.SY math-ph math.MP math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An effective approach to modeling non-Markovian quantum systems is to embed a principal (quantum) system of interest into a larger quantum system. A widely employed embedding is one that uses another quantum system, referred to as the auxiliary system, which is coupled to the principal system, and both the principal and auxiliary can be coupled to quantum white noise processes. The principal and auxiliary together form a quantum Markov system and the quantum white noises act as a bath (environment) for this system.
Recently it was shown that the conditional evolution of the principal system in this embedding under continuous monitoring by a travelling quantum probe can be expressed as a system of coupled stochastic differential equations (SDEs) that involve only operators of the principal system. The reduced conditional state of the principal only (conditioned on the measurement outcomes) are determined by the "diagonal" blocks of this coupled systems of SDEs. It is shown here that the "off-diagonal" blocks can be exactly eliminated up to their initial conditions, leaving a reduced closed system of SDEs for the diagonal blocks only. Under additional conditions the off-diagonal initial conditions can be made to vanish. This new closed system of equations, which includes an integration term involving a two-time stochastic kernel, represents the non-Markovian stochastic dynamics of the principal system under continuous-measurement. The system of equations determine the reduced conditional state of the principal only and may be viewed as a stochastic Nakajima-Zwanzig type of equation for continuously monitored non-Markovian quantum systems. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:52:37 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Nurdin",
"Hendra I.",
""
]
] |
2505.22071 | Ruiqi Shu | Ruiqi Shu, Yuan Gao, Hao Wu, Ruijian Gou, Yanfei Xiang, Fan Xu, Qingsong Wen, Xian Wu, Xiaomeng Huang | Ocean-E2E: Hybrid Physics-Based and Data-Driven Global Forecasting of Extreme Marine Heatwaves with End-to-End Neural Assimilation | null | null | null | null | physics.geo-ph | http://creativecommons.org/licenses/by/4.0/ | This work focuses on the end-to-end forecast of global extreme marine heatwaves (MHWs), which are unusually warm sea surface temperature events with profound impacts on marine ecosystems. Accurate prediction of extreme MHWs has significant scientific and financial worth. However, existing methods still have certain limitations, especially in the most extreme MHWs. In this study, to address these issues, based on the physical nature of MHWs, we created a novel hybrid data-driven and numerical MHWs forecast framework Ocean-E2E, which is capable of 40-day accurate MHW forecasting with end-to-end data assimilation. Our framework significantly improves the forecast ability of extreme MHWs by explicitly modeling the effect of oceanic mesoscale advection and air-sea interaction based on a differentiable dynamic kernel. Furthermore, Ocean-E2E is capable of end-to-end MHWs forecast and regional high-resolution prediction using neural data assimilation approaches, allowing our framework to operate completely independently of numerical models while demonstrating high assimilation stability and accuracy, outperforming the current state-of-the-art ocean numerical forecasting-assimilation models. Experimental results show that the proposed framework performs excellently on global-to-regional scales and short-to-long-term forecasts, especially in those most extreme MHWs. Overall, our model provides a framework for forecasting and understanding MHWs and other climate extremes. Our codes are available at https://github.com/ChiyodaMomo01/Ocean-E2E. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:52:41 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Shu",
"Ruiqi",
""
],
[
"Gao",
"Yuan",
""
],
[
"Wu",
"Hao",
""
],
[
"Gou",
"Ruijian",
""
],
[
"Xiang",
"Yanfei",
""
],
[
"Xu",
"Fan",
""
],
[
"Wen",
"Qingsong",
""
],
[
"Wu",
"Xian",
""
],
[
"Huang",
"Xiaomeng",
""
]
] |
