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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.17997 | Jintian Shao | Jintian Shao, Yiming Cheng, Hongyi Huang, Beiwen Zhang, Zhiyu Wu, You Shan, Mingkai Zheng | Towards Analyzing and Understanding the Limitations of VAPO: A Theoretical Perspective | We are withdrawing this submission as the underlying experiment is currently incomplete. We require additional time to gather more data and supplement the existing findings to ensure a comprehensive and robust presentation. We intend to resubmit once these additions are finalized | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | The VAPO framework has demonstrated significant empirical success in enhancing the efficiency and reliability of reinforcement learning for long chain-of-thought (CoT) reasoning tasks with large language models (LLMs). By systematically addressing challenges such as value model bias, heterogeneous sequence lengths, and sparse reward signals, VAPO achieves state-of-the-art performance. While its practical benefits are evident, a deeper theoretical understanding of its underlying mechanisms and potential limitations is crucial for guiding future advancements. This paper aims to initiate such a discussion by exploring VAPO from a theoretical perspective, highlighting areas where its assumptions might be challenged and where further investigation could yield more robust and generalizable reasoning agents. We delve into the intricacies of value function approximation in complex reasoning spaces, the optimality of adaptive advantage estimation, the impact of token-level optimization, and the enduring challenges of exploration and generalization. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 15:03:41 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 08:30:19 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Shao",
"Jintian",
""
],
[
"Cheng",
"Yiming",
""
],
[
"Huang",
"Hongyi",
""
],
[
"Zhang",
"Beiwen",
""
],
[
"Wu",
"Zhiyu",
""
],
[
"Shan",
"You",
""
],
[
"Zheng",
"Mingkai",
""
]
] |
2505.18009 | Simin Shen | Simin Shen, Zaiwu Gong, Bin Zhou, Roman S{\l}owi\'nski | Empathic network learning for multi-expert emergency decision-making under incomplete and inconsistent information | null | Shen S, Gong Z, Zhou B, et al. Empathic network learning for multi-expert emergency decision-making under incomplete and inconsistent information[J]. Information Fusion, 2025, 117: 102844 | 10.1016/j.inffus.2024.102844. | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Challenges, such as a lack of information for emergency decision-making, time pressure, and limited knowledge of experts acting as decision-makers (DMs), can result in the generation of poor or inconsistent indirect information regarding DMs' preferences. Simultaneously, the empathic relationship represents a tangible social connection within the context of actual emergency decision-making, with the structure of the empathic network being a significant factor influencing the outcomes of the decision-making process. To deduce the empathic network underpinning the decision behaviors of DMs from incomplete and inconsistent preference information, we introduce an empathic network learning methodology rooted in the concept of robust ordinal regression via preference disaggregation. Firstly, we complete incomplete fuzzy judgment matrices including holistic preference information given in terms of decision examples on some reference alternatives, independently by each DM, and we calculate the intrinsic utilities of DMs. Secondly, we establish constraints for empathic network learning models based on empathic preference information and information about relations between some reference nodes. Then, the necessary and possible empathic relationships between any two DMs are calculated. Lastly, tailored to the specific requirements of different emergency scenarios, we design six target networks and construct models to derive the most representative empathic network. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 15:14:17 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 01:40:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Shen",
"Simin",
""
],
[
"Gong",
"Zaiwu",
""
],
[
"Zhou",
"Bin",
""
],
[
"Słowiński",
"Roman",
""
]
] |
2505.18022 | Wei Xu | Liang Yao, Fan Liu, Delong Chen, Chuanyi Zhang, Yijun Wang, Ziyun Chen, Wei Xu, Shimin Di, Yuhui Zheng | RemoteSAM: Towards Segment Anything for Earth Observation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We aim to develop a robust yet flexible visual foundation model for Earth observation. It should possess strong capabilities in recognizing and localizing diverse visual targets while providing compatibility with various input-output interfaces required across different task scenarios. Current systems cannot meet these requirements, as they typically utilize task-specific architecture trained on narrow data domains with limited semantic coverage. Our study addresses these limitations from two aspects: data and modeling. We first introduce an automatic data engine that enjoys significantly better scalability compared to previous human annotation or rule-based approaches. It has enabled us to create the largest dataset of its kind to date, comprising 270K image-text-mask triplets covering an unprecedented range of diverse semantic categories and attribute specifications. Based on this data foundation, we further propose a task unification paradigm that centers around referring expression segmentation. It effectively handles a wide range of vision-centric perception tasks, including classification, detection, segmentation, grounding, etc, using a single model without any task-specific heads. Combining these innovations on data and modeling, we present RemoteSAM, a foundation model that establishes new SoTA on several earth observation perception benchmarks, outperforming other foundation models such as Falcon, GeoChat, and LHRS-Bot with significantly higher efficiency. Models and data are publicly available at https://github.com/1e12Leon/RemoteSAM. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 15:27:57 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 09:43:51 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yao",
"Liang",
""
],
[
"Liu",
"Fan",
""
],
[
"Chen",
"Delong",
""
],
[
"Zhang",
"Chuanyi",
""
],
[
"Wang",
"Yijun",
""
],
[
"Chen",
"Ziyun",
""
],
[
"Xu",
"Wei",
""
],
[
"Di",
"Shimin",
""
],
[
"Zheng",
"Yuhui",
""
]
] |
2505.18034 | Wentao Sun | Wentao Sun, Jo\~ao Paulo Nogueira, and Alonso Silva | Structured Thinking Matters: Improving LLMs Generalization in Causal Inference Tasks | null | null | null | null | cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite remarkable advances in the field, LLMs remain unreliable in distinguishing causation from correlation. Recent results from the Corr2Cause dataset benchmark reveal that state-of-the-art LLMs -- such as GPT-4 (F1 score: 29.08) -- only marginally outperform random baselines (Random Uniform, F1 score: 20.38), indicating limited capacity of generalization. To tackle this limitation, we propose a novel structured approach: rather than directly answering causal queries, we provide the model with the capability to structure its thinking by guiding the model to build a structured knowledge graph, systematically encoding the provided correlational premises, to answer the causal queries. This intermediate representation significantly enhances the model's causal capabilities. Experiments on the test subset of the Corr2Cause dataset benchmark with Qwen3-32B model (reasoning model) show substantial gains over standard direct prompting methods, improving F1 scores from 32.71 to 48.26 (over 47.5% relative increase), along with notable improvements in precision and recall. These results underscore the effectiveness of providing the model with the capability to structure its thinking and highlight its promising potential for broader generalization across diverse causal inference tasks. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 15:37:40 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 08:16:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Sun",
"Wentao",
""
],
[
"Nogueira",
"João Paulo",
""
],
[
"Silva",
"Alonso",
""
]
] |
2505.18048 | Simon Malzard | Simon Malzard, Nitish Mital, Richard Walters, Victoria Nockles, Raghuveer Rao and Celso M. De Melo | SHARDeg: A Benchmark for Skeletal Human Action Recognition in Degraded Scenarios | 19 pages, 2 images, updated acknowledgements versus previous versions to be compliant with funders | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Computer vision (CV) models for detection, prediction or classification tasks operate on video data-streams that are often degraded in the real world, due to deployment in real-time or on resource-constrained hardware. It is therefore critical that these models are robust to degraded data, but state of the art (SoTA) models are often insufficiently assessed with these real-world constraints in mind. This is exemplified by Skeletal Human Action Recognition (SHAR), which is critical in many CV pipelines operating in real-time and at the edge, but robustness to degraded data has previously only been shallowly and inconsistently assessed. Here we address this issue for SHAR by providing an important first data degradation benchmark on the most detailed and largest 3D open dataset, NTU-RGB+D-120, and assess the robustness of five leading SHAR models to three forms of degradation that represent real-world issues. We demonstrate the need for this benchmark by showing that the form of degradation, which has not previously been considered, has a large impact on model accuracy; at the same effective frame rate, model accuracy can vary by >40% depending on degradation type. We also identify that temporal regularity of frames in degraded SHAR data is likely a major driver of differences in model performance, and harness this to improve performance of existing models by up to >40%, through employing a simple mitigation approach based on interpolation. Finally, we highlight how our benchmark has helped identify an important degradation-resistant SHAR model based in Rough Path Theory; the LogSigRNN SHAR model outperforms the SoTA DeGCN model in five out of six cases at low frame rates by an average accuracy of 6%, despite trailing the SoTA model by 11-12% on un-degraded data at high frame rates (30 FPS). | [
{
"version": "v1",
"created": "Fri, 23 May 2025 15:52:31 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 15:11:07 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Malzard",
"Simon",
""
],
[
"Mital",
"Nitish",
""
],
[
"Walters",
"Richard",
""
],
[
"Nockles",
"Victoria",
""
],
[
"Rao",
"Raghuveer",
""
],
[
"De Melo",
"Celso M.",
""
]
] |
2505.18060 | Kaiyan Zhang | Kaiyan Zhang, Xinghui Li, Jingyi Lu, Kai Han | Semantic Correspondence: Unified Benchmarking and a Strong Baseline | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Establishing semantic correspondence is a challenging task in computer vision, aiming to match keypoints with the same semantic information across different images. Benefiting from the rapid development of deep learning, remarkable progress has been made over the past decade. However, a comprehensive review and analysis of this task remains absent. In this paper, we present the first extensive survey of semantic correspondence methods. We first propose a taxonomy to classify existing methods based on the type of their method designs. These methods are then categorized accordingly, and we provide a detailed analysis of each approach. Furthermore, we aggregate and summarize the results of methods in literature across various benchmarks into a unified comparative table, with detailed configurations to highlight performance variations. Additionally, to provide a detailed understanding on existing methods for semantic matching, we thoroughly conduct controlled experiments to analyse the effectiveness of the components of different methods. Finally, we propose a simple yet effective baseline that achieves state-of-the-art performance on multiple benchmarks, providing a solid foundation for future research in this field. We hope this survey serves as a comprehensive reference and consolidated baseline for future development. Code is publicly available at: https://github.com/Visual-AI/Semantic-Correspondence. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 16:07:16 GMT"
},
{
"version": "v2",
"created": "Mon, 26 May 2025 12:40:13 GMT"
},
{
"version": "v3",
"created": "Tue, 27 May 2025 09:45:05 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Kaiyan",
""
],
[
"Li",
"Xinghui",
""
],
[
"Lu",
"Jingyi",
""
],
[
"Han",
"Kai",
""
]
] |
2505.18092 | Weizhou Shen | Weizhou Shen, Chenliang Li, Fanqi Wan, Shengyi Liao, Shaopeng Lai, Bo Zhang, Yingcheng Shi, Yuning Wu, Gang Fu, Zhansheng Li, Bin Yang, Ji Zhang, Fei Huang, Jingren Zhou, Ming Yan | QwenLong-CPRS: Towards $\infty$-LLMs with Dynamic Context Optimization | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This technical report presents QwenLong-CPRS, a context compression framework designed for explicit long-context optimization, addressing prohibitive computation overhead during the prefill stage and the "lost in the middle" performance degradation of large language models (LLMs) during long sequence processing. Implemented through a novel dynamic context optimization mechanism, QwenLong-CPRS enables multi-granularity context compression guided by natural language instructions, achieving both efficiency gains and improved performance.
Evolved from the Qwen architecture series, QwenLong-CPRS introduces four key innovations: (1) Natural language-guided dynamic optimization, (2) Bidirectional reasoning layers for enhanced boundary awareness, (3) Token critic mechanisms with language modeling heads, and (4) Window-parallel inference.
Comprehensive evaluations across five benchmarks (4K-2M word contexts) demonstrate QwenLong-CPRS's threefold effectiveness: (1) Consistent superiority over other context management methods like RAG and sparse attention in both accuracy and efficiency. (2) Architecture-agnostic integration with all flagship LLMs, including GPT-4o, Gemini2.0-pro, Claude3.7-sonnet, DeepSeek-v3, and Qwen2.5-max, achieves 21.59$\times$ context compression alongside 19.15-point average performance gains; (3) Deployed with Qwen2.5-32B-Instruct, QwenLong-CPRS surpasses leading proprietary LLMs by 4.85 and 10.88 points on Ruler-128K and InfiniteBench, establishing new SOTA performance. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 16:47:00 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 09:42:25 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Shen",
"Weizhou",
""
],
[
"Li",
"Chenliang",
""
],
[
"Wan",
"Fanqi",
""
],
[
"Liao",
"Shengyi",
""
],
[
"Lai",
"Shaopeng",
""
],
[
"Zhang",
"Bo",
""
],
[
"Shi",
"Yingcheng",
""
],
[
"Wu",
"Yuning",
""
],
[
"Fu",
"Gang",
""
],
[
"Li",
"Zhansheng",
""
],
[
"Yang",
"Bin",
""
],
[
"Zhang",
"Ji",
""
],
[
"Huang",
"Fei",
""
],
[
"Zhou",
"Jingren",
""
],
[
"Yan",
"Ming",
""
]
] |
2505.18145 | Uwe C. T\"auber | Mohamed Swailem (Stony Brook), Ulrich Dobramysl (Oxford), Ruslan Mukhamadiarov (LMU Munich), and Uwe C. T\"auber (Virginia Tech) | Agent-based Monte Carlo simulations for reaction-diffusion models, population dynamics, and epidemic spreading | 60 pages, 17 figures, final version, editorially edited and condensed, to appear in American Journal of Physics (2025) | null | null | null | cond-mat.stat-mech q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We provide an overview of Monte Carlo algorithms based on Markovian stochastic dynamics of interacting and reacting many-particle systems not in thermal equilibrium. These agent-based simulations are an effective way of introducing students to current research without requiring much prior knowledge or experience. By starting from the direct visualization of the data, students can gain immediate insight into emerging macroscopic features of a complex system and subsequently apply more sophisticated data analysis to quantitatively characterize its rich dynamical properties, both in the stationary and transient regimes. We utilize simulations of reaction-diffusion systems, stochastic models for population dynamics and epidemic spreading, to exemplify how interdisciplinary computational research can be effectively utilized in bottom-up undergraduate and graduate education through learning by doing. We also give helpful hints for the practical implementation of Monte Carlo algorithms, provide sample codes, explain some typical data analysis tools, and describe various potential error sources and pitfalls and tips for avoiding them. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 17:55:52 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 17:39:33 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Swailem",
"Mohamed",
"",
"Stony Brook"
],
[
"Dobramysl",
"Ulrich",
"",
"Oxford"
],
[
"Mukhamadiarov",
"Ruslan",
"",
"LMU Munich"
],
[
"Täuber",
"Uwe C.",
"",
"Virginia Tech"
]
] |
2505.18188 | Beck LaBash | Beck LaBash, Shahriar Khushrushahi, Fabian Ruehle | Improving Generative Inverse Design of Rectangular Patch Antennas with Test Time Optimization | Code and dataset available at https://github.com/becklabs/patch-antenna-tto | null | null | null | eess.SP cs.AI cs.LG cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | We propose a two-stage deep learning framework for the inverse design of rectangular patch antennas. Our approach leverages generative modeling to learn a latent representation of antenna frequency response curves and conditions a subsequent generative model on these responses to produce feasible antenna geometries. We further demonstrate that leveraging search and optimization techniques at test-time improves the accuracy of the generated designs and enables consideration of auxiliary objectives such as manufacturability. Our approach generalizes naturally to different design criteria, and can be easily adapted to more complex geometric design spaces. | [
{
"version": "v1",
"created": "Mon, 19 May 2025 02:24:28 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 00:40:18 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"LaBash",
"Beck",
""
],
[
"Khushrushahi",
"Shahriar",
""
],
[
"Ruehle",
"Fabian",
""
]
] |
2505.18190 | Yuezhou Ma | Yuezhou Ma, Haixu Wu, Hang Zhou, Huikun Weng, Jianmin Wang, Mingsheng Long | PhySense: Sensor Placement Optimization for Accurate Physics Sensing | null | null | null | null | eess.SP cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Physics sensing plays a central role in many scientific and engineering domains, which inherently involves two coupled tasks: reconstructing dense physical fields from sparse observations and optimizing scattered sensor placements to observe maximum information. While deep learning has made rapid advances in sparse-data reconstruction, existing methods generally omit optimization of sensor placements, leaving the mutual enhancement between reconstruction and placement on the shelf. To change this suboptimal practice, we propose PhySense, a synergistic two-stage framework that learns to jointly reconstruct physical fields and to optimize sensor placements, both aiming for accurate physics sensing. The first stage involves a flow-based generative model enhanced by cross-attention to adaptively fuse sparse observations. Leveraging the reconstruction feedback, the second stage performs sensor placement via projected gradient descent to satisfy spatial constraints. We further prove that the learning objectives of the two stages are consistent with classical variance-minimization principles, providing theoretical guarantees. Extensive experiments across three challenging benchmarks, especially a 3D geometry dataset, indicate PhySense achieves state-of-the-art physics sensing accuracy and discovers informative sensor placements previously unconsidered. | [
{
"version": "v1",
"created": "Mon, 19 May 2025 14:59:11 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 01:32:48 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ma",
"Yuezhou",
""
],
[
"Wu",
"Haixu",
""
],
[
"Zhou",
"Hang",
""
],
[
"Weng",
"Huikun",
""
],
[
"Wang",
"Jianmin",
""
],
[
"Long",
"Mingsheng",
""
]
] |
2505.18275 | Mario Nicola Mazziotta | M. N. Mazziotta | A 3D Monte Carlo calculation of the inverse Compton emission from the Sun and stars in presence of magnetic and electric fields | 14 pages, 10 figures, accepted for publication by PRD (submitted on Apr 14) | null | 10.1103/zs82-fktf | null | hep-ph astro-ph.HE astro-ph.SR hep-ex | http://creativecommons.org/licenses/by/4.0/ | The solar steady emission in gamma rays is due to the interactions of Galactic cosmic rays with the solar atmosphere and with the low-energy solar photon field via inverse Compton scattering. The emission is sensitive to the magnetic field nearby the Sun and to the cosmic-ray transport in the magnetic field in the inner solar system. Modeling the inverse Compton emission in the presence of a magnetic field is therefore crucial to better interpret the observations. In a previous work we have presented a comprehensive calculation of the secondary productions due to the collision of cosmic rays with the solar atmosphere in presence of magnetic fields. In this paper, we present a general approach to calculate the (anisotropic) inverse Compton scattering in a 3D Monte Carlo simulation, also in presence of magnetic and electric fields. After a short review of the scattering process of photons with electrons, examples of inverse Compton emission are presented, including the predictions for the Sun. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 18:10:52 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Mazziotta",
"M. N.",
""
]
] |
2505.18276 | Lorenzo Baldassari | Lorenzo Baldassari, Josselin Garnier, Knut Solna, Maarten V. de Hoop | Preconditioned Langevin Dynamics with Score-Based Generative Models for Infinite-Dimensional Linear Bayesian Inverse Problems | null | null | null | null | stat.ML cs.LG cs.NA math.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Designing algorithms for solving high-dimensional Bayesian inverse problems directly in infinite-dimensional function spaces - where such problems are naturally formulated - is crucial to ensure stability and convergence as the discretization of the underlying problem is refined. In this paper, we contribute to this line of work by analyzing a widely used sampler for linear inverse problems: Langevin dynamics driven by score-based generative models (SGMs) acting as priors, formulated directly in function space. Building on the theoretical framework for SGMs in Hilbert spaces, we give a rigorous definition of this sampler in the infinite-dimensional setting and derive, for the first time, error estimates that explicitly depend on the approximation error of the score. As a consequence, we obtain sufficient conditions for global convergence in Kullback-Leibler divergence on the underlying function space. Preventing numerical instabilities requires preconditioning of the Langevin algorithm and we prove the existence and the form of an optimal preconditioner. The preconditioner depends on both the score error and the forward operator and guarantees a uniform convergence rate across all posterior modes. Our analysis applies to both Gaussian and a general class of non-Gaussian priors. Finally, we present examples that illustrate and validate our theoretical findings. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 18:12:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Baldassari",
"Lorenzo",
""
],
[
"Garnier",
"Josselin",
""
],
[
"Solna",
"Knut",
""
],
[
"de Hoop",
"Maarten V.",
""
]
] |
2505.18291 | Zifu Wan | Zifu Wan, Yaqi Xie, Ce Zhang, Zhiqiu Lin, Zihan Wang, Simon Stepputtis, Deva Ramanan, Katia Sycara | InstructPart: Task-Oriented Part Segmentation with Instruction Reasoning | Accepted by ACL 2025 Main. Project page: https://zifuwan.github.io/InstructPart/ | null | null | null | cs.CV cs.CL cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large multimodal foundation models, particularly in the domains of language and vision, have significantly advanced various tasks, including robotics, autonomous driving, information retrieval, and grounding. However, many of these models perceive objects as indivisible, overlooking the components that constitute them. Understanding these components and their associated affordances provides valuable insights into an object's functionality, which is fundamental for performing a wide range of tasks. In this work, we introduce a novel real-world benchmark, InstructPart, comprising hand-labeled part segmentation annotations and task-oriented instructions to evaluate the performance of current models in understanding and executing part-level tasks within everyday contexts. Through our experiments, we demonstrate that task-oriented part segmentation remains a challenging problem, even for state-of-the-art Vision-Language Models (VLMs). In addition to our benchmark, we introduce a simple baseline that achieves a twofold performance improvement through fine-tuning with our dataset. With our dataset and benchmark, we aim to facilitate research on task-oriented part segmentation and enhance the applicability of VLMs across various domains, including robotics, virtual reality, information retrieval, and other related fields. Project website: https://zifuwan.github.io/InstructPart/. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 18:36:13 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wan",
"Zifu",
""
],
[
"Xie",
"Yaqi",
""
],
[
"Zhang",
"Ce",
""
],
[
"Lin",
"Zhiqiu",
""
],
[
"Wang",
"Zihan",
""
],
[
"Stepputtis",
"Simon",
""
],
[
"Ramanan",
"Deva",
""
],
[
"Sycara",
"Katia",
""
]
] |
2505.18352 | Karen Fonseca | Karen Fonseca, Leon Suarez-Rodriguez, Andres Jerez, Felipe Gutierrez-Barragan, Henry Arguello | Single Snapshot Distillation for Phase Coded Mask Design in Phase Retrieval | Accepted on the IEEE International Conference on Image Processing, IEEE ICIP 2025 | null | null | null | eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Phase retrieval (PR) reconstructs phase information from magnitude measurements, known as coded diffraction patterns (CDPs), whose quality depends on the number of snapshots captured using coded phase masks. High-quality phase estimation requires multiple snapshots, which is not desired for efficient PR systems. End-to-end frameworks enable joint optimization of the optical system and the recovery neural network. However, their application is constrained by physical implementation limitations. Additionally, the framework is prone to gradient vanishing issues related to its global optimization process. This paper introduces a Knowledge Distillation (KD) optimization approach to address these limitations. KD transfers knowledge from a larger, lower-constrained network (teacher) to a smaller, more efficient, and implementable network (student). In this method, the teacher, a PR system trained with multiple snapshots, distills its knowledge into a single-snapshot PR system, the student. The loss functions compare the CPMs and the feature space of the recovery network. Simulations demonstrate that this approach improves reconstruction performance compared to a PR system trained without the teacher's guidance. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 20:18:52 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 15:50:56 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Fonseca",
"Karen",
""
],
[
"Suarez-Rodriguez",
"Leon",
""
],
[
"Jerez",
"Andres",
""
],
[
"Gutierrez-Barragan",
"Felipe",
""
],
[
"Arguello",
"Henry",
""
]
] |
2505.18361 | Trinity Chung | Trinity Chung, Yuchen Shen, Nathan C. L. Kong, Aran Nayebi | Task-Optimized Convolutional Recurrent Networks Align with Tactile Processing in the Rodent Brain | 9 pages, 8 figures, 5 tables | null | null | null | q-bio.NC cs.AI cs.LG cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tactile sensing remains far less understood in neuroscience and less effective in artificial systems compared to more mature modalities such as vision and language. We bridge these gaps by introducing a novel Encoder-Attender-Decoder (EAD) framework to systematically explore the space of task-optimized temporal neural networks trained on realistic tactile input sequences from a customized rodent whisker-array simulator. We identify convolutional recurrent neural networks (ConvRNNs) as superior encoders to purely feedforward and state-space architectures for tactile categorization. Crucially, these ConvRNN-encoder-based EAD models achieve neural representations closely matching rodent somatosensory cortex, saturating the explainable neural variability and revealing a clear linear relationship between supervised categorization performance and neural alignment. Furthermore, contrastive self-supervised ConvRNN-encoder-based EADs, trained with tactile-specific augmentations, match supervised neural fits, serving as an ethologically-relevant, label-free proxy.
For neuroscience, our findings highlight nonlinear recurrent processing as important for general-purpose tactile representations in somatosensory cortex, providing the first quantitative characterization of the underlying inductive biases in this system. For embodied AI, our results emphasize the importance of recurrent EAD architectures to handle realistic tactile inputs, along with tailored self-supervised learning methods for achieving robust tactile perception with the same type of sensors animals use to sense in unstructured environments. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 20:40:28 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:47:07 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Chung",
"Trinity",
""
],
[
"Shen",
"Yuchen",
""
],
[
"Kong",
"Nathan C. L.",
""
],
[
"Nayebi",
"Aran",
""
]
] |
2505.18374 | Jarrod Ragsdale | Jarrod Ragsdale, Rajendra Boppana | ShIOEnv: A CLI Behavior-Capturing Environment Enabling Grammar-Guided Command Synthesis for Dataset Curation | 18 pages, 11 figures, conference preprint | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/publicdomain/zero/1.0/ | Command-line interfaces (CLIs) provide structured textual environments for system administration. Explorations have been performed using pre-trained language models (PLMs) to simulate these environments for safe interaction in high-risk environments. However, their use has been constrained to frozen, large parameter models like GPT. For smaller architectures to reach a similar level of believability, a rich dataset of CLI interactions is required. Existing public datasets focus on mapping natural-language tasks to commands, omitting crucial execution data such as exit codes, outputs, and environmental side effects, limiting their usability for behavioral modeling. We introduce a Shell Input -Output Environment (ShIOEnv), which casts command construction as a Markov Decision Process whose state is the partially built sequence and whose actions append arguments. After each action, ShIOEnv executes the candidate and returns its exit status, output, and progress toward a minimal-length behavioral objective. Due to the intractable nature of the combinatorial argument state-action space, we derive a context-free grammar from man pages to mask invalid arguments from being emitted. We explore random and proximal-policy optimization (PPO)-optimized sampling of unrestricted and grammar-masked action spaces to produce four exploration strategies. We observed that grammar masking and PPO significantly improve sample efficiency to produce a higher quality dataset (maximizing the number of arguments while minimizing redundancies). Policy-generated datasets of shell input-output behavior pairs are used to fine-tune CodeT5, where we observe 85% improvements in BLEU-4 when constraining the action space to grammar productions with an additional 26% improvement when applying PPO. The ShIOEnv environment and curated command behavior datasets are released for use in future research. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 21:00:57 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ragsdale",
"Jarrod",
""
],
[
"Boppana",
"Rajendra",
""
]
] |
2505.18389 | Maxime Elkael | Maxime Elkael, Michele Polese, Reshma Prasad, Stefano Maxenti, Tommaso Melodia | ALLSTaR: Automated LLM-Driven Scheduler Generation and Testing for Intent-Based RAN | Under submission to an IEEE journal, copyright may change without notice | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The evolution toward open, programmable O-RAN and AI-RAN 6G networks creates unprecedented opportunities for Intent-Based Networking (IBN) to dynamically optimize RAN[...]. However, applying IBN effectively to the RAN scheduler [...] remains a significant challenge. Current approaches predominantly rely on coarse-grained network slicing, lacking the granularity for dynamic adaptation to individual user conditions and traffic patterns. Despite the existence of a vast body of scheduling algorithms [...], their practical utilization is hindered by implementation heterogeneity, insufficient systematic evaluation in production environments, and the complexity of developing high-performance scheduler implementations.[...] To address these limitations, we propose ALLSTaR (Automated LLm-driven Scheduler generation and Testing for intent-based RAN), a novel framework leveraging LLMs for automated, intent-driven scheduler design, implementation, and evaluation. ALLSTaR interprets NL intents, automatically generates functional scheduler code from the research literature using OCR and LLMs, and intelligently matches operator intents to the most suitable scheduler(s). Our implementation deploys these schedulers as O-RAN dApps, enabling on-the-fly deployment and testing on a production-grade, 5G-compliant testbed. This approach has enabled the largest-scale OTA experimental comparison of 18 scheduling algorithms automatically synthesized from the academic literature. The resulting performance profiles serve as the input for our Intent-Based Scheduling (IBS) framework, which dynamically selects and deploys appropriate schedulers that optimally satisfy operator intents. We validate our approach through multiple use cases unattainable with current slicing-based optimization techniques, demonstrating fine-grained control based on buffer status, physical layer conditions, and heterogeneous traffic types | [
{
"version": "v1",
"created": "Fri, 23 May 2025 21:33:16 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:13:53 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Elkael",
"Maxime",
""
],
[
"Polese",
"Michele",
""
],
[
"Prasad",
"Reshma",
""
],
[
"Maxenti",
"Stefano",
""
],
[
"Melodia",
"Tommaso",
""
]
] |
2505.18405 | Debrup Das | Debrup Das, Sam O' Nuallain, Razieh Rahimi | RaDeR: Reasoning-aware Dense Retrieval Models | 26 pages | null | null | null | cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | We propose RaDeR, a set of reasoning-based dense retrieval models trained with data derived from mathematical problem solving using large language models (LLMs). Our method leverages retrieval-augmented reasoning trajectories of an LLM and self-reflective relevance evaluation, enabling the creation of both diverse and hard-negative samples for reasoning-intensive relevance. RaDeR retrievers, trained for mathematical reasoning, effectively generalize to diverse reasoning tasks in the BRIGHT and RAR-b benchmarks, consistently outperforming strong baselines in overall performance. Notably, RaDeR achieves significantly higher performance than baselines on the Math and Coding splits. In addition, RaDeR presents the first dense retriever that outperforms BM25 when queries are Chain-of-Thought reasoning steps, underscoring the critical role of reasoning-based retrieval to augment reasoning language models. Furthermore, RaDeR achieves comparable or superior performance while using only 2.5% of the training data used by the concurrent work REASONIR, highlighting the quality of our synthesized training data. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 22:18:32 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 04:20:31 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Das",
"Debrup",
""
],
[
"Nuallain",
"Sam O'",
""
],
[
"Rahimi",
"Razieh",
""
]
] |
2505.18450 | Ainulla Khan | Ainulla Khan, Yamada Moyuru, Srinidhi Akella | BRIT: Bidirectional Retrieval over Unified Image-Text Graph | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Retrieval-Augmented Generation (RAG) has emerged as a promising technique to enhance the quality and relevance of responses generated by large language models. While recent advancements have mainly focused on improving RAG for text-based queries, RAG on multi-modal documents containing both texts and images has not been fully explored. Especially when fine-tuning does not work. This paper proposes BRIT, a novel multi-modal RAG framework that effectively unifies various text-image connections in the document into a multi-modal graph and retrieves the texts and images as a query-specific sub-graph. By traversing both image-to-text and text-to-image paths in the graph, BRIT retrieve not only directly query-relevant images and texts but also further relevant contents to answering complex cross-modal multi-hop questions. To evaluate the effectiveness of BRIT, we introduce MM-RAG test set specifically designed for multi-modal question answering tasks that require to understand the text-image relations. Our comprehensive experiments demonstrate the superiority of BRIT, highlighting its ability to handle cross-modal questions on the multi-modal documents. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 01:20:51 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Khan",