2505.22072 | Shujie Hu | Shujie HU, Xurong Xie, Mengzhe Geng, Jiajun Deng, Huimeng Wang, Guinan Li, Chengxi Deng, Tianzi Wang, Mingyu Cui, Helen Meng, Xunying Liu | On-the-fly Routing for Zero-shot MoE Speaker Adaptation of Speech Foundation Models for Dysarthric Speech Recognition | Accepted by Interspeech 2025 | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes a novel MoE-based speaker adaptation framework for foundation models based dysarthric speech recognition. This approach enables zero-shot adaptation and real-time processing while incorporating domain knowledge. Speech impairment severity and gender conditioned adapter experts are dynamically combined using on-the-fly predicted speaker-dependent routing parameters. KL-divergence is used to further enforce diversity among experts and their generalization to unseen speakers. Experimental results on the UASpeech corpus suggest that on-the-fly MoE-based adaptation produces statistically significant WER reductions of up to 1.34% absolute (6.36% relative) over the unadapted baseline HuBERT/WavLM models. Consistent WER reductions of up to 2.55% absolute (11.44% relative) and RTF speedups of up to 7 times are obtained over batch-mode adaptation across varying speaker-level data quantities. The lowest published WER of 16.35% (46.77% on very low intelligibility) is obtained. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:52:58 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"HU",
"Shujie",
""
],
[
"Xie",
"Xurong",
""
],
[
"Geng",
"Mengzhe",
""
],
[
"Deng",
"Jiajun",
""
],
[
"Wang",
"Huimeng",
""
],
[
"Li",
"Guinan",
""
],
[
"Deng",
"Chengxi",
""
],
[
"Wang",
"Tianzi",
""
],
[
"Cui",
"Mingyu",
""
],
[
"Meng",
"Helen",
""
],
[
"Liu",
"Xunying",
""
]
] |
2505.22073 | Megan Li | Megan Li, Wendy Bickersteth, Ningjing Tang, Jason Hong, Lorrie Cranor, Hong Shen, Hoda Heidari | A Closer Look at the Existing Risks of Generative AI: Mapping the Who, What, and How of Real-World Incidents | 21 pages including references and appendices, 5 figures, submitted to Conference on AI, Ethics, and Society (AIES 2025) | null | null | null | cs.CY | http://creativecommons.org/licenses/by/4.0/ | Due to its general-purpose nature, Generative AI is applied in an ever-growing set of domains and tasks, leading to an expanding set of risks of harm impacting people, communities, society, and the environment. These risks may arise due to failures during the design and development of the technology, as well as during its release, deployment, or downstream usages and appropriations of its outputs. In this paper, building on prior taxonomies of AI risks, harms, and failures, we construct a taxonomy specifically for Generative AI failures and map them to the harms they precipitate. Through a systematic analysis of 499 publicly reported incidents, we describe what harms are reported, how they arose, and who they impact. We report the prevalence of each type of harm, underlying failure mode, and harmed stakeholder, as well as their common co-occurrences. We find that most reported incidents are caused by use-related issues but bring harm to parties beyond the end user(s) of the Generative AI system at fault, and that the landscape of Generative AI harms is distinct from that of traditional AI. Our work offers actionable insights to policymakers, developers, and Generative AI users. In particular, we call for the prioritization of non-technical risk and harm mitigation strategies, including public disclosures and education and careful regulatory stances. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:54:06 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Megan",
""
],
[
"Bickersteth",
"Wendy",
""
],
[
"Tang",
"Ningjing",
""
],
[
"Hong",
"Jason",
""
],
[
"Cranor",
"Lorrie",
""
],
[
"Shen",
"Hong",
""
],
[
"Heidari",
"Hoda",
""
]
] |
2505.22074 | Sebastian Otte | Co\c{s}ku Can Horuz, Geoffrey Kasenbacher, Saya Higuchi, Sebastian Kairat, Jendrik Stoltz, Moritz Pesl, Bernhard A. Moser, Christoph Linse, Thomas Martinetz, Sebastian Otte | The Resurrection of the ReLU | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modeling sophisticated activation functions within deep learning architectures has evolved into a distinct research direction. Functions such as GELU, SELU, and SiLU offer smooth gradients and improved convergence properties, making them popular choices in state-of-the-art models. Despite this trend, the classical ReLU remains appealing due to its simplicity, inherent sparsity, and other advantageous topological characteristics. However, ReLU units are prone to becoming irreversibly inactive - a phenomenon known as the dying ReLU problem - which limits their overall effectiveness. In this work, we introduce surrogate gradient learning for ReLU (SUGAR) as a novel, plug-and-play regularizer for deep architectures. SUGAR preserves the standard ReLU function during the forward pass but replaces its derivative in the backward pass with a smooth surrogate that avoids zeroing out gradients. We demonstrate that SUGAR, when paired with a well-chosen surrogate function, substantially enhances generalization performance over convolutional network architectures such as VGG-16 and ResNet-18, providing sparser activations while effectively resurrecting dead ReLUs. Moreover, we show that even in modern architectures like Conv2NeXt and Swin Transformer - which typically employ GELU - substituting these with SUGAR yields competitive and even slightly superior performance. These findings challenge the prevailing notion that advanced activation functions are necessary for optimal performance. Instead, they suggest that the conventional ReLU, particularly with appropriate gradient handling, can serve as a strong, versatile revived classic across a broad range of deep learning vision models. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:55:51 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Horuz",
"Coşku Can",
""
],
[
"Kasenbacher",
"Geoffrey",
""
],
[
"Higuchi",
"Saya",
""
],
[
"Kairat",
"Sebastian",
""
],
[
"Stoltz",
"Jendrik",
""
],
[
"Pesl",
"Moritz",
""
],
[
"Moser",
"Bernhard A.",
""
],
[
"Linse",
"Christoph",
""
],
[
"Martinetz",
"Thomas",
""
],
[
"Otte",
"Sebastian",
""
]
] |
2505.22075 | Xiaoxing Ren | Xiaoxing Ren, Alessio Moreschini, Zhongda Chu, Yulong Gao, Thomas Parisini | Data-Driven Adjustable Robust Optimization | null | null | null | null | math.OC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we develop a two-stage data-driven approach to address the adjustable robust optimization problem, where the uncertainty set is adjustable to manage infeasibility caused by significant or poorly quantified uncertainties. In the first stage, we synthesize an uncertainty set to ensure the feasibility of the problem as much as possible using the collected uncertainty samples. In the second stage, we find the optimal solution while ensuring that the constraints are satisfied under the new uncertainty set. This approach enlarges the feasible state set, at the expense of the risk of possible constraint violation. We analyze two scenarios: one where the uncertainty is non-stochastic, and another where the uncertainty is stochastic but with unknown probability distribution, leading to a distributionally robust optimization problem. In the first case, we scale the uncertainty set and find the best subset that fits the uncertainty samples. In the second case, we employ the Wasserstein metric to quantify uncertainty based on training data, and for polytope uncertainty sets, we further provide a finite program reformulation of the problem. The effectiveness of the proposed methods is demonstrated through an optimal power flow problem. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:56:08 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Ren",
"Xiaoxing",
""
],
[
"Moreschini",
"Alessio",
""
],
[
"Chu",
"Zhongda",
""
],
[
"Gao",
"Yulong",
""
],
[
"Parisini",
"Thomas",
""
]
] |
2505.22076 | Timon Ziegenbein | Maja Stahl, Timon Ziegenbein, Joonsuk Park, Henning Wachsmuth | ArgInstruct: Specialized Instruction Fine-Tuning for Computational Argumentation | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Training large language models (LLMs) to follow instructions has significantly enhanced their ability to tackle unseen tasks. However, despite their strong generalization capabilities, instruction-following LLMs encounter difficulties when dealing with tasks that require domain knowledge. This work introduces a specialized instruction fine-tuning for the domain of computational argumentation (CA). The goal is to enable an LLM to effectively tackle any unseen CA tasks while preserving its generalization capabilities. Reviewing existing CA research, we crafted natural language instructions for 105 CA tasks to this end. On this basis, we developed a CA-specific benchmark for LLMs that allows for a comprehensive evaluation of LLMs' capabilities in solving various CA tasks. We synthesized 52k CA-related instructions, adapting the self-instruct process to train a CA-specialized instruction-following LLM. Our experiments suggest that CA-specialized instruction fine-tuning significantly enhances the LLM on both seen and unseen CA tasks. At the same time, performance on the general NLP tasks of the SuperNI benchmark remains stable. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:58:29 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Stahl",
"Maja",
""
],
[
"Ziegenbein",
"Timon",
""
],
[
"Park",
"Joonsuk",
""
],
[
"Wachsmuth",
"Henning",
""
]
] |