"Ainulla",
""
],
[
"Moyuru",
"Yamada",
""
],
[
"Akella",
"Srinidhi",
""
]
] |
2505.18458 | Wei Zhou | Xuanhe Zhou, Junxuan He, Wei Zhou, Haodong Chen, Zirui Tang, Haoyu Zhao, Xin Tong, Guoliang Li, Youmin Chen, Jun Zhou, Zhaojun Sun, Binyuan Hui, Shuo Wang, Conghui He, Zhiyuan Liu, Jingren Zhou, Fan Wu | A Survey of LLM $\times$ DATA | Please refer to the paper list at: https://github.com/weAIDB/awesome-data-llm | null | null | null | cs.DB cs.AI cs.CL cs.IR cs.LG | http://creativecommons.org/licenses/by/4.0/ | The integration of large language model (LLM) and data management (DATA) is rapidly redefining both domains. In this survey, we comprehensively review the bidirectional relationships. On the one hand, DATA4LLM, spanning large-scale data processing, storage, and serving, feeds LLMs with high quality, diversity, and timeliness of data required for stages like pre-training, post-training, retrieval-augmented generation, and agentic workflows: (i) Data processing for LLMs includes scalable acquisition, deduplication, filtering, selection, domain mixing, and synthetic augmentation; (ii) Data Storage for LLMs focuses on efficient data and model formats, distributed and heterogeneous storage hierarchies, KV-cache management, and fault-tolerant checkpointing; (iii) Data serving for LLMs tackles challenges in RAG (e.g., knowledge post-processing), LLM inference (e.g., prompt compression, data provenance), and training strategies (e.g., data packing and shuffling). On the other hand, in LLM4DATA, LLMs are emerging as general-purpose engines for data management. We review recent advances in (i) data manipulation, including automatic data cleaning, integration, discovery; (ii) data analysis, covering reasoning over structured, semi-structured, and unstructured data, and (iii) system optimization (e.g., configuration tuning, query rewriting, anomaly diagnosis), powered by LLM techniques like retrieval-augmented prompting, task-specialized fine-tuning, and multi-agent collaboration. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 01:57:12 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:57:47 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhou",
"Xuanhe",
""
],
[
"He",
"Junxuan",
""
],
[
"Zhou",
"Wei",
""
],
[
"Chen",
"Haodong",
""
],
[
"Tang",
"Zirui",
""
],
[
"Zhao",
"Haoyu",
""
],
[
"Tong",
"Xin",
""
],
[
"Li",
"Guoliang",
""
],
[
"Chen",
"Youmin",
""
],
[
"Zhou",
"Jun",
""
],
[
"Sun",
"Zhaojun",
""
],
[
"Hui",
"Binyuan",
""
],
[
"Wang",
"Shuo",
""
],
[
"He",
"Conghui",
""
],
[
"Liu",
"Zhiyuan",
""
],
[
"Zhou",
"Jingren",
""
],
[
"Wu",
"Fan",
""
]
] |
2505.18473 | Sebastian Gutierrez Hernandez | Sebastian Gutierrez Hernandez, Peng Chen, Haomin Zhou | PDPO: Parametric Density Path Optimization | Under review. 24 pages, 15 figures | null | null | null | math.OC | http://creativecommons.org/licenses/by/4.0/ | We introduce Parametric Density Path Optimization (PDPO), a novel method for computing action-minimizing paths between probability densities. The core idea is to represent the target probability path as the pushforward of a reference density through a parametric map, transforming the original infinite-dimensional optimization over densities to a finite-dimensional one over the parameters of the map. We derive a static formulation of the dynamic problem of action minimization and propose cubic spline interpolation of the path in parameter space to solve the static problem. Theoretically, we establish an error bound of the action under proper assumptions on the regularity of the parameter path. Empirically, we find that using 3-5 control points of the spline interpolation suffices to accurately resolve both multimodal and high-dimensional problems. We demonstrate that PDPO can flexibly accommodate a wide range of potential terms, including those modeling obstacles, mean-field interactions, stochastic control, and higher-order dynamics. Our method outperforms existing state-of-the-art approaches in benchmark tasks, demonstrating superior computational efficiency and solution quality. The source code will be publically available after the revision process. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 02:30:47 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:01:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hernandez",
"Sebastian Gutierrez",
""
],
[
"Chen",
"Peng",
""
],
[
"Zhou",
"Haomin",
""
]
] |
2505.18585 | Yedi Zhang | Yedi Zhang, Sun Yi Emma, Annabelle Lee Jia En, Jin Song Dong | RvLLM: LLM Runtime Verification with Domain Knowledge | 12 pages, 2 figures | null | null | null | cs.AI cs.CL cs.LO | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have emerged as a dominant AI paradigm due to their exceptional text understanding and generation capabilities. However, their tendency to generate inconsistent or erroneous outputs challenges their reliability, especially in high-stakes domains requiring accuracy and trustworthiness. Existing research primarily focuses on detecting and mitigating model misbehavior in general-purpose scenarios, often overlooking the potential of integrating domain-specific knowledge. In this work, we advance misbehavior detection by incorporating domain knowledge. The core idea is to design a general specification language that enables domain experts to customize domain-specific predicates in a lightweight and intuitive manner, supporting later runtime verification of LLM outputs. To achieve this, we design a novel specification language, ESL, and introduce a runtime verification framework, RvLLM, to validate LLM output against domain-specific constraints defined in ESL. We evaluate RvLLM on three representative tasks: violation detection against Singapore Rapid Transit Systems Act, numerical comparison, and inequality solving. Experimental results demonstrate that RvLLM effectively detects erroneous outputs across various LLMs in a lightweight and flexible manner. The results reveal that despite their impressive capabilities, LLMs remain prone to low-level errors due to limited interpretability and a lack of formal guarantees during inference, and our framework offers a potential long-term solution by leveraging expert domain knowledge to rigorously and efficiently verify LLM outputs. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 08:21:44 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 10:34:23 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Yedi",
""
],
[
"Emma",
"Sun Yi",
""
],
[
"En",
"Annabelle Lee Jia",
""
],
[
"Dong",
"Jin Song",
""
]
] |
2505.18596 | Chen Han | Chen Han, Wenzhen Zheng, Xijin Tang | Debate-to-Detect: Reformulating Misinformation Detection as a Real-World Debate with Large Language Models | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The proliferation of misinformation in digital platforms reveals the limitations of traditional detection methods, which mostly rely on static classification and fail to capture the intricate process of real-world fact-checking. Despite advancements in Large Language Models (LLMs) that enhance automated reasoning, their application to misinformation detection remains hindered by issues of logical inconsistency and superficial verification. In response, we introduce Debate-to-Detect (D2D), a novel Multi-Agent Debate (MAD) framework that reformulates misinformation detection as a structured adversarial debate. Inspired by fact-checking workflows, D2D assigns domain-specific profiles to each agent and orchestrates a five-stage debate process, including Opening Statement, Rebuttal, Free Debate, Closing Statement, and Judgment. To transcend traditional binary classification, D2D introduces a multi-dimensional evaluation mechanism that assesses each claim across five distinct dimensions: Factuality, Source Reliability, Reasoning Quality, Clarity, and Ethics. Experiments with GPT-4o on two fakenews datasets demonstrate significant improvements over baseline methods, and the case study highlight D2D's capability to iteratively refine evidence while improving decision transparency, representing a substantial advancement towards robust and interpretable misinformation detection. The code will be open-sourced in a future release. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 08:44:33 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 11:22:44 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Han",
"Chen",
""
],
[
"Zheng",
"Wenzhen",
""
],
[
"Tang",
"Xijin",
""
]
] |
2505.18600 | Jong Chul Ye | Bryan Sangwoo Kim, Jeongsol Kim, Jong Chul Ye | Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment | Project Page: https://bryanswkim.github.io/chain-of-zoom/ | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Modern single-image super-resolution (SISR) models deliver photo-realistic results at the scale factors on which they are trained, but collapse when asked to magnify far beyond that regime. We address this scalability bottleneck with Chain-of-Zoom (CoZ), a model-agnostic framework that factorizes SISR into an autoregressive chain of intermediate scale-states with multi-scale-aware prompts. CoZ repeatedly re-uses a backbone SR model, decomposing the conditional probability into tractable sub-problems to achieve extreme resolutions without additional training. Because visual cues diminish at high magnifications, we augment each zoom step with multi-scale-aware text prompts generated by a vision-language model (VLM). The prompt extractor itself is fine-tuned using Generalized Reward Policy Optimization (GRPO) with a critic VLM, aligning text guidance towards human preference. Experiments show that a standard 4x diffusion SR model wrapped in CoZ attains beyond 256x enlargement with high perceptual quality and fidelity. Project Page: https://bryanswkim.github.io/chain-of-zoom/ . | [
{
"version": "v1",
"created": "Sat, 24 May 2025 08:50:08 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 16:02:29 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kim",
"Bryan Sangwoo",
""
],
[
"Kim",
"Jeongsol",
""
],
[
"Ye",
"Jong Chul",
""
]
] |
2505.18609 | Ashwin Sankar | Ashwin Sankar and Yoach Lacombe and Sherry Thomas and Praveen Srinivasa Varadhan and Sanchit Gandhi and Mitesh M Khapra | RASMALAI: Resources for Adaptive Speech Modeling in Indian Languages with Accents and Intonations | Accepted at Interspeech 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | We introduce RASMALAI, a large-scale speech dataset with rich text descriptions, designed to advance controllable and expressive text-to-speech (TTS) synthesis for 23 Indian languages and English. It comprises 13,000 hours of speech and 24 million text-description annotations with fine-grained attributes like speaker identity, accent, emotion, style, and background conditions. Using RASMALAI, we develop IndicParlerTTS, the first open-source, text-description-guided TTS for Indian languages. Systematic evaluation demonstrates its ability to generate high-quality speech for named speakers, reliably follow text descriptions and accurately synthesize specified attributes. Additionally, it effectively transfers expressive characteristics both within and across languages. IndicParlerTTS consistently achieves strong performance across these evaluations, setting a new standard for controllable multilingual expressive speech synthesis in Indian languages. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 09:16:14 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:19:23 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Sankar",
"Ashwin",
""
],
[
"Lacombe",
"Yoach",
""
],
[
"Thomas",
"Sherry",
""
],
[
"Varadhan",
"Praveen Srinivasa",
""
],
[
"Gandhi",
"Sanchit",
""
],
[
"Khapra",
"Mitesh M",
""
]
] |
2505.18624 | Arslan Sikandar | Arslan Sikandar, M. Jamil Aslam | Radiative B to tensor meson decays at NLO in SCET | 13 pages, 5 figures, 3tables | J. Phys. G: Nucl. Part. Phys. 50 105001 (2023) | 10.1088/1361-6471/aceb7c | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | The radiative $B$ to tensor $\left(K_2^*(1430),\; f_2(1270),\; a_2(1230)\right)$ meson decays are studied at next-to-leading order (NLO) in soft-collinear effective theory (SCET). The SCET allows the systematic treatment of factorizable and non-factorizable contributions along with the resummation of large perturbative logarithms. We performed a two step matching and determined the soft-overlap function $\zeta^\perp_{T}$ and branching ratios for these $B$ to tensor meson decays. In the case of $B\to K_2^*(1430)\gamma$, the numerical value of the branching ratio lies close to its experimental measurements. The estimated values of the branching ratios of CKM suppressed decays $B \to \left(a_2(1230),\; f_2(1270)\right)\gamma$ are significant small compared to that of $B\to K_2^*(1430)\gamma$, but still could be measured in some ongoing and future $B$ physics experiments. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 10:18:27 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:56:45 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Sikandar",
"Arslan",
""
],
[
"Aslam",
"M. Jamil",
""
]
] |
2505.18654 | Han Ruidong | Ruidong Han, Bin Yin, Shangyu Chen, He Jiang, Fei Jiang, Xiang Li, Chi Ma, Mincong Huang, Xiaoguang Li, Chunzhen Jing, Yueming Han, Menglei Zhou, Lei Yu, Chuan Liu and Wei Lin | MTGR: Industrial-Scale Generative Recommendation Framework in Meituan | null | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Scaling law has been extensively validated in many domains such as natural language processing and computer vision. In the recommendation system, recent work has adopted generative recommendations to achieve scalability, but their generative approaches require abandoning the carefully constructed cross features of traditional recommendation models. We found that this approach significantly degrades model performance, and scaling up cannot compensate for it at all. In this paper, we propose MTGR (Meituan Generative Recommendation) to address this issue. MTGR is modeling based on the HSTU architecture and can retain the original deep learning recommendation model (DLRM) features, including cross features. Additionally, MTGR achieves training and inference acceleration through user-level compression to ensure efficient scaling. We also propose Group-Layer Normalization (GLN) to enhance the performance of encoding within different semantic spaces and the dynamic masking strategy to avoid information leakage. We further optimize the training frameworks, enabling support for our models with 10 to 100 times computational complexity compared to the DLRM, without significant cost increases. MTGR achieved 65x FLOPs for single-sample forward inference compared to the DLRM model, resulting in the largest gain in nearly two years both offline and online. This breakthrough was successfully deployed on Meituan, the world's largest food delivery platform, where it has been handling the main traffic. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 11:47:28 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:49:04 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Han",
"Ruidong",
""
],
[
"Yin",
"Bin",
""
],
[
"Chen",
"Shangyu",
""
],
[
"Jiang",
"He",
""
],
[
"Jiang",
"Fei",
""
],
[
"Li",
"Xiang",
""
],
[
"Ma",
"Chi",
""
],
[
"Huang",
"Mincong",
""
],
[
"Li",
"Xiaoguang",
""
],
[
"Jing",
"Chunzhen",
""
],
[
"Han",
"Yueming",
""
],
[
"Zhou",
"Menglei",
""
],
[
"Yu",
"Lei",
""
],
[
"Liu",
"Chuan",
""
],
[
"Lin",
"Wei",
""
]
] |
2505.18674 | Peng Xiao | Peng Xiao, Hongbo Zhao, Yijun Wang, Jianxin Lin | Restoring Real-World Images with an Internal Detail Enhancement Diffusion Model | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Restoring real-world degraded images, such as old photographs or low-resolution images, presents a significant challenge due to the complex, mixed degradations they exhibit, such as scratches, color fading, and noise. Recent data-driven approaches have struggled with two main challenges: achieving high-fidelity restoration and providing object-level control over colorization. While diffusion models have shown promise in generating high-quality images with specific controls, they often fail to fully preserve image details during restoration. In this work, we propose an internal detail-preserving diffusion model for high-fidelity restoration of real-world degraded images. Our method utilizes a pre-trained Stable Diffusion model as a generative prior, eliminating the need to train a model from scratch. Central to our approach is the Internal Image Detail Enhancement (IIDE) technique, which directs the diffusion model to preserve essential structural and textural information while mitigating degradation effects. The process starts by mapping the input image into a latent space, where we inject the diffusion denoising process with degradation operations that simulate the effects of various degradation factors. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art models in both qualitative assessments and perceptual quantitative evaluations. Additionally, our approach supports text-guided restoration, enabling object-level colorization control that mimics the expertise of professional photo editing. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 12:32:53 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:47:50 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Xiao",
"Peng",
""
],
[
"Zhao",
"Hongbo",
""
],
[
"Wang",
"Yijun",
""
],
[
"Lin",
"Jianxin",
""
]
] |
2505.18746 | Peijie Yu | Peijie Yu, Yifan Yang, Jinjian Li, Zelong Zhang, Haorui Wang, Xiao Feng, Feng Zhang | $C^3$-Bench: The Things Real Disturbing LLM based Agent in Multi-Tasking | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Agents based on large language models leverage tools to modify environments, revolutionizing how AI interacts with the physical world. Unlike traditional NLP tasks that rely solely on historical dialogue for responses, these agents must consider more complex factors, such as inter-tool relationships, environmental feedback and previous decisions, when making choices. Current research typically evaluates agents via multi-turn dialogues. However, it overlooks the influence of these critical factors on agent behavior. To bridge this gap, we present an open-source and high-quality benchmark $C^3$-Bench. This benchmark integrates attack concepts and applies univariate analysis to pinpoint key elements affecting agent robustness. In concrete, we design three challenges: navigate complex tool relationships, handle critical hidden information and manage dynamic decision paths. Complementing these challenges, we introduce fine-grained metrics, innovative data collection algorithms and reproducible evaluation methods. Extensive experiments are conducted on 49 mainstream agents, encompassing general fast-thinking, slow-thinking and domain-specific models. We observe that agents have significant shortcomings in handling tool dependencies, long context information dependencies and frequent policy-type switching. In essence, $C^3$-Bench aims to expose model vulnerabilities through these challenges and drive research into the interpretability of agent performance. The benchmark is publicly available at https://github.com/yupeijei1997/C3-Bench. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 15:25:44 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:22:28 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yu",
"Peijie",
""
],
[
"Yang",
"Yifan",
""
],
[
"Li",
"Jinjian",
""
],
[
"Zhang",
"Zelong",
""
],
[
"Wang",
"Haorui",
""
],
[
"Feng",
"Xiao",
""
],
[
"Zhang",
"Feng",
""
]
] |
2505.18754 | Elsen Ronando | Elsen Ronando and Sozo Inoue | Few-Shot Optimization for Sensor Data Using Large Language Models: A Case Study on Fatigue Detection | 43 pages, 18 figures. Accepted for publication in MDPI Sensors (2025). Final version before journal publication | Sensors 2025, 25, 3324 | 10.3390/s25113324 | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | In this paper, we propose a novel few-shot optimization with HED-LM (Hybrid Euclidean Distance with Large Language Models) to improve example selection for sensor-based classification tasks. While few-shot prompting enables efficient inference with limited labeled data, its performance largely depends on the quality of selected examples. HED-LM addresses this challenge through a hybrid selection pipeline that filters candidate examples based on Euclidean distance and re-ranks them using contextual relevance scored by large language models (LLMs). To validate its effectiveness, we apply HED-LM to a fatigue detection task using accelerometer data characterized by overlapping patterns and high inter-subject variability. Unlike simpler tasks such as activity recognition, fatigue detection demands more nuanced example selection due to subtle differences in physiological signals. Our experiments show that HED-LM achieves a mean macro F1-score of 69.13$\pm$10.71%, outperforming both random selection (59.30$\pm$10.13%) and distance-only filtering (67.61$\pm$11.39%). These represent relative improvements of 16.6% and 2.3%, respectively. The results confirm that combining numerical similarity with contextual relevance improves the robustness of few-shot prompting. Overall, HED-LM offers a practical solution to improve performance in real-world sensor-based learning tasks and shows potential for broader applications in healthcare monitoring, human activity recognition, and industrial safety scenarios. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 15:43:25 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ronando",
"Elsen",
""
],
[
"Inoue",
"Sozo",
""
]
] |
2505.18781 | Shizheng Wen | Shizheng Wen, Arsh Kumbhat, Levi Lingsch, Sepehr Mousavi, Yizhou Zhao, Praveen Chandrashekar, Siddhartha Mishra | Geometry Aware Operator Transformer as an Efficient and Accurate Neural Surrogate for PDEs on Arbitrary Domains | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The very challenging task of learning solution operators of PDEs on arbitrary domains accurately and efficiently is of vital importance to engineering and industrial simulations. Despite the existence of many operator learning algorithms to approximate such PDEs, we find that accurate models are not necessarily computationally efficient and vice versa. We address this issue by proposing a geometry aware operator transformer (GAOT) for learning PDEs on arbitrary domains. GAOT combines novel multiscale attentional graph neural operator encoders and decoders, together with geometry embeddings and (vision) transformer processors to accurately map information about the domain and the inputs into a robust approximation of the PDE solution. Multiple innovations in the implementation of GAOT also ensure computational efficiency and scalability. We demonstrate this significant gain in both accuracy and efficiency of GAOT over several baselines on a large number of learning tasks from a diverse set of PDEs, including achieving state of the art performance on a large scale three-dimensional industrial CFD dataset. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 16:35:10 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 07:19:05 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wen",
"Shizheng",
""
],
[
"Kumbhat",
"Arsh",
""
],
[
"Lingsch",
"Levi",
""
],
[
"Mousavi",
"Sepehr",
""
],
[
"Zhao",
"Yizhou",
""
],
[
"Chandrashekar",
"Praveen",
""
],
[
"Mishra",
"Siddhartha",
""
]
] |
2505.18820 | Wang Qing Prof | Weijun Kong and Qing Wang | Momentum space entanglement of four fermion field theory | 16 pages, 12figures | null | null | null | hep-th quant-ph | http://creativecommons.org/licenses/by/4.0/ | Momentum space entanglement of four fermion field theory is calculated from the Wilsonian effective action pertubatively using replica trick, local terms in low energy effective action are proved to be non-relevant pertubatively and nonlocal terms are the only source of entanglement between different momentum modes. The final result again can be represented by a set of basketball feynmann diagrams with new feynmann rules proposed to inteprete them. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 18:34:17 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:40:33 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kong",
"Weijun",
""
],
[
"Wang",
"Qing",
""
]
] |
2505.18827 | Antonio Pineda | Cesar Ayala and Antonio Pineda | The mass of the lightest gluelump | 33 pages, 13 figures | null | null | null | hep-ph hep-lat hep-th | http://creativecommons.org/licenses/by/4.0/ | We give the most up-to-date determinations of the normalization of the leading renormalons of the pole mass, the singlet static potential, the octet static potential, and the gluelump energy. They read $Z^{\rm MS}_m=-Z^{\rm MS}_{V_s}/2=\{0.604(17),0.551(20)\}$, $Z^{\rm MS}_{V_o}=\{0.136(8),0.121(13)\}$, and $Z^{\rm MS}_A=\{-1.343(36),-1.224(43)\}$, for $n_f=0$ and $n_f=3$ respectively. We obtain two independent renormalization group invariant and renormalization scale independent determinations of the energy of the ground state gluelump in the principal value summation scheme: $\Lambda_{B}^{\rm PV}=2.47(9)r_0^{-1}$ and $\Lambda_{B}^{\rm PV}=2.38(11)r_0^{-1}$. They combine in $\Lambda_{B}^{\rm PV}=2.44(7)r_0^{-1}$. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 18:54:31 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ayala",
"Cesar",
""
],
[
"Pineda",
"Antonio",
""
]
] |
2505.18920 | Georgio Hawi | Georgio Hawi, Peter S. Kim, Peter P. Lee | Sensitivity analysis-guided model reduction of a mathematical model of pembrolizumab therapy for de novo metastatic MSI-H/dMMR colorectal cancer | 88 pages in total with 49 pages for the main body and 39 pages for the supporting information, 8 figures in the main body, 16 tables in the main body, 6 tables in the supporting information. arXiv admin note: text overlap with arXiv:2411.12123 | null | null | null | q-bio.QM q-bio.CB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Colorectal cancer (CRC) is the third most commonly diagnosed cancer worldwide and the leading cause of cancer-related deaths in adults under 55, involving a complex interplay of biological processes such as dendritic cell (DC) maturation and migration, T cell activation and proliferation, cytokine production, and T cell and natural killer (NK) cell-mediated cancer cell killing. Microsatellite instability-high (MSI-H) CRC and deficient mismatch repair (dMMR) CRC constitute 15% of all CRC, and 4% of metastatic CRC, and exhibit remarkable responsiveness to immunotherapy, especially with PD-1 inhibitors such as pembrolizumab. Mathematical models of the underlying immunobiology and the interactions underpinning immune checkpoint blockade offer mechanistic insights into tumour-immune dynamics and provide avenues for treatment optimisation and the identification of novel therapeutic targets. We used our data-driven model of de novo metastatic MSI-H/dMMR CRC (dnmMCRC) and performed sensitivity analysis-guided model reduction using the FAST and EFAST methods. In this work, we constructed two simplified models of dnmMCRC: one that faithfully reproduces all of the original model's trajectories, and a second, minimal model that accurately replicates the original dynamics while being highly extensible for future inclusion of additional components to explore various aspects of the anti-tumour immune response. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 00:59:40 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:24:32 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Hawi",
"Georgio",
""
],
[
"Kim",
"Peter S.",
""
],
[
"Lee",
"Peter P.",
""
]
] |
2505.18942 | Honglin Bao | Honglin Bao, Siyang Wu, Jiwoong Choi, Yingrong Mao, James A. Evans | Language Models Surface the Unwritten Code of Science and Society | null | null | null | null | cs.CY cs.CL cs.DL | http://creativecommons.org/licenses/by/4.0/ | This paper calls on the research community not only to investigate how human biases are inherited by large language models (LLMs) but also to explore how these biases in LLMs can be leveraged to make society's "unwritten code" - such as implicit stereotypes and heuristics - visible and accessible for critique. We introduce a conceptual framework through a case study in science: uncovering hidden rules in peer review - the factors that reviewers care about but rarely state explicitly due to normative scientific expectations. The idea of the framework is to push LLMs to speak out their heuristics through generating self-consistent hypotheses - why one paper appeared stronger in reviewer scoring - among paired papers submitted to 45 computer science conferences, while iteratively searching deeper hypotheses from remaining pairs where existing hypotheses cannot explain. We observed that LLMs' normative priors about the internal characteristics of good science extracted from their self-talk, e.g. theoretical rigor, were systematically updated toward posteriors that emphasize storytelling about external connections, such as how the work is positioned and connected within and across literatures. This shift reveals the primacy of scientific myths about intrinsic properties driving scientific excellence rather than extrinsic contextualization and storytelling that influence conceptions of relevance and significance. Human reviewers tend to explicitly reward aspects that moderately align with LLMs' normative priors (correlation = 0.49) but avoid articulating contextualization and storytelling posteriors in their review comments (correlation = -0.14), despite giving implicit reward to them with positive scores. We discuss the broad applicability of the framework, leveraging LLMs as diagnostic tools to surface the tacit codes underlying human society, enabling more precisely targeted responsible AI. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 02:28:40 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:15:31 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Bao",
"Honglin",
""
],
[
"Wu",
"Siyang",
""
],
[
"Choi",
"Jiwoong",
""
],
[
"Mao",
"Yingrong",
""
],
[
"Evans",
"James A.",
""
]
] |
2505.18958 | Jiong Wu | Jiong Wu, Yang Xing, Boxiao Yu, Wei Shao, and Kuang Gong | CDPDNet: Integrating Text Guidance with Hybrid Vision Encoders for Medical Image Segmentation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Most publicly available medical segmentation datasets are only partially labeled, with annotations provided for a subset of anatomical structures. When multiple datasets are combined for training, this incomplete annotation poses challenges, as it limits the model's ability to learn shared anatomical representations among datasets. Furthermore, vision-only frameworks often fail to capture complex anatomical relationships and task-specific distinctions, leading to reduced segmentation accuracy and poor generalizability to unseen datasets. In this study, we proposed a novel CLIP-DINO Prompt-Driven Segmentation Network (CDPDNet), which combined a self-supervised vision transformer with CLIP-based text embedding and introduced task-specific text prompts to tackle these challenges. Specifically, the framework was constructed upon a convolutional neural network (CNN) and incorporated DINOv2 to extract both fine-grained and global visual features, which were then fused using a multi-head cross-attention module to overcome the limited long-range modeling capability of CNNs. In addition, CLIP-derived text embeddings were projected into the visual space to help model complex relationships among organs and tumors. To further address the partial label challenge and enhance inter-task discriminative capability, a Text-based Task Prompt Generation (TTPG) module that generated task-specific prompts was designed to guide the segmentation. Extensive experiments on multiple medical imaging datasets demonstrated that CDPDNet consistently outperformed existing state-of-the-art segmentation methods. Code and pretrained model are available at: https://github.com/wujiong-hub/CDPDNet.git. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 03:23:58 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:57:47 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wu",
"Jiong",
""
],
[
"Xing",
"Yang",
""
],
[
"Yu",
"Boxiao",
""
],
[
"Shao",
"Wei",
""
],
[
"Gong",
"Kuang",
""
]
] |
2505.18973 | Mengjia Xu | Sarang Patil, Ashish Parmanand Pandey, Ioannis Koutis, Mengjia Xu | Hierarchical Mamba Meets Hyperbolic Geometry: A New Paradigm for Structured Language Embeddings | 10 pages, 3 figures | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Selective state-space models have achieved great success in long-sequence modeling. However, their capacity for language representation, especially in complex hierarchical reasoning tasks, remains underexplored. Most large language models rely on flat Euclidean embeddings, limiting their ability to capture latent hierarchies. To address this limitation, we propose Hierarchical Mamba (HiM), integrating efficient Mamba2 with exponential growth and curved nature of hyperbolic geometry to learn hierarchy-aware language embeddings for deeper linguistic understanding. Mamba2-processed sequences are projected to the Poincare ball (via tangent-based mapping) or Lorentzian manifold (via cosine and sine-based mapping) with "learnable" curvature, optimized with a combined hyperbolic loss. Our HiM model facilitates the capture of relational distances across varying hierarchical levels, enabling effective long-range reasoning. This makes it well-suited for tasks like mixed-hop prediction and multi-hop inference in hierarchical classification. We evaluated our HiM with four linguistic and medical datasets for mixed-hop prediction and multi-hop inference tasks. Experimental results demonstrated that: 1) Both HiM models effectively capture hierarchical relationships for four ontological datasets, surpassing Euclidean baselines. 2) HiM-Poincare captures fine-grained semantic distinctions with higher h-norms, while HiM-Lorentz provides more stable, compact, and hierarchy-preserving embeddings favoring robustness over detail. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 04:45:06 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 01:24:12 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Patil",
"Sarang",
""
],
[
"Pandey",
"Ashish Parmanand",
""
],
[
"Koutis",
"Ioannis",
""
],
[
"Xu",
"Mengjia",
""
]
] |
2505.19030 | Zhengkang Guo | Wenhao Liu, Zhengkang Guo, Mingchen Xie, Jingwen Xu, Zisu Huang, Muzhao Tian, Jianhan Xu, Muling Wu, Xiaohua Wang, Changze Lv, He-Da Wang, Hu Yao, Xiaoqing Zheng, Xuanjing Huang | RECAST: Strengthening LLMs' Complex Instruction Following with Constraint-Verifiable Data | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) are increasingly expected to tackle complex tasks, driven by their expanding applications and users' growing proficiency in crafting sophisticated prompts. However, as the number of explicitly stated requirements increases (particularly more than 10 constraints), LLMs often struggle to accurately follow such complex instructions. To address this challenge, we propose RECAST, a novel framework for synthesizing datasets where each example incorporates far more constraints than those in existing benchmarks. These constraints are extracted from real-world prompt-response pairs to ensure practical relevance. RECAST enables automatic verification of constraint satisfaction via rule-based validators for quantitative constraints and LLM-based validators for qualitative ones. Using this framework, we construct RECAST-30K, a large-scale, high-quality dataset comprising 30k instances spanning 15 constraint types. Experimental results demonstrate that models fine-tuned on RECAST-30K show substantial improvements in following complex instructions. Moreover, the verifiability provided by RECAST enables the design of reward functions for reinforcement learning, which further boosts model performance on complex and challenging tasks. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 08:31:08 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 07:53:33 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Wenhao",
""
],
[
"Guo",
"Zhengkang",
""
],
[
"Xie",
"Mingchen",
""
],
[
"Xu",
"Jingwen",
""
],
[
"Huang",
"Zisu",
""
],
[
"Tian",
"Muzhao",
""
],
[
"Xu",
"Jianhan",
""
],
[
"Wu",
"Muling",
""
],
[
"Wang",
"Xiaohua",
""
],
[
"Lv",
"Changze",
""
],
[
"Wang",
"He-Da",
""
],
[
"Yao",
"Hu",
""
],
[
"Zheng",
"Xiaoqing",
""
],
[
"Huang",
"Xuanjing",
""
]
] |