2505.22077 | Leon Abelmann | Indujan Sivanesarajaha and Leon Abelmann and Uwe Hartmann | Influence of thickness on magnetic properties of RF-sputtered amorphous CoNbZr thin films | null | null | null | null | cond-mat.mtrl-sci physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | Amorphous sputtered Co-based thin films are widely used as soft magnetic materials in applications such as sensors, inductors and magnetic flux concentrators. The magnetic properties of these films can be controlled by deposition parameters like film thickness, argon pressure, deposition rate and others. In this study, we present a detailed investigation of the magnetic properties of RF-sputtered Co$_{91}$Nb$_7$Zr$_2$ films with thicknesses ranging from 52 nm to 1040 nm. These amorphous films exhibit an average saturation magnetisation of 1.01(4) MA/m. As the film thickness increases, there is a significant decrease in coercivity, remanent-to-saturation magnetisation ratio M$_r$/M$_s$, and maximum permeability. The change in macroscopic magnetic properties is also reflected by the domain structure. At a thickness of 52 nm, the remanent domain state shows irregular domains, while films thicknesses above 208 nm exhibit flux-closure domain structures instead. The thickness-dependent modifications are attributed to the transition between N\'eel and Bloch type domain walls, which is expected to occur at approximately 84 nm. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 07:58:36 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Sivanesarajaha",
"Indujan",
""
],
[
"Abelmann",
"Leon",
""
],
[
"Hartmann",
"Uwe",
""
]
] |
2505.22078 | Pauline Vidal | Pauline Vidal and Emily Bourne and Virginie Grandgirard and Michel Mehrenberger and Eric Sonnendr\"ucker | Local cubic spline interpolation for Vlasov-type equations on a multi-patch geometry | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | We present a semi-Lagrangian method for the numerical resolution of Vlasov-type equations on multi-patch meshes. We employ a local cubic spline interpolation with Hermite boundary conditions between the patches. The derivative reconstruction is adapted to cope with non-uniform meshes as well as non-conforming situations. In the conforming case, there are no longer any constraints on the number of points for each patch; however, a small global system must now be solved. In that case, the local spline representations coincide with the corresponding global spline reconstruction. Alternatively, we can choose not to apply the global system and the derivatives can be approximated. The influence of the most distant points diminishes as the number of points per patch increases. For uniform per patch configurations, a study of the explicit and asymptotic behavior of this influence has been led. The method is validated using a two-dimensional guiding-center model with an O-point. All the numerical results are carried out in the Gyselalib++ library. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:00:15 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Vidal",
"Pauline",
""
],
[
"Bourne",
"Emily",
""
],
[
"Grandgirard",
"Virginie",
""
],
[
"Mehrenberger",
"Michel",
""
],
[
"Sonnendrücker",
"Eric",
""
]
] |
2505.22079 | Hanbin Ko | Hanbin Ko, Chang-Min Park | Bringing CLIP to the Clinic: Dynamic Soft Labels and Negation-Aware Learning for Medical Analysis | 16 pages (8 main, 2 references, 6 appendix), 13 figures. Accepted to CVPR 2025. This author-accepted manuscript includes an expanded ethics/data user agreement section. The final version will appear in the Proceedings of CVPR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The development of large-scale image-text pair datasets has significantly advanced self-supervised learning in Vision-Language Processing (VLP). However, directly applying general-domain architectures such as CLIP to medical data presents challenges, particularly in handling negations and addressing the inherent data imbalance of medical datasets. To address these issues, we propose a novel approach that integrates clinically-enhanced dynamic soft labels and medical graphical alignment, thereby improving clinical comprehension and the applicability of contrastive loss in medical contexts. Furthermore, we introduce negation-based hard negatives to deepen the model's understanding of the complexities of clinical language. Our approach is easily integrated into the medical CLIP training pipeline and achieves state-of-the-art performance across multiple tasks, including zero-shot, fine-tuned classification, and report retrieval. To comprehensively evaluate our model's capacity for understanding clinical language, we introduce CXR-Align, a benchmark uniquely designed to evaluate the understanding of negation and clinical information within chest X-ray (CXR) datasets. Experimental results demonstrate that our proposed methods are straightforward to implement and generalize effectively across contrastive learning frameworks, enhancing medical VLP capabilities and advancing clinical language understanding in medical imaging. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:00:18 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Ko",