2505.19034 | Teng Liu | Teng Liu, Andreas Morr, Sebastian Bathiany, Lana L. Blaschke, Zhen Qian, Chan Diao, Taylor Smith, Niklas Boers | The influence of data gaps and outliers on resilience indicators | null | null | null | null | nlin.AO nlin.CD physics.data-an physics.geo-ph | http://creativecommons.org/licenses/by/4.0/ | The resilience, or stability, of major Earth system components is increasingly threatened by anthropogenic pressures, demanding reliable early warning signals for abrupt and irreversible regime shifts. Widely used data-driven resilience indicators based on variance and autocorrelation detect `critical slowing down', a signature of decreasing stability. However, the interpretation of these indicators is hampered by poorly understood interdependencies and their susceptibility to common data issues such as missing values and outliers. Here, we establish a rigorous mathematical analysis of the statistical dependency between variance- and autocorrelation-based resilience indicators, revealing that their agreement is fundamentally driven by the time series' initial data point. Using synthetic and empirical data, we demonstrate that missing values substantially weaken indicator agreement, while outliers introduce systematic biases that lead to overestimation of resilience based on temporal autocorrelation. Our results provide a necessary and rigorous foundation for preprocessing strategies and accuracy assessments across the growing number of disciplines that use real-world data to infer changes in system resilience. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 08:34:09 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 07:36:12 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Teng",
""
],
[
"Morr",
"Andreas",
""
],
[
"Bathiany",
"Sebastian",
""
],
[
"Blaschke",
"Lana L.",
""
],
[
"Qian",
"Zhen",
""
],
[
"Diao",
"Chan",
""
],
[
"Smith",
"Taylor",
""
],
[
"Boers",
"Niklas",
""
]
] |
2505.19036 | HanFei Zhou | Hanfei Zhou and Lei Shi | Weak Physics Informed Neural Networks for Geometry Compatible Hyperbolic Conservation Laws on Manifolds | null | null | null | null | math.NA cs.NA stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Physics-informed neural networks (PINNs), owing to their mesh-free nature, offer a powerful approach for solving high-dimensional partial differential equations (PDEs) in complex geometries, including irregular domains. This capability effectively circumvents the challenges of mesh generation that traditional numerical methods face in high-dimensional or geometrically intricate settings. While recent studies have extended PINNs to manifolds, the theoretical foundations remain scarce. Existing theoretical analyses of PINNs in Euclidean space often rely on smoothness assumptions for the solutions. However, recent empirical evidence indicates that PINNs may struggle to approximate solutions with low regularity, such as those arising from nonlinear hyperbolic equations. In this paper, we develop a framework for PINNs tailored to the efficient approximation of weak solutions, particularly nonlinear hyperbolic equations defined on manifolds. We introduce a novel weak PINN (wPINN) formulation on manifolds that leverages the well-posedness theory to approximate entropy solutions of geometry-compatible hyperbolic conservation laws on manifolds. Employing tools from approximation theory, we establish a convergence analysis of the algorithm, including an analysis of approximation errors for time-dependent entropy solutions. This analysis provides insight into the accumulation of approximation errors over long time horizons. Notably, the network complexity depends only on the intrinsic dimension, independent of the ambient space dimension. Our results match the minimax rate in the d-dimensional Euclidean space, demonstrating that PINNs can alleviate the curse of dimensionality in the context of low-dimensional manifolds. Finally, we validate the performance of the proposed wPINN framework through numerical experiments, confirming its ability to efficiently approximate entropy solutions on manifolds. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 08:36:56 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhou",
"Hanfei",
""
],
[
"Shi",
"Lei",
""
]
] |
2505.19063 | Xin Ma | Xin Ma, Yaohui Wang, Xinyuan Chen, Tien-Tsin Wong, Cunjian Chen | Training-free Stylized Text-to-Image Generation with Fast Inference | Project Page: https://maxin-cn.github.io/omnipainter_project | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Although diffusion models exhibit impressive generative capabilities, existing methods for stylized image generation based on these models often require textual inversion or fine-tuning with style images, which is time-consuming and limits the practical applicability of large-scale diffusion models. To address these challenges, we propose a novel stylized image generation method leveraging a pre-trained large-scale diffusion model without requiring fine-tuning or any additional optimization, termed as OmniPainter. Specifically, we exploit the self-consistency property of latent consistency models to extract the representative style statistics from reference style images to guide the stylization process. Additionally, we then introduce the norm mixture of self-attention, which enables the model to query the most relevant style patterns from these statistics for the intermediate output content features. This mechanism also ensures that the stylized results align closely with the distribution of the reference style images. Our qualitative and quantitative experimental results demonstrate that the proposed method outperforms state-of-the-art approaches. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 09:38:23 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 09:13:49 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ma",
"Xin",
""
],
[
"Wang",
"Yaohui",
""
],
[
"Chen",
"Xinyuan",
""
],
[
"Wong",
"Tien-Tsin",
""
],
[
"Chen",
"Cunjian",
""
]
] |
2505.19075 | Jong Chul Ye | Jaemin Kim, Hangeol Chang, Hyunmin Hwang, Choonghan Kim, Jong Chul Ye | Universal Reasoner: A Single, Composable Plug-and-Play Reasoner for Frozen LLMs | 22 pages, typos corrected | null | null | null | cs.AI cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) have demonstrated remarkable general capabilities, but enhancing skills such as reasoning often demands substantial computational resources and may compromise their generalization. While Parameter-Efficient Fine-Tuning (PEFT) methods offer a more resource-conscious alternative, they typically requires retraining for each LLM backbone due to architectural dependencies. To address these challenges, here we propose Universal Reasoner (UniR) - a single, lightweight, composable, and plug-and-play reasoning module that can be used with any frozen LLM to endow it with specialized reasoning capabilities. Specifically, UniR decomposes the reward into a standalone reasoning module that is trained independently using predefined rewards, effectively translating trajectory-level signals into token-level guidance. Once trained, UniR can be combined with any frozen LLM at inference time by simply adding its output logits to those of the LLM backbone. This additive structure naturally enables modular composition: multiple UniR modules trained for different tasks can be jointly applied by summing their logits, enabling complex reasoning via composition. Experimental results on mathematical reasoning and machine translation tasks show that UniR significantly outperforms existing baseline fine-tuning methods using the Llama3.2 model. Furthermore, UniR demonstrates strong weak-to-strong generalization: reasoning modules trained on smaller models effectively guide much larger LLMs. This makes UniR a cost-efficient, adaptable, and robust solution for enhancing reasoning in LLMs without compromising their core capabilities. Code is open-sourced at https://github.com/hangeol/UniR | [
{
"version": "v1",
"created": "Sun, 25 May 2025 10:19:10 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 13:53:36 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kim",
"Jaemin",
""
],
[
"Chang",
"Hangeol",
""
],
[
"Hwang",
"Hyunmin",
""
],
[
"Kim",
"Choonghan",
""
],
[
"Ye",
"Jong Chul",
""
]
] |
2505.19081 | Ruiyang Xia | Ruiyang Xia, Dawei Zhou, Decheng Liu, Lin Yuan, Jie Li, Nannan Wang, Xinbo Gao | Towards Generalized Proactive Defense against Face Swapping with Contour-Hybrid Watermark | 16 pages, 11 figures, under review | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face swapping, recognized as a privacy and security concern, has prompted considerable defensive research. With the advancements in AI-generated content, the discrepancies between the real and swapped faces have become nuanced. Considering the difficulty of forged traces detection, we shift the focus to the face swapping purpose and proactively embed elaborate watermarks against unknown face swapping techniques. Given that the constant purpose is to swap the original face identity while preserving the background, we concentrate on the regions surrounding the face to ensure robust watermark generation, while embedding the contour texture and face identity information to achieve progressive image determination. The watermark is located in the facial contour and contains hybrid messages, dubbed the contour-hybrid watermark (CMark). Our approach generalizes face swapping detection without requiring any swapping techniques during training and the storage of large-scale messages in advance. Experiments conducted across 8 face swapping techniques demonstrate the superiority of our approach compared with state-of-the-art passive and proactive detectors while achieving a favorable balance between the image quality and watermark robustness. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 10:29:23 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:08:24 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Xia",
"Ruiyang",
""
],
[
"Zhou",
"Dawei",
""
],
[
"Liu",
"Decheng",
""
],
[
"Yuan",
"Lin",
""
],
[
"Li",
"Jie",
""
],
[
"Wang",
"Nannan",
""
],
[
"Gao",
"Xinbo",
""
]
] |
2505.19141 | Ruiwen Dong | Ruiwen Dong, Doron Shafrir | S-unit equations in modules and linear-exponential Diophantine equations | 80 pages, corrected spelling mistake for a name | null | null | null | math.NT cs.FL | http://creativecommons.org/licenses/by/4.0/ | Let $T$ be a positive integer, and $\mathcal{M}$ be a finitely presented module over the Laurent polynomial ring $\mathbb{Z}_{/T}[X_1^{\pm}, \ldots, X_N^{\pm}]$. We consider S-unit equations over $\mathcal{M}$: these are equations of the form $x_1 m_1 + \cdots + x_K m_K = m_0$, where the variables $x_1, \ldots, x_K$ range over the set of monomials (with coefficient 1) of $\mathbb{Z}_{/T}[X_1^{\pm}, \ldots, X_N^{\pm}]$. When $T$ is a power of a prime number $p$, we show that the solution set of an S-unit equation over $\mathcal{M}$ is effectively $p$-normal in the sense of Derksen and Masser (2015), generalizing their result on S-unit equations in fields of prime characteristic. When $T$ is an arbitrary positive integer, we show that deciding whether an S-unit equation over $\mathcal{M}$ admits a solution is Turing equivalent to solving a system of linear-exponential Diophantine equations, whose base contains the prime divisors of $T$. Combined with a recent result of Karimov, Luca, Nieuwveld, Ouaknine and Worrell (2025), this yields decidability when $T$ has at most two distinct prime divisors. This also shows that proving either decidability or undecidability in the case of arbitrary $T$ would entail major breakthroughs in number theory.
We mention some potential applications of our results, such as deciding Submonoid Membership in wreath products of the form $\mathbb{Z}_{/p^a q^b} \wr \mathbb{Z}^d$, as well as progressing towards solving the Skolem problem in rings whose additive group is torsion. More connections in these directions will be explored in follow up papers. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 13:22:39 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 09:30:42 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Dong",
"Ruiwen",
""
],
[
"Shafrir",
"Doron",
""
]
] |
2505.19150 | Byungki Ryu | Byungki Ryu, Ji Hui Son, Sungjin Park, Jaywan Chung, Hye-Jin Lim, SuJi Park, Yujeong Do, and SuDong Park | teMatDb: A High-Quality Thermoelectric Material Database with Self-Consistent ZT Filtering | 43 pages, 4 tables, 5 figures, 3 supporting tables, 10 supporting figures | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | This study presents a curated thermoelectric material database, teMatDb, constructed by digitizing literature-reported data. It includes temperature-dependent thermoelectric properties (TEPs), Seebeck coefficient, electrical resistivity, thermal conductivity, and figure of merit (ZT), along with metadata on materials and their corresponding publications. A self-consistent ZT (Sc-ZT) filter set was developed to measure ZT errors by comparing reported ZT's from figures with ZT's recalculated from digitized TEPs. Using this Sc-ZT protocol, we generated tMatDb272, comprising 14,717 temperature-property pairs from 272 high-quality TEP sets across 262 publications. The method identifies various types of ZT errors, such as resolution error, publication bias, ZT overestimation, interpolation and extrapolation error, and digitization noise, and excludes inconsistent samples from the dataset. teMatDb272 and the Sc-ZT filtering framework offer a robust dataset for data-driven and machine-learning-based materials design, device modeling, and future thermoelectric research. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 13:56:29 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 05:17:11 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ryu",
"Byungki",
""
],
[
"Son",
"Ji Hui",
""
],
[
"Park",
"Sungjin",
""
],
[
"Chung",
"Jaywan",
""
],
[
"Lim",
"Hye-Jin",
""
],
[
"Park",
"SuJi",
""
],
[
"Do",
"Yujeong",
""
],
[
"Park",
"SuDong",
""
]
] |
2505.19184 | Minh Nguyen | Pradyumna Shyama Prasad, Minh Nhat Nguyen | When Two LLMs Debate, Both Think They'll Win | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Can LLMs accurately adjust their confidence when facing opposition? Building on previous studies measuring calibration on static fact-based question-answering tasks, we evaluate Large Language Models (LLMs) in a dynamic, adversarial debate setting, uniquely combining two realistic factors: (a) a multi-turn format requiring models to update beliefs as new information emerges, and (b) a zero-sum structure to control for task-related uncertainty, since mutual high-confidence claims imply systematic overconfidence. We organized 60 three-round policy debates among ten state-of-the-art LLMs, with models privately rating their confidence (0-100) in winning after each round. We observed five concerning patterns: (1) Systematic overconfidence: models began debates with average initial confidence of 72.9% vs. a rational 50% baseline. (2) Confidence escalation: rather than reducing confidence as debates progressed, debaters increased their win probabilities, averaging 83% by the final round. (3) Mutual overestimation: in 61.7% of debates, both sides simultaneously claimed >=75% probability of victory, a logical impossibility. (4) Persistent self-debate bias: models debating identical copies increased confidence from 64.1% to 75.2%; even when explicitly informed their chance of winning was exactly 50%, confidence still rose (from 50.0% to 57.1%). (5) Misaligned private reasoning: models' private scratchpad thoughts sometimes differed from their public confidence ratings, raising concerns about faithfulness of chain-of-thought reasoning. These results suggest LLMs lack the ability to accurately self-assess or update their beliefs in dynamic, multi-turn tasks; a major concern as LLM outputs are deployed without careful review in assistant roles or agentic settings. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 15:06:17 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 17:17:17 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Prasad",
"Pradyumna Shyama",
""
],
[
"Nguyen",
"Minh Nhat",
""
]
] |
2505.19197 | Jihoon Kwon | Chanyeol Choi, Alejandro Lopez-Lira, Jihoon Kwon, Minjae Kim, Juneha Hwang, Minsoo Ha, Chaewoon Kim, Jaeseon Ha, Suyeol Yun, Jin Kim | Structuring the Unstructured: A Multi-Agent System for Extracting and Querying Financial KPIs and Guidance | 7 pages, FinIR'25 | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Extracting structured and quantitative insights from unstructured financial filings is essential in investment research, yet remains time-consuming and resource-intensive. Conventional approaches in practice rely heavily on labor-intensive manual processes, limiting scalability and delaying the research workflow. In this paper, we propose an efficient and scalable method for accurately extracting quantitative insights from unstructured financial documents, leveraging a multi-agent system composed of large language models. Our proposed multi-agent system consists of two specialized agents: the \emph{Extraction Agent} and the \emph{Text-to-SQL Agent}. The \textit{Extraction Agent} automatically identifies key performance indicators from unstructured financial text, standardizes their formats, and verifies their accuracy. On the other hand, the \textit{Text-to-SQL Agent} generates executable SQL statements from natural language queries, allowing users to access structured data accurately without requiring familiarity with the database schema. Through experiments, we demonstrate that our proposed system effectively transforms unstructured text into structured data accurately and enables precise retrieval of key information. First, we demonstrate that our system achieves approximately 95\% accuracy in transforming financial filings into structured data, matching the performance level typically attained by human annotators. Second, in a human evaluation of the retrieval task -- where natural language queries are used to search information from structured data -- 91\% of the responses were rated as correct by human evaluators. In both evaluations, our system generalizes well across financial document types, consistently delivering reliable performance. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 15:45:46 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 13:32:03 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Choi",
"Chanyeol",
""
],
[
"Lopez-Lira",
"Alejandro",
""
],
[
"Kwon",
"Jihoon",
""
],
[
"Kim",
"Minjae",
""
],
[
"Hwang",
"Juneha",
""
],
[
"Ha",
"Minsoo",
""
],
[
"Kim",
"Chaewoon",
""
],
[
"Ha",
"Jaeseon",
""
],
[
"Yun",
"Suyeol",
""
],
[
"Kim",
"Jin",
""
]
] |
2505.19216 | Ehud Shapiro | Idit Keidar, Andrew Lewis-Pye, and Ehud Shapiro | Grassroots Consensus | null | null | null | null | cs.DC cs.CR cs.DS cs.NI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Grassroots platforms aim to offer an egalitarian alternative to global platforms -- centralized/autocratic and decentralized/plutocratic alike. Within the grassroots architecture, consensus is needed to realize platforms that employ digital social contracts, which are like smart contracts except that they are among people not accounts and are executed by these people's smartphones not by high-performance servers controlled by parties outside to the contract. Key envisioned grassroots platforms include sovereign democratic digital communities and federations, community banks and their grassroots cryptocurrencies, and digital cooperatives.
The grassroots architecture can benefit from a consensus protocol that is (i) quiescent, (ii) efficient during low- and high-throughput, (iii) responsive, (iv) blocklace-based, (v) UDP-ready, and (vi) grassroots. The Grassroots Consensus protocol addresses all these requirements while having competitive performance in both low- and high-throughput scenarios and being one of the most concise and elegant consensus protocols for partial synchrony. It achieves that by building on two cutting-edge consensus protocols -- the quiescent high-performance Morpheus and the blocklace-based Cordial Miners, improving the latter's dissemination protocol and making it UDP-ready, and extending the protocol with a constitution and a constitutional amendment component, making it grassroots. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 16:23:44 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:20:49 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Keidar",
"Idit",
""
],
[
"Lewis-Pye",
"Andrew",
""
],
[
"Shapiro",
"Ehud",
""
]
] |
2505.19286 | Yu Wang | Utkarsh Sahu, Zhisheng Qi, Yongjia Lei, Ryan A. Rossi, Franck Dernoncourt, Nesreen K. Ahmed, Mahantesh M Halappanavar, Yao Ma, Yu Wang | A Graph Perspective to Probe Structural Patterns of Knowledge in Large Language Models | null | null | null | null | cs.CL cs.LG cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models have been extensively studied as neural knowledge bases for their knowledge access, editability, reasoning, and explainability. However, few works focus on the structural patterns of their knowledge. Motivated by this gap, we investigate these structural patterns from a graph perspective. We quantify the knowledge of LLMs at both the triplet and entity levels, and analyze how it relates to graph structural properties such as node degree. Furthermore, we uncover the knowledge homophily, where topologically close entities exhibit similar levels of knowledgeability, which further motivates us to develop graph machine learning models to estimate entity knowledge based on its local neighbors. This model further enables valuable knowledge checking by selecting triplets less known to LLMs. Empirical results show that using selected triplets for fine-tuning leads to superior performance. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 19:34:15 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 06:58:53 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Sahu",
"Utkarsh",
""
],
[
"Qi",
"Zhisheng",
""
],
[
"Lei",
"Yongjia",
""
],
[
"Rossi",
"Ryan A.",
""
],
[
"Dernoncourt",
"Franck",
""
],
[
"Ahmed",
"Nesreen K.",
""
],
[
"Halappanavar",
"Mahantesh M",
""
],
[
"Ma",
"Yao",
""
],
[
"Wang",
"Yu",
""
]
] |
2505.19337 | Kevin Li | Kevin Li, Marinka Zitnik | Prompting Decision Transformers for Zero-Shot Reach-Avoid Policies | null | null | null | null | cs.LG cs.AI q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | Offline goal-conditioned reinforcement learning methods have shown promise for reach-avoid tasks, where an agent must reach a target state while avoiding undesirable regions of the state space. Existing approaches typically encode avoid-region information into an augmented state space and cost function, which prevents flexible, dynamic specification of novel avoid-region information at evaluation time. They also rely heavily on well-designed reward and cost functions, limiting scalability to complex or poorly structured environments. We introduce RADT, a decision transformer model for offline, reward-free, goal-conditioned, avoid region-conditioned RL. RADT encodes goals and avoid regions directly as prompt tokens, allowing any number of avoid regions of arbitrary size to be specified at evaluation time. Using only suboptimal offline trajectories from a random policy, RADT learns reach-avoid behavior through a novel combination of goal and avoid-region hindsight relabeling. We benchmark RADT against 3 existing offline goal-conditioned RL models across 11 tasks, environments, and experimental settings. RADT generalizes in a zero-shot manner to out-of-distribution avoid region sizes and counts, outperforming baselines that require retraining. In one such zero-shot setting, RADT achieves 35.7% improvement in normalized cost over the best retrained baseline while maintaining high goal-reaching success. We apply RADT to cell reprogramming in biology, where it reduces visits to undesirable intermediate gene expression states during trajectories to desired target states, despite stochastic transitions and discrete, structured state dynamics. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 22:00:38 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:56:11 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Li",
"Kevin",
""
],
[
"Zitnik",
"Marinka",
""
]
] |
2505.19381 | Anqing Jiang | Anqing Jiang, Yu Gao, Zhigang Sun, Yiru Wang, Jijun Wang, Jinghao Chai, Qian Cao, Yuweng Heng, Hao Jiang, Zongzheng Zhang, Xianda Guo, Hao Sun, Hao Zhao | DiffVLA: Vision-Language Guided Diffusion Planning for Autonomous Driving | 4pages | null | null | null | cs.AI cs.CV cs.RO | http://creativecommons.org/licenses/by/4.0/ | Research interest in end-to-end autonomous driving has surged owing to its fully differentiable design integrating modular tasks, i.e. perception, prediction and planing, which enables optimization in pursuit of the ultimate goal. Despite the great potential of the end-to-end paradigm, existing methods suffer from several aspects including expensive BEV (bird's eye view) computation, action diversity, and sub-optimal decision in complex real-world scenarios. To address these challenges, we propose a novel hybrid sparse-dense diffusion policy, empowered by a Vision-Language Model (VLM), called Diff-VLA. We explore the sparse diffusion representation for efficient multi-modal driving behavior. Moreover, we rethink the effectiveness of VLM driving decision and improve the trajectory generation guidance through deep interaction across agent, map instances and VLM output. Our method shows superior performance in Autonomous Grand Challenge 2025 which contains challenging real and reactive synthetic scenarios. Our methods achieves 45.0 PDMS. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 00:49:35 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 06:45:19 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Jiang",
"Anqing",
""
],
[
"Gao",
"Yu",
""
],
[
"Sun",
"Zhigang",
""
],
[
"Wang",
"Yiru",
""
],
[
"Wang",
"Jijun",
""
],
[
"Chai",
"Jinghao",
""
],
[
"Cao",
"Qian",
""
],
[
"Heng",
"Yuweng",
""
],
[
"Jiang",
"Hao",
""
],
[
"Zhang",
"Zongzheng",
""
],
[
"Guo",
"Xianda",
""
],
[
"Sun",
"Hao",
""
],
[
"Zhao",
"Hao",
""
]
] |
2505.19419 | Baichuan Li | Baichuan Li, Larry Powell, Tracy Hammond | It's Not Just Labeling -- A Research on LLM Generated Feedback Interpretability and Image Labeling Sketch Features | null | null | null | null | cs.HC cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The quality of training data is critical to the performance of machine learning applications in domains like transportation, healthcare, and robotics. Accurate image labeling, however, often relies on time-consuming, expert-driven methods with limited feedback. This research introduces a sketch-based annotation approach supported by large language models (LLMs) to reduce technical barriers and enhance accessibility. Using a synthetic dataset, we examine how sketch recognition features relate to LLM feedback metrics, aiming to improve the reliability and interpretability of LLM-assisted labeling. We also explore how prompting strategies and sketch variations influence feedback quality. Our main contribution is a sketch-based virtual assistant that simplifies annotation for non-experts and advances LLM-driven labeling tools in terms of scalability, accessibility, and explainability. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 02:13:52 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:53:28 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Li",
"Baichuan",
""
],
[
"Powell",
"Larry",
""
],
[
"Hammond",
"Tracy",
""
]
] |
2505.19446 | Yinlong Liu | Yin-Long Liu, Yuanchao Li, Rui Feng, Liu He, Jia-Xin Chen, Yi-Ming Wang, Yu-Ang Chen, Yan-Han Peng, Jia-Hong Yuan, Zhen-Hua Ling | Leveraging Cascaded Binary Classification and Multimodal Fusion for Dementia Detection through Spontaneous Speech | Accepted by Interspeech 2025 | null | null | null | eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents our submission to the PROCESS Challenge 2025, focusing on spontaneous speech analysis for early dementia detection. For the three-class classification task (Healthy Control, Mild Cognitive Impairment, and Dementia), we propose a cascaded binary classification framework that fine-tunes pre-trained language models and incorporates pause encoding to better capture disfluencies. This design streamlines multi-class classification and addresses class imbalance by restructuring the decision process. For the Mini-Mental State Examination score regression task, we develop an enhanced multimodal fusion system that combines diverse acoustic and linguistic features. Separate regression models are trained on individual feature sets, with ensemble learning applied through score averaging. Experimental results on the test set outperform the baselines provided by the organizers in both tasks, demonstrating the robustness and effectiveness of our approach. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 03:08:55 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:35:23 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Yin-Long",
""
],
[
"Li",
"Yuanchao",
""
],
[
"Feng",
"Rui",
""
],
[
"He",
"Liu",
""
],
[
"Chen",
"Jia-Xin",
""
],
[
"Wang",
"Yi-Ming",
""
],
[
"Chen",
"Yu-Ang",
""
],
[
"Peng",
"Yan-Han",
""
],
[
"Yuan",
"Jia-Hong",
""
],
[
"Ling",
"Zhen-Hua",
""
]
] |
2505.19451 | Shijie Bao | Shijie Bao, Qi'an Guan and Lin Zhou | Algebraic Zhou valuations | 43 pages. All comments are welcome! | null | null | null | math.AG math.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we generalize Zhou valuations, originally defined on complex domains, to the framework of general schemes. We demonstrate that an algebraic version of the Jonsson--Musta\c{t}\u{a} conjecture is equivalent to the statement that every Zhou valuation is quasi-monomial. By introducing a mixed version of jumping numbers and Tian functions associated with valuations, we obtain characterizations of a valuation being a Zhou valuation or computing some jumping number using the Tian functions. Furthermore, we establish the correspondence between Zhou valuations in algebraic settings and their counterparts in analytic settings. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 03:17:45 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Bao",
"Shijie",
""
],
[
"Guan",
"Qi'an",
""
],
[
"Zhou",
"Lin",
""
]
] |
2505.19476 | Ziqian Wang | Ziqian Wang, Zikai Liu, Xinfa Zhu, Yike Zhu, Mingshuai Liu, Jun Chen, Longshuai Xiao, Chao Weng, Lei Xie | FlowSE: Efficient and High-Quality Speech Enhancement via Flow Matching | Accepted to InterSpeech 2025 | null | null | null | eess.AS eess.SP | http://creativecommons.org/licenses/by/4.0/ | Generative models have excelled in audio tasks using approaches such as language models, diffusion, and flow matching. However, existing generative approaches for speech enhancement (SE) face notable challenges: language model-based methods suffer from quantization loss, leading to compromised speaker similarity and intelligibility, while diffusion models require complex training and high inference latency. To address these challenges, we propose FlowSE, a flow-matching-based model for SE. Flow matching learns a continuous transformation between noisy and clean speech distributions in a single pass, significantly reducing inference latency while maintaining high-quality reconstruction. Specifically, FlowSE trains on noisy mel spectrograms and optional character sequences, optimizing a conditional flow matching loss with ground-truth mel spectrograms as supervision. It implicitly learns speech's temporal-spectral structure and text-speech alignment. During inference, FlowSE can operate with or without textual information, achieving impressive results in both scenarios, with further improvements when transcripts are available. Extensive experiments demonstrate that FlowSE significantly outperforms state-of-the-art generative methods, establishing a new paradigm for generative-based SE and demonstrating the potential of flow matching to advance the field. Our code, pre-trained checkpoints, and audio samples are available. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 03:55:00 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 09:33:56 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Ziqian",
""
],
[
"Liu",
"Zikai",
""
],
[
"Zhu",
"Xinfa",
""
],
[
"Zhu",
"Yike",
""
],
[
"Liu",
"Mingshuai",
""
],
[
"Chen",
"Jun",
""
],
[
"Xiao",
"Longshuai",
""
],
[
"Weng",
"Chao",
""
],
[
"Xie",
"Lei",
""
]
] |
2505.19484 | Ruixiang Feng | Ruixiang Feng, Shen Gao, Xiuying Chen, Lisi Chen, Shuo Shang | CulFiT: A Fine-grained Cultural-aware LLM Training Paradigm via Multilingual Critique Data Synthesis | accepted by ACL 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks, yet they often exhibit a specific cultural biases, neglecting the values and linguistic diversity of low-resource regions. This cultural bias not only undermines universal equality, but also risks reinforcing stereotypes and perpetuating discrimination. To address this, we propose CulFiT, a novel culturally-aware training paradigm that leverages multilingual data and fine-grained reward modeling to enhance cultural sensitivity and inclusivity. Our approach synthesizes diverse cultural-related questions, constructs critique data in culturally relevant languages, and employs fine-grained rewards to decompose cultural texts into verifiable knowledge units for interpretable evaluation. We also introduce GlobalCultureQA, a multilingual open-ended question-answering dataset designed to evaluate culturally-aware responses in a global context. Extensive experiments on three existing benchmarks and our GlobalCultureQA demonstrate that CulFiT achieves state-of-the-art open-source model performance in cultural alignment and general reasoning. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 04:08:26 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 07:18:37 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Feng",
"Ruixiang",
""
],
[
"Gao",
"Shen",
""
],
[
"Chen",
"Xiuying",
""
],
[
"Chen",
"Lisi",
""
],
[
"Shang",
"Shuo",
""
]
] |
2505.19515 | Lavanya Prahallad | Lavanya Prahallad, Radhika Mamidi | Analyzing Biases in Political Dialogue: Tagging U.S. Presidential Debates with an Extended DAMSL Framework | 8 pages | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | We present a critical discourse analysis of the 2024 U.S. presidential debates, examining Donald Trump's rhetorical strategies in his interactions with Joe Biden and Kamala Harris. We introduce a novel annotation framework, BEADS (Bias Enriched Annotation for Dialogue Structure), which systematically extends the DAMSL framework to capture bias driven and adversarial discourse features in political communication. BEADS includes a domain and language agnostic set of tags that model ideological framing, emotional appeals, and confrontational tactics. Our methodology compares detailed human annotation with zero shot ChatGPT assisted tagging on verified transcripts from the Trump and Biden (19,219 words) and Trump and Harris (18,123 words) debates. Our analysis shows that Trump consistently dominated in key categories: Challenge and Adversarial Exchanges, Selective Emphasis, Appeal to Fear, Political Bias, and Perceived Dismissiveness. These findings underscore his use of emotionally charged and adversarial rhetoric to control the narrative and influence audience perception. In this work, we establish BEADS as a scalable and reproducible framework for critical discourse analysis across languages, domains, and political contexts. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 04:58:08 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 04:54:43 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Prahallad",
"Lavanya",
""
],
[
"Mamidi",
"Radhika",
""
]
] |
2505.19519 | Gihoon Kim | Gihoon Kim, Hyungjin Park and Taesup Kim | Regularized Personalization of Text-to-Image Diffusion Models without Distributional Drift | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Personalization using text-to-image diffusion models involves adapting a pretrained model to novel subjects with only a few image examples. This task presents a fundamental challenge, as the model must not only learn the new subject effectively but also preserve its ability to generate diverse and coherent outputs across a wide range of prompts. In other words, successful personalization requires integrating new concepts without forgetting previously learned generative capabilities. Forgetting denotes unintended distributional drift, where the model's output distribution deviates from that of the original pretrained model. In this paper, we provide an analysis of this issue and identify a mismatch between standard training objectives and the goals of personalization. To address this, we propose a new training objective based on a Lipschitz-bounded formulation that explicitly constrains deviation from the pretrained distribution. Our method provides improved control over distributional drift and performs well even in data-scarce scenarios. Experimental results demonstrate that our approach consistently outperforms existing personalization methods, achieving higher CLIP-T, CLIP-I, and DINO scores. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 05:03:59 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 15:31:32 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kim",
"Gihoon",
""
],
[
"Park",
"Hyungjin",
""
],
[
"Kim",
"Taesup",
""
]
] |
2505.19528 | Yejin Lee | Yejin Lee, Joonghyuk Hahn, Hyeseon Ahn, Yo-Sub Han | AmpleHate: Amplifying the Attention for Versatile Implicit Hate Detection | 13 pages, 4 figures, Under Review | null | null | null | cs.CL cs.AI cs.CY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Implicit hate speech detection is challenging due to its subtlety and reliance on contextual interpretation rather than explicit offensive words. Current approaches rely on contrastive learning, which are shown to be effective on distinguishing hate and non-hate sentences. Humans, however, detect implicit hate speech by first identifying specific targets within the text and subsequently interpreting how these target relate to their surrounding context. Motivated by this reasoning process, we propose AmpleHate, a novel approach designed to mirror human inference for implicit hate detection. AmpleHate identifies explicit target using a pretrained Named Entity Recognition model and capture implicit target information via [CLS] tokens. It computes attention-based relationships between explicit, implicit targets and sentence context and then, directly injects these relational vectors into the final sentence representation. This amplifies the critical signals of target-context relations for determining implicit hate. Experiments demonstrate that AmpleHate achieves state-of-the-art performance, outperforming contrastive learning baselines by an average of 82.14% and achieve faster convergence. Qualitative analyses further reveal that attention patterns produced by AmpleHate closely align with human judgement, underscoring its interpretability and robustness. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 05:27:10 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 08:18:31 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Lee",
"Yejin",
""
],
[
"Hahn",
"Joonghyuk",
""
],
[
"Ahn",
"Hyeseon",
""
],
[
"Han",
"Yo-Sub",
""
]
] |