"Hanbin",
""
],
[
"Park",
"Chang-Min",
""
]
] |
2505.22080 | Tomoo Kikuchi | Tomoo Kikuchi and Lien Pham | Trade Networks and the Rise of a Global Currency | null | null | null | null | econ.TH | http://creativecommons.org/licenses/by/4.0/ | We develop a model of strategic competition between global currencies. Issuers choose their commitment levels for currency internationalization, while users -- interconnected in a trade network -- choose their usage of each currency for trade settlement. Our theoretical findings highlight not only the advantage of the status-quo issuer in maintaining dominance, but also the conditions under which an emerging issuer can attract users, potentially leading to a multi-currency payment system. The network centrality of users plays a key role in shaping both their currency choices and the strategic commitment levels of issuers. Our framework offers testable implications for the share of global currencies for trade settlement by linking the network structure, the strategy of issuers and the currency choice of users. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:00:55 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Kikuchi",
"Tomoo",
""
],
[
"Pham",
"Lien",
""
]
] |
2505.22081 | Shun Sato | Shun Sato, Issei Sato | Can Test-time Computation Mitigate Memorization Bias in Neural Symbolic Regression? | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Symbolic regression aims to discover mathematical equations that fit given numerical data. It has been applied in various fields of scientific research, such as producing human-readable expressions that explain physical phenomena. Recently, Neural symbolic regression (NSR) methods that involve Transformers pre-trained on large-scale synthetic datasets have gained attention. While these methods offer advantages such as short inference time, they suffer from low performance, particularly when the number of input variables is large. In this study, we hypothesized that this limitation stems from the memorization bias of Transformers in symbolic regression. We conducted a quantitative evaluation of this bias in Transformers using a synthetic dataset and found that Transformers rarely generate expressions not present in the training data. Additional theoretical analysis reveals that this bias arises from the Transformer's inability to construct expressions compositionally while verifying their numerical validity. We finally examined if tailoring test-time strategies can lead to reduced memorization bias and better performance. We empirically demonstrate that providing additional information to the model at test time can significantly mitigate memorization bias. On the other hand, we also find that reducing memorization bias does not necessarily correlate with improved performance. These findings contribute to a deeper understanding of the limitations of NSR approaches and offer a foundation for designing more robust, generalizable symbolic regression methods. Code is available at https://github.com/Shun-0922/Mem-Bias-NSR . | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:01:25 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Sato",
"Shun",
""
],
[
"Sato",
"Issei",
""
]
] |
2505.22082 | Maret Einasto | Maret Einasto | Galaxy Superclusters and Their Complexes in the Cosmic Web | 50 pages, 22 figures, Universe, 2025, 11, 167 | Universe, 2025, 11, 167 | 10.3390/universe11060167 | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The richest and largest structures in the cosmic web are galaxy superclusters, their complexes (associations of several almost connected very rich superclusters), and planes. Superclusters represent a special environment where the evolution of galaxies and galaxy groups and clusters differs from the evolution of these systems in a low-density environment. The richest galaxy clusters reside in superclusters. The richest superclusters in the nearby Universe form a quasiregular pattern with the characteristic distance between superclusters 120 - 140 $h^{-1}$Mpc. Moreover, superclusters in the nearby Universe lie in two huge perpendicular planes with the extent of several hundreds of megaparsecs, the Local Supercluster plane and the Dominant supercluster plane. The origin of these patterns in the supercluster distribution is not yet clear, and it is an open question whether the presence of such structures can be explained within the $\Lambda$CDM cosmological model. This review presents a brief story of superclusters, their discovery, definitions, main properties, and large-scale distribution. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:04:49 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Einasto",