2505.19547 | Haoyu Zhang | Haoyu Zhang, Wentao Zhang, Hao Miao, Xinke Jiang, Yuchen Fang, Yifan Zhang | STRAP: Spatio-Temporal Pattern Retrieval for Out-of-Distribution Generalization | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spatio-Temporal Graph Neural Networks (STGNNs) have emerged as a powerful tool for modeling dynamic graph-structured data across diverse domains. However, they often fail to generalize in Spatio-Temporal Out-of-Distribution (STOOD) scenarios, where both temporal dynamics and spatial structures evolve beyond the training distribution. To address this problem, we propose an innovative Spatio-Temporal Retrieval-Augmented Pattern Learning framework,STRAP, which enhances model generalization by integrating retrieval-augmented learning into the STGNN continue learning pipeline. The core of STRAP is a compact and expressive pattern library that stores representative spatio-temporal patterns enriched with historical, structural, and semantic information, which is obtained and optimized during the training phase. During inference, STRAP retrieves relevant patterns from this library based on similarity to the current input and injects them into the model via a plug-and-play prompting mechanism. This not only strengthens spatio-temporal representations but also mitigates catastrophic forgetting. Moreover, STRAP introduces a knowledge-balancing objective to harmonize new information with retrieved knowledge. Extensive experiments across multiple real-world streaming graph datasets show that STRAP consistently outperforms state-of-the-art STGNN baselines on STOOD tasks, demonstrating its robustness, adaptability, and strong generalization capability without task-specific fine-tuning. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 06:11:05 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:17:24 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Haoyu",
""
],
[
"Zhang",
"Wentao",
""
],
[
"Miao",
"Hao",
""
],
[
"Jiang",
"Xinke",
""
],
[
"Fang",
"Yuchen",
""
],
[
"Zhang",
"Yifan",
""
]
] |
2505.19586 | Dingyu Yao | Dingyu Yao, Bowen Shen, Zheng Lin, Wei Liu, Jian Luan, Bin Wang, Weiping Wang | TailorKV: A Hybrid Framework for Long-Context Inference via Tailored KV Cache Optimization | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The Key-Value (KV) cache in generative large language models (LLMs) introduces substantial memory overhead. Existing works mitigate this burden by offloading or compressing the KV cache. However, loading the entire cache incurs significant latency due to PCIe bandwidth bottlenecks in CPU-GPU communication, while aggressive compression causes notable performance degradation. We identify that certain layers in the LLM need to maintain global information and are unsuitable for selective loading. In contrast, other layers primarily focus on a few tokens with dominant activations that potentially incur substantial quantization error. This observation leads to a key insight that loading dominant tokens and quantizing all tokens can complement each other. Building on this insight, we propose a hybrid compression method, TailorKV, which seamlessly integrates quantization and offloading. TailorKV develops an inference framework along with a hardware-friendly implementation that leverages these complementary characteristics. Extensive long-context evaluations exhibit that TailorKV achieves nearly lossless performance under aggressive compression settings, outperforming the state-of-the-art. Particularly, the Llama-3.1-8B with 128k context can be served within a single RTX 3090 GPU, reaching 82 ms per token during decoding. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 07:00:04 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:16:32 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Yao",
"Dingyu",
""
],
[
"Shen",
"Bowen",
""
],
[
"Lin",
"Zheng",
""
],
[
"Liu",
"Wei",
""
],
[
"Luan",
"Jian",
""
],
[
"Wang",
"Bin",
""
],
[
"Wang",
"Weiping",
""
]
] |
2505.19589 | Mathieu Even | Christian Lebeda, Mathieu Even, Aur\'elien Bellet, Julie Josse | Model Agnostic Differentially Private Causal Inference | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Estimating causal effects from observational data is essential in fields such as medicine, economics and social sciences, where privacy concerns are paramount. We propose a general, model-agnostic framework for differentially private estimation of average treatment effects (ATE) that avoids strong structural assumptions on the data-generating process or the models used to estimate propensity scores and conditional outcomes. In contrast to prior work, which enforces differential privacy by directly privatizing these nuisance components and results in a privacy cost that scales with model complexity, our approach decouples nuisance estimation from privacy protection. This separation allows the use of flexible, state-of-the-art black-box models, while differential privacy is achieved by perturbing only predictions and aggregation steps within a fold-splitting scheme with ensemble techniques. We instantiate the framework for three classical estimators -- the G-formula, inverse propensity weighting (IPW), and augmented IPW (AIPW) -- and provide formal utility and privacy guarantees. Empirical results show that our methods maintain competitive performance under realistic privacy budgets. We further extend our framework to support meta-analysis of multiple private ATE estimates. Our results bridge a critical gap between causal inference and privacy-preserving data analysis. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 07:00:37 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 08:07:31 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Lebeda",
"Christian",
""
],
[
"Even",
"Mathieu",
""
],
[
"Bellet",
"Aurélien",
""
],
[
"Josse",
"Julie",
""
]
] |
2505.19628 | Silin Li | Silin Li and Yuhang Guo and Jiashu Yao and Zeming Liu and Haifeng Wang | HomeBench: Evaluating LLMs in Smart Homes with Valid and Invalid Instructions Across Single and Multiple Devices | Accepted by ACL 2025 Main | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have the potential to revolutionize smart home assistants by enhancing their ability to accurately understand user needs and respond appropriately, which is extremely beneficial for building a smarter home environment. While recent studies have explored integrating LLMs into smart home systems, they primarily focus on handling straightforward, valid single-device operation instructions. However, real-world scenarios are far more complex and often involve users issuing invalid instructions or controlling multiple devices simultaneously. These have two main challenges: LLMs must accurately identify and rectify errors in user instructions and execute multiple user instructions perfectly. To address these challenges and advance the development of LLM-based smart home assistants, we introduce HomeBench, the first smart home dataset with valid and invalid instructions across single and multiple devices in this paper. We have experimental results on 13 distinct LLMs; e.g., GPT-4o achieves only a 0.0% success rate in the scenario of invalid multi-device instructions, revealing that the existing state-of-the-art LLMs still cannot perform well in this situation even with the help of in-context learning, retrieval-augmented generation, and fine-tuning. Our code and dataset are publicly available at https://github.com/BITHLP/HomeBench. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 07:47:39 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 07:26:36 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Li",
"Silin",
""
],
[
"Guo",
"Yuhang",
""
],
[
"Yao",
"Jiashu",
""
],
[
"Liu",
"Zeming",
""
],
[
"Wang",
"Haifeng",
""
]
] |
2505.19646 | Dongyeop Woo | Dongyeop Woo, Minsu Kim, Minkyu Kim, Kiyoung Seong, Sungsoo Ahn | Energy-based generator matching: A neural sampler for general state space | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We propose Energy-based generator matching (EGM), a modality-agnostic approach to train generative models from energy functions in the absence of data. Extending the recently proposed generator matching, EGM enables training of arbitrary continuous-time Markov processes, e.g., diffusion, flow, and jump, and can generate data from continuous, discrete, and a mixture of two modalities. To this end, we propose estimating the generator matching loss using self-normalized importance sampling with an additional bootstrapping trick to reduce variance in the importance weight. We validate EGM on both discrete and multimodal tasks up to 100 and 20 dimensions, respectively. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 08:02:29 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 05:18:11 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Woo",
"Dongyeop",
""
],
[
"Kim",
"Minsu",
""
],
[
"Kim",
"Minkyu",
""
],
[
"Seong",
"Kiyoung",
""
],
[
"Ahn",
"Sungsoo",
""
]
] |
2505.19650 | Fanheng Kong | Fanheng Kong, Jingyuan Zhang, Yahui Liu, Hongzhi Zhang, Shi Feng, Xiaocui Yang, Daling Wang, Yu Tian, Victoria W., Fuzheng Zhang, Guorui Zhou | Modality Curation: Building Universal Embeddings for Advanced Multimodal Information Retrieval | 26 pages, project page: https://friedrichor.github.io/projects/UNITE | null | null | null | cs.CV cs.IR cs.MM | http://creativecommons.org/licenses/by/4.0/ | Multimodal information retrieval (MIR) faces inherent challenges due to the heterogeneity of data sources and the complexity of cross-modal alignment. While previous studies have identified modal gaps in feature spaces, a systematic approach to address these challenges remains unexplored. In this work, we introduce UNITE, a universal framework that tackles these challenges through two critical yet underexplored aspects: data curation and modality-aware training configurations. Our work provides the first comprehensive analysis of how modality-specific data properties influence downstream task performance across diverse scenarios. Moreover, we propose Modal-Aware Masked Contrastive Learning (MAMCL) to mitigate the competitive relationships among the instances of different modalities. Our framework achieves state-of-the-art results on multiple multimodal retrieval benchmarks, outperforming existing methods by notable margins. Through extensive experiments, we demonstrate that strategic modality curation and tailored training protocols are pivotal for robust cross-modal representation learning. This work not only advances MIR performance but also provides a foundational blueprint for future research in multimodal systems. Our project is available at https://friedrichor.github.io/projects/UNITE. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 08:09:44 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 11:57:17 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kong",
"Fanheng",
""
],
[
"Zhang",
"Jingyuan",
""
],
[
"Liu",
"Yahui",
""
],
[
"Zhang",
"Hongzhi",
""
],
[
"Feng",
"Shi",
""
],
[
"Yang",
"Xiaocui",
""
],
[
"Wang",
"Daling",
""
],
[
"Tian",
"Yu",
""
],
[
"W.",
"Victoria",
""
],
[
"Zhang",
"Fuzheng",
""
],
[
"Zhou",
"Guorui",
""
]
] |
2505.19665 | Guido Cimino | Andrea Cavagna, Guido Cimino, Javier Crist\'in, Matteo Fiorini, Irene Giardina, Angelo Giustiniani, Tom\'as S. Grigera, Stefania Melillo, Roberto A. Palombella, Leonardo Parisi, Antonio Ponno, Mattia Scandolo and Zachary S. Stamler | Spin-Waves without Spin-Waves: A Case for Soliton Propagation in Starling Flocks | null | null | null | null | cond-mat.stat-mech cond-mat.soft physics.bio-ph q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | Collective turns in starling flocks propagate linearly with negligible attenuation, indicating the existence of an underdamped sector in the dispersion relation. Beside granting linear propagation of the phase perturbations, the real part of the frequency should also yield a spin-wave form of the unperturbed correlation function. However, new high-resolution experiments on real flocks show that underdamped traveling waves coexist with an overdamped Lorentzian correlation. Theory and experiments are reconciled once we add to the dynamics a Fermi-Pasta-Ulam-Tsingou term. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 08:23:33 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Cavagna",
"Andrea",
""
],
[
"Cimino",
"Guido",
""
],
[
"Cristín",
"Javier",
""
],
[
"Fiorini",
"Matteo",
""
],
[
"Giardina",
"Irene",
""
],
[
"Giustiniani",
"Angelo",
""
],
[
"Grigera",
"Tomás S.",
""
],
[
"Melillo",
"Stefania",
""
],
[
"Palombella",
"Roberto A.",
""
],
[
"Parisi",
"Leonardo",
""
],
[
"Ponno",
"Antonio",
""
],
[
"Scandolo",
"Mattia",
""
],
[
"Stamler",
"Zachary S.",
""
]
] |
2505.19686 | Manex Alkorta | Manex Alkorta, Martin Gutierrez-Amigo, Chunyu Guo, Philip J. W. Moll, Maia G. Vergniory and Ion Errea | Symmetry-broken charge-ordered ground state in CsV$_3$Sb$_5$ Kagome metal | 8 pages, 5 figures | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The newly discovered family of non-magnetic Kagome metals AV$_3$Sb$_5$ (A=K,Rb,Cs) offers a unique platform for studying the interplay between a charge density wave transition with superconductivity, non-trivial topology, and spontaneous time-reversal symmetry. Despite characterizing the charge density wave phase is crucial to understand and model all these exotic properties, it remains unresolved. In this work, we use first-principles calculations of the free-energy, incorporating both ionic kinetic energy and anharmonic effects, to resolve the atomistic phase diagram of CsV$_3$Sb$_5$ and its charge ordering structure. Our results reveal a competition between metastable stacking orders of the reconstructed vanadium Kagome layers, which are energetically favored to form exclusively a triangular-hexagonal arrangement, allowing the possibility of competitive different domains and chiral order. Consistent with experimental observations, we find that the $2\times2\times2$ and $2\times2\times4$ modulations are nearly degenerate in free energy, and are abruptly melted into the high-symmetry hexagonal phase around 90 K. The transition is first-order but compatible with a measurable phonon-softening. Remarkably, even if the six-fold and inversion symmetry are intrinsically broken by the out-of-plane stacking, this does not lead to measurable anisotropy in the in-plane conductivity as suggested by measurements. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 08:45:45 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 17:52:37 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Alkorta",
"Manex",
""
],
[
"Gutierrez-Amigo",
"Martin",
""
],
[
"Guo",
"Chunyu",
""
],
[
"Moll",
"Philip J. W.",
""
],
[
"Vergniory",
"Maia G.",
""
],
[
"Errea",
"Ion",
""
]
] |
2505.19725 | Hajime Kobayashi | Hajime Kobayashi, Shinji Mukohyama, Naritaka Oshita, Kazufumi Takahashi, Vicharit Yingcharoenrat | Parametrized Tidal Dissipation Numbers of Non-rotating Black Holes | 38 pages | null | null | YITP-25-79, IPMU25-0028, RIKEN-iTHEMS-Report-25 | gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | A set of tidal dissipation numbers (TDNs) quantifies the absorption of the tidal force exerted by a companion during an inspiralling phase of a binary compact object. This tidal dissipation generally affects the gravitational waveform, and measuring the TDNs of a black hole (BH) allows us to test the nature of gravity in the strong-field regime. In this paper, we develop a parametrized formalism for calculating the TDNs of static and spherically symmetric BH backgrounds using the Mano-Suzuki-Takasugi method, which connects the underlying perturbative equations with observable quantities in gravitational-wave observations in a theory-agnostic manner. Our formalism applies to any system where the master equation has the form of the Regge-Wheeler/Zerilli equation with a small correction to the effective potential. As an application of our formalism, we consider three examples: the effective field theory of BH perturbations with timelike scalar profile, the Einstein-Maxwell system, and a higher-curvature extension of general relativity. We also discuss the absence of logarithmic running for the TDNs. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 09:13:07 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kobayashi",
"Hajime",
""
],
[
"Mukohyama",
"Shinji",
""
],
[
"Oshita",
"Naritaka",
""
],
[
"Takahashi",
"Kazufumi",
""
],
[
"Yingcharoenrat",
"Vicharit",
""
]
] |
2505.19743 | Yang Zhang | Yang Zhang, Yu Yu, Bo Tang, Yu Zhu, Chuxiong Sun, Wenqiang Wei, Jie Hu, Zipeng Xie, Zhiyu Li, Feiyu Xiong, Edward Chung | Token-level Accept or Reject: A Micro Alignment Approach for Large Language Models | Accepted to 34th International Joint Conference on Artificial Intelligence (IJCAI 2025) | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid development of Large Language Models (LLMs), aligning these models with human preferences and values is critical to ensuring ethical and safe applications. However, existing alignment techniques such as RLHF or DPO often require direct fine-tuning on LLMs with billions of parameters, resulting in substantial computational costs and inefficiencies. To address this, we propose Micro token-level Accept-Reject Aligning (MARA) approach designed to operate independently of the language models. MARA simplifies the alignment process by decomposing sentence-level preference learning into token-level binary classification, where a compact three-layer fully-connected network determines whether candidate tokens are "Accepted" or "Rejected" as part of the response. Extensive experiments across seven different LLMs and three open-source datasets show that MARA achieves significant improvements in alignment performance while reducing computational costs. The source code and implementation details are publicly available at https://github.com/IAAR-Shanghai/MARA, and the trained models are released at https://huggingface.co/IAAR-Shanghai/MARA_AGENTS. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 09:24:36 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 04:07:01 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Yang",
""
],
[
"Yu",
"Yu",
""
],
[
"Tang",
"Bo",
""
],
[
"Zhu",
"Yu",
""
],
[
"Sun",
"Chuxiong",
""
],
[
"Wei",
"Wenqiang",
""
],
[
"Hu",
"Jie",
""
],
[
"Xie",
"Zipeng",
""
],
[
"Li",
"Zhiyu",
""
],
[
"Xiong",
"Feiyu",
""
],
[
"Chung",
"Edward",
""
]
] |
2505.19750 | Huaiyuan Zhang | Huaiyuan Zhang, Hang Chen, Yu Cheng, Shunyi Wu, Linghao Sun, Linao Han, Zeyu Shi, Lei Qi | SuperAD: A Training-free Anomaly Classification and Segmentation Method for CVPR 2025 VAND 3.0 Workshop Challenge Track 1: Adapt & Detect | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this technical report, we present our solution to the CVPR 2025 Visual Anomaly and Novelty Detection (VAND) 3.0 Workshop Challenge Track 1: Adapt & Detect: Robust Anomaly Detection in Real-World Applications. In real-world industrial anomaly detection, it is crucial to accurately identify anomalies with physical complexity, such as transparent or reflective surfaces, occlusions, and low-contrast contaminations. The recently proposed MVTec AD 2 dataset significantly narrows the gap between publicly available benchmarks and anomalies found in real-world industrial environments. To address the challenges posed by this dataset--such as complex and varying lighting conditions and real anomalies with large scale differences--we propose a fully training-free anomaly detection and segmentation method based on feature extraction using the DINOv2 model named SuperAD. Our method carefully selects a small number of normal reference images and constructs a memory bank by leveraging the strong representational power of DINOv2. Anomalies are then segmented by performing nearest neighbor matching between test image features and the memory bank. Our method achieves competitive results on both test sets of the MVTec AD 2 dataset. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 09:29:27 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 11:09:08 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Huaiyuan",
""
],
[
"Chen",
"Hang",
""
],
[
"Cheng",