"Maret",
""
]
] |
2505.22083 | H. L. Dao | H. L. Dao | Hyperbolic recurrent neural network as the first type of non-Euclidean neural quantum state ansatz | null | null | null | null | quant-ph cond-mat.dis-nn cs.LG physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we introduce the first type of non-Euclidean neural quantum state (NQS) ansatz, in the form of the hyperbolic GRU (a variant of recurrent neural networks (RNNs)), to be used in the Variational Monte Carlo method of approximating the ground state wavefunction for quantum many-body systems. In particular, we examine the performances of NQS ansatzes constructed from both conventional or Euclidean RNN/GRU and from hyperbolic GRU in the prototypical settings of the one- and two-dimensional transverse field Ising models (TFIM) of up to 100 spins and the one-dimensional Heisenberg $J_1J_2$ and $J_1J_2J_3$ systems of up 50 spins. By virtue of the fact that, for all of the experiments performed in this work, hyperbolic GRU can yield performances comparable to or better than Euclidean RNNs, which have been extensively studied in these settings in the literature, our work is a proof-of-concept for the viability of hyperbolic GRU as the first type of non-Euclidean NQS ansatz for quantum many-body systems. Furthermore, in settings where the Hamiltonian displays a clear hierarchical interaction structure, such as the 1D Heisenberg $J_1J_2$ & $J_1J_2J_3$ systems with the 1st, 2nd and even 3rd nearest neighbor interactions, our results show that hyperbolic GRU definitively outperforms its Euclidean version in all instances. The fact that these results are reminiscent of the established ones from natural language processing where hyperbolic GRU almost always outperforms Euclidean RNNs when the training data exhibit a tree-like or hierarchical structure leads us to hypothesize that hyperbolic GRU NQS ansatz would likely outperform Euclidean RNN/GRU NQS ansatz in quantum spin systems that involve different degrees of nearest neighbor interactions. Finally, with this work, we hope to initiate future studies of other types of non-Euclidean NQS beyond hyperbolic GRU. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:06:25 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Dao",
"H. L.",
""
]
] |
2505.22084 | Julie Tores | Julie Tores, Elisa Ancarani, Lucile Sassatelli, Hui-Yin Wu, Clement Bergman, Lea Andolfi, Victor Ecrement, Remy Sun, Frederic Precioso, Thierry Devars, Magali Guaresi, Virginie Julliard, Sarah Lecossais | MObyGaze: a film dataset of multimodal objectification densely annotated by experts | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Characterizing and quantifying gender representation disparities in audiovisual storytelling contents is necessary to grasp how stereotypes may perpetuate on screen. In this article, we consider the high-level construct of objectification and introduce a new AI task to the ML community: characterize and quantify complex multimodal (visual, speech, audio) temporal patterns producing objectification in films. Building on film studies and psychology, we define the construct of objectification in a structured thesaurus involving 5 sub-constructs manifesting through 11 concepts spanning 3 modalities. We introduce the Multimodal Objectifying Gaze (MObyGaze) dataset, made of 20 movies annotated densely by experts for objectification levels and concepts over freely delimited segments: it amounts to 6072 segments over 43 hours of video with fine-grained localization and categorization. We formulate different learning tasks, propose and investigate best ways to learn from the diversity of labels among a low number of annotators, and benchmark recent vision, text and audio models, showing the feasibility of the task. We make our code and our dataset available to the community and described in the Croissant format: https://anonymous.4open.science/r/MObyGaze-F600/. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:07:28 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Tores",
"Julie",
""
],
[
"Ancarani",
"Elisa",
""
],
[
"Sassatelli",
"Lucile",
""
],
[
"Wu",
"Hui-Yin",
""
],
[
"Bergman",
"Clement",
""
],
[
"Andolfi",
"Lea",
""
],
[
"Ecrement",
"Victor",
""
],
[
"Sun",
"Remy",
""
],
[
"Precioso",
"Frederic",
""
],
[
"Devars",
"Thierry",
""
],
[
"Guaresi",
"Magali",
""
],
[
"Julliard",
"Virginie",
""
],
[
"Lecossais",
"Sarah",
""
]
] |
2505.22085 | Julian Kranz | Arnulf Jentzen and Julian Kranz and Adrian Riekert | PADAM: Parallel averaged Adam reduces the error for stochastic optimization in scientific machine learning | 38 pages, 13 figures | null | null | null | math.OC cs.LG cs.NA math.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Averaging techniques such as Ruppert--Polyak averaging and exponential movering averaging (EMA) are powerful approaches to accelerate optimization procedures of stochastic gradient descent (SGD) optimization methods such as the popular ADAM optimizer. However, depending on the specific optimization problem under consideration, the type and the parameters for the averaging need to be adjusted to achieve the smallest optimization error. In this work we propose an averaging approach, which we refer to as parallel averaged ADAM (PADAM), in which we compute parallely different averaged variants of ADAM and during the training process dynamically select the variant with the smallest optimization error. A central feature of this approach is that this procedure requires no more gradient evaluations than the usual ADAM optimizer as each of the averaged trajectories relies on the same underlying ADAM trajectory and thus on the same underlying gradients. We test the proposed PADAM optimizer in 13 stochastic optimization and deep neural network (DNN) learning problems and compare its performance with known optimizers from the literature such as standard SGD, momentum SGD, Adam with and without EMA, and ADAMW. In particular, we apply the compared optimizers to physics-informed neural network, deep Galerkin, deep backward stochastic differential equation and deep Kolmogorov approximations for boundary value partial differential equation problems from scientific machine learning, as well as to DNN approximations for optimal control and optimal stopping problems. In nearly all of the considered examples PADAM achieves, sometimes among others and sometimes exclusively, essentially the smallest optimization error. This work thus strongly suggest to consider PADAM for scientific machine learning problems and also motivates further research for adaptive averaging procedures within the training of DNNs. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:07:34 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Jentzen",
"Arnulf",
""
],
[
"Kranz",
"Julian",
""
],
[
"Riekert",
"Adrian",
""
]
] |
2505.22086 | Jia Xiong | Runkai Li, Jia Xiong, Xi Wang | iDSE: Navigating Design Space Exploration in High-Level Synthesis Using LLMs | null | null | null | null | cs.AR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | High-Level Synthesis (HLS) serves as an agile hardware development tool that streamlines the circuit design by abstracting the register transfer level into behavioral descriptions, while allowing designers to customize the generated microarchitectures through optimization directives. However, the combinatorial explosion of possible directive configurations yields an intractable design space. Traditional design space exploration (DSE) methods, despite adopting heuristics or constructing predictive models to accelerate Pareto-optimal design acquisition, still suffer from prohibitive exploration costs and suboptimal results. Addressing these concerns, we introduce iDSE, the first LLM-aided DSE framework that leverages HLS design quality perception to effectively navigate the design space. iDSE intelligently pruns the design space to guide LLMs in calibrating representative initial sampling designs, expediting convergence toward the Pareto front. By exploiting the convergent and divergent thinking patterns inherent in LLMs for hardware optimization, iDSE achieves multi-path refinement of the design quality and diversity. Extensive experiments demonstrate that iDSE outperforms heuristic-based DSE methods by 5.1$\times$$\sim$16.6$\times$ in proximity to the reference Pareto front, matching NSGA-II with only 4.6% of the explored designs. Our work demonstrates the transformative potential of LLMs in scalable and efficient HLS design optimization, offering new insights into multiobjective optimization challenges. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 08:08:57 GMT"
}
] | 2025-05-29T00:00:00 | [
[
"Li",
"Runkai",
""
],
[
"Xiong",
"Jia",
""
],
[
"Wang",
"Xi",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.