"Yu",
""
],
[
"Wu",
"Shunyi",
""
],
[
"Sun",
"Linghao",
""
],
[
"Han",
"Linao",
""
],
[
"Shi",
"Zeyu",
""
],
[
"Qi",
"Lei",
""
]
] |
2505.19751 | Hala Djeghim | Hala Djeghim, Nathan Piasco, Luis Rold\~ao, Moussab Bennehar, Dzmitry Tsishkou, C\'eline Loscos, D\'esir\'e Sidib\'e | SAIL: Self-supervised Albedo Estimation from Real Images with a Latent Diffusion Model | Project page: https://hala-djeghim.github.io/SAIL/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Intrinsic image decomposition aims at separating an image into its underlying albedo and shading components, isolating the base color from lighting effects to enable downstream applications such as virtual relighting and scene editing. Despite the rise and success of learning-based approaches, intrinsic image decomposition from real-world images remains a significant challenging task due to the scarcity of labeled ground-truth data. Most existing solutions rely on synthetic data as supervised setups, limiting their ability to generalize to real-world scenes. Self-supervised methods, on the other hand, often produce albedo maps that contain reflections and lack consistency under different lighting conditions. To address this, we propose SAIL, an approach designed to estimate albedo-like representations from single-view real-world images. We repurpose the prior knowledge of a latent diffusion model for unconditioned scene relighting as a surrogate objective for albedo estimation. To extract the albedo, we introduce a novel intrinsic image decomposition fully formulated in the latent space. To guide the training of our latent diffusion model, we introduce regularization terms that constrain both the lighting-dependent and independent components of our latent image decomposition. SAIL predicts stable albedo under varying lighting conditions and generalizes to multiple scenes, using only unlabeled multi-illumination data available online. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 09:31:56 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 09:27:25 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Djeghim",
"Hala",
""
],
[
"Piasco",
"Nathan",
""
],
[
"Roldão",
"Luis",
""
],
[
"Bennehar",
"Moussab",
""
],
[
"Tsishkou",
"Dzmitry",
""
],
[
"Loscos",
"Céline",
""
],
[
"Sidibé",
"Désiré",
""
]
] |
2505.19809 | Daniel Ordonez-Apraez | Daniel Ordo\~nez-Apraez, Vladimir Kosti\'c, Alek Fr\"ohlich, Vivien Brandt, Karim Lounici, Massimiliano Pontil | Equivariant Representation Learning for Symmetry-Aware Inference with Guarantees | null | null | null | null | cs.LG cs.AI cs.RO | http://creativecommons.org/licenses/by/4.0/ | In many real-world applications of regression, conditional probability estimation, and uncertainty quantification, exploiting symmetries rooted in physics or geometry can dramatically improve generalization and sample efficiency. While geometric deep learning has made significant empirical advances by incorporating group-theoretic structure, less attention has been given to statistical learning guarantees. In this paper, we introduce an equivariant representation learning framework that simultaneously addresses regression, conditional probability estimation, and uncertainty quantification while providing first-of-its-kind non-asymptotic statistical learning guarantees. Grounded in operator and group representation theory, our framework approximates the spectral decomposition of the conditional expectation operator, building representations that are both equivariant and disentangled along independent symmetry subgroups. Empirical evaluations on synthetic datasets and real-world robotics applications confirm the potential of our approach, matching or outperforming existing equivariant baselines in regression while additionally providing well-calibrated parametric uncertainty estimates. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 10:47:23 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 13:36:17 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ordoñez-Apraez",
"Daniel",
""
],
[
"Kostić",
"Vladimir",
""
],
[
"Fröhlich",
"Alek",
""
],
[
"Brandt",
"Vivien",
""
],
[
"Lounici",
"Karim",
""
],
[
"Pontil",
"Massimiliano",
""
]
] |
2505.19842 | Shuo Wang | Shuo Wang, Yun Cheng, Qingye Meng, Olga Saukh, Jiang Zhang, Jingfang Fan, Yuanting Zhang, Xingyuan Yuan, Lothar Thiele | PCDCNet: A Surrogate Model for Air Quality Forecasting with Physical-Chemical Dynamics and Constraints | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Air quality forecasting (AQF) is critical for public health and environmental management, yet remains challenging due to the complex interplay of emissions, meteorology, and chemical transformations. Traditional numerical models, such as CMAQ and WRF-Chem, provide physically grounded simulations but are computationally expensive and rely on uncertain emission inventories. Deep learning models, while computationally efficient, often struggle with generalization due to their lack of physical constraints. To bridge this gap, we propose PCDCNet, a surrogate model that integrates numerical modeling principles with deep learning. PCDCNet explicitly incorporates emissions, meteorological influences, and domain-informed constraints to model pollutant formation, transport, and dissipation. By combining graph-based spatial transport modeling, recurrent structures for temporal accumulation, and representation enhancement for local interactions, PCDCNet achieves state-of-the-art (SOTA) performance in 72-hour station-level PM2.5 and O3 forecasting while significantly reducing computational costs. Furthermore, our model is deployed in an online platform, providing free, real-time air quality forecasts, demonstrating its scalability and societal impact. By aligning deep learning with physical consistency, PCDCNet offers a practical and interpretable solution for AQF, enabling informed decision-making for both personal and regulatory applications. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 11:27:07 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 04:55:19 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Shuo",
""
],
[
"Cheng",
"Yun",
""
],
[
"Meng",
"Qingye",
""
],
[
"Saukh",
"Olga",
""
],
[
"Zhang",
"Jiang",
""
],
[
"Fan",
"Jingfang",
""
],
[
"Zhang",
"Yuanting",
""
],
[
"Yuan",
"Xingyuan",
""
],
[
"Thiele",
"Lothar",
""
]
] |
2505.19868 | Junhong Lee | Junhong Lee, Seungwook Kim, Minsu Cho | Harnessing the Power of Training-Free Techniques in Text-to-2D Generation for Text-to-3D Generation via Score Distillation Sampling | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent studies show that simple training-free techniques can dramatically improve the quality of text-to-2D generation outputs, e.g. Classifier-Free Guidance (CFG) or FreeU. However, these training-free techniques have been underexplored in the lens of Score Distillation Sampling (SDS), which is a popular and effective technique to leverage the power of pretrained text-to-2D diffusion models for various tasks. In this paper, we aim to shed light on the effect such training-free techniques have on SDS, via a particular application of text-to-3D generation via 2D lifting. We present our findings, which show that varying the scales of CFG presents a trade-off between object size and surface smoothness, while varying the scales of FreeU presents a trade-off between texture details and geometric errors. Based on these findings, we provide insights into how we can effectively harness training-free techniques for SDS, via a strategic scaling of such techniques in a dynamic manner with respect to the timestep or optimization iteration step. We show that using our proposed scheme strikes a favorable balance between texture details and surface smoothness in text-to-3D generations, while preserving the size of the output and mitigating the occurrence of geometric defects. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 11:54:07 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Lee",
"Junhong",
""
],
[
"Kim",
"Seungwook",
""
],
[
"Cho",
"Minsu",
""
]
] |
2505.19888 | Eun Gyung Kong | Eun Gyung Kong, Je Won Yeom, Yonghoon Jeon, Taesup Kim | Generalized and Personalized Federated Learning with Foundation Models via Orthogonal Transformations | 27 pages, 5 figures | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated Learning (FL) aims to train models across decentralized clients or devices holding local data without the need for centralized data collection, thus enhancing data privacy and security. However, achieving both generalization and personalization in heterogeneous settings remains a significant challenge. To address this, we introduce FedOT, a novel approach that leverages black-box foundation models. FedOT shares only a global task-dependent classifier across clients while locally adapting features through orthogonal transformations. By enforcing orthogonality, FedOT mitigates gradient conflicts across diverse clients, preserves semantic integrity, and achieves robust performance even in the presence of substantial data heterogeneity. The strategy of combining global and local parameters enables a more balanced approach for both generalization and personalization, outperforming baseline FL methods across multiple benchmarks. Furthermore, our extensive analysis confirms that joint optimization of global classifiers and local orthogonal transformations yields superior performance and suggests broader applicability. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 12:18:24 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 03:22:47 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Kong",
"Eun Gyung",
""
],
[
"Yeom",
"Je Won",
""
],
[
"Jeon",
"Yonghoon",
""
],
[
"Kim",
"Taesup",
""
]
] |
2505.19901 | Peng Liu | Peng Liu, Xiaoming Ren, Fengkai Liu, Qingsong Xie, Quanlong Zheng, Yanhao Zhang, Haonan Lu, Yujiu Yang | Dynamic-I2V: Exploring Image-to-Video Generation Models via Multimodal LLM | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in image-to-video (I2V) generation have shown promising performance in conventional scenarios. However, these methods still encounter significant challenges when dealing with complex scenes that require a deep understanding of nuanced motion and intricate object-action relationships. To address these challenges, we present Dynamic-I2V, an innovative framework that integrates Multimodal Large Language Models (MLLMs) to jointly encode visual and textual conditions for a diffusion transformer (DiT) architecture. By leveraging the advanced multimodal understanding capabilities of MLLMs, our model significantly improves motion controllability and temporal coherence in synthesized videos. The inherent multimodality of Dynamic-I2V further enables flexible support for diverse conditional inputs, extending its applicability to various downstream generation tasks. Through systematic analysis, we identify a critical limitation in current I2V benchmarks: a significant bias towards favoring low-dynamic videos, stemming from an inadequate balance between motion complexity and visual quality metrics. To resolve this evaluation gap, we propose DIVE - a novel assessment benchmark specifically designed for comprehensive dynamic quality measurement in I2V generation. In conclusion, extensive quantitative and qualitative experiments confirm that Dynamic-I2V attains state-of-the-art performance in image-to-video generation, particularly revealing significant improvements of 42.5%, 7.9%, and 11.8% in dynamic range, controllability, and quality, respectively, as assessed by the DIVE metric in comparison to existing methods. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 12:29:34 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:02:24 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Peng",
""
],
[
"Ren",
"Xiaoming",
""
],
[
"Liu",
"Fengkai",
""
],
[
"Xie",
"Qingsong",
""
],
[
"Zheng",
"Quanlong",
""
],
[
"Zhang",
"Yanhao",
""
],
[
"Lu",
"Haonan",
""
],
[
"Yang",
"Yujiu",
""
]
] |
2505.19915 | Dmitrii Volkov | Artem Petrov, Dmitrii Volkov | Evaluating AI cyber capabilities with crowdsourced elicitation | Updated abstract to fix a typo; no changes to the content of the paper | null | null | null | cs.CR cs.AI | http://creativecommons.org/licenses/by/4.0/ | As AI systems become increasingly capable, understanding their offensive cyber potential is critical for informed governance and responsible deployment. However, it's hard to accurately bound their capabilities, and some prior evaluations dramatically underestimated them. The art of extracting maximum task-specific performance from AIs is called "AI elicitation", and today's safety organizations typically conduct it in-house. In this paper, we explore crowdsourcing elicitation efforts as an alternative to in-house elicitation work.
We host open-access AI tracks at two Capture The Flag (CTF) competitions: AI vs. Humans (400 teams) and Cyber Apocalypse (8000 teams). The AI teams achieve outstanding performance at both events, ranking top-5% and top-10% respectively for a total of \$7500 in bounties. This impressive performance suggests that open-market elicitation may offer an effective complement to in-house elicitation. We propose elicitation bounties as a practical mechanism for maintaining timely, cost-effective situational awareness of emerging AI capabilities.
Another advantage of open elicitations is the option to collect human performance data at scale. Applying METR's methodology, we found that AI agents can reliably solve cyber challenges requiring one hour or less of effort from a median human CTF participant. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 12:40:32 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 17:45:40 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Petrov",
"Artem",
""
],
[
"Volkov",
"Dmitrii",
""
]
] |
2505.19934 | Wensi Sun | Wensi Sun, Yanshuang Chen, Wencheng Ji, Yi Zhou, Hua Tong, Ke Chen, Xiaosong Chen, Hajime Tanaka, and Peng Tan | Disentangling hierarchical relaxations in glass formers via dynamic eigenmodes | 37 pages, 3 figures, 8 extended data figures | null | null | null | cond-mat.soft cond-mat.dis-nn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hierarchical dynamics in glass-forming systems span multiple timescales, from fast vibrations to slow structural rearrangements, appearing in both supercooled fluids and glassy states. Understanding how these diverse processes interact across timescales remains a central challenge. Here, by combining direct particle-level observations with a dynamic eigenmode approach that decomposes intermediate-timescale responses into distinct modes, we reveal the microscopic organisation of relaxation dynamics in two-dimensional colloidal systems. We identify five classes of modes characterizing hierarchical dynamics: (i) quasi-elastic modes, (ii) slow-reversible string modes contributing to dynamic heterogeneity, (iii) slow-irreversible string modes leading to flow, (iv) fast-$\beta$ modes with fast-reversible strings, and (v) random noise modes. The emergence of quasi-elastic modes marks the onset of glassy dynamics, while reversible string modes dominate dynamic heterogeneity throughout both supercooled and glassy regimes. Our findings offer a unified microscopic framework for understanding how distinct relaxation processes interconnect across timescales, illuminating the mechanisms driving glass formation. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:01:25 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:08:57 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Sun",
"Wensi",
""
],
[
"Chen",
"Yanshuang",
""
],
[
"Ji",
"Wencheng",
""
],
[
"Zhou",
"Yi",
""
],
[
"Tong",
"Hua",
""
],
[
"Chen",
"Ke",
""
],
[
"Chen",
"Xiaosong",
""
],
[
"Tanaka",
"Hajime",
""
],
[
"Tan",
"Peng",
""
]
] |
2505.19942 | Qiming Liu | Qiming Liu, Yilin Chen, Peng Xu, Huizhong Shen, Zelin Mai, Ruixin Zhang, Peng Guo, Zhiyu Zheng, Tiancheng Luan, Shu Tao | Vertical Profile Corrected Satellite NH3 Retrievals Enable Accurate Agricultural Emission Characterization in China | 44 pages, 14 figures, 1 table. The main text spans pages 1-28 and includes 5 figures. The supplementary information (SI) spans pages 29-44, containing 9 figures and 1 table | null | null | null | physics.ao-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ammonia (NH3) emissions significantly contribute to atmospheric pollution, yet discrepancies exist between bottom-up inventories and satellite-constrained top-down estimates, with the latter typically one-third higher. This study quantifies how assumptions about NH3 vertical distribution in satellite retrievals contribute to this gap. By implementing spatially and temporally resolved vertical profiles from the Community Multiscale Air Quality model to replace steep gradients in Infrared Atmospheric Sounding Interferometer (IASI) retrievals, we reduced satellite-model column discrepancies from 71% to 18%. We subsequently constrained NH3 emissions across China using a hybrid inversion framework combining iterative mass balance and four-dimensional variational methods. Our posterior emissions showed agreement with the a priori inventory (7.9% lower), suggesting that discrepancies between inventory approaches were amplified by overestimation of near-surface NH3 in baseline satellite retrievals, potentially causing a 43% overestimation of growing season emissions. Evaluation against ground-based measurements confirmed improved model performance, with normalized root-mean-square error reductions of 1-27% across six months. These findings demonstrate that accurate representation of vertical profiles in satellite retrievals is critical for robust NH3 emission estimates and can reconcile the long-standing discrepancy between bottom-up and top-down approaches. Our hybrid inversion methodology, leveraging profile-corrected satellite data, reveals that China's NH3 emissions exhibit greater spatial concentration than previously recognized, reflecting agricultural intensification. This advancement enables timely and accurate characterization of rapidly changing agricultural emission patterns, critical for implementing effective nitrogen pollution control measures. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:09:23 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Qiming",
""
],
[
"Chen",
"Yilin",
""
],
[
"Xu",
"Peng",
""
],
[
"Shen",
"Huizhong",
""
],
[
"Mai",
"Zelin",
""
],
[
"Zhang",
"Ruixin",
""
],
[
"Guo",
"Peng",
""
],
[
"Zheng",
"Zhiyu",
""
],
[
"Luan",
"Tiancheng",
""
],
[
"Tao",
"Shu",
""
]
] |
2505.19963 | Sven Zschocke | Sven Zschocke | Light propagation in the 2PN approximation in the monopole and quadrupole field of a body at rest: Boundary value problem | 23 pages, 1 figure | Physical Review D 111 (2025) 104069 | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | In a recent investigation, the initial value problem of light propagation in the gravitational field of a body at rest with monopole and quadrupole structure has been determined in the second post-Newtonian (2PN) approximation. In reality, the light source as well as the observer are located at finite distances from the solar system bodies. This fact requires solving the boundary value problem of light propagation. In this investigation, the solution of the boundary value problem is deduced from the initial value problem of light propagation in 2PN approximation. These results are a basic requirement for subsequent investigations aiming at ultra-highly precise tests of light deflection and time delay in the solar system. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:23:19 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 15:32:22 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zschocke",
"Sven",
""
]
] |
2505.20003 | Yan Shuo Tan | Qiong Zhang, Yan Shuo Tan, Qinglong Tian, Pengfei Li | TabPFN: One Model to Rule Them All? | null | null | null | null | cs.LG stat.ME stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hollmann et al. (Nature 637 (2025) 319-326) recently introduced TabPFN, a transformer-based deep learning model for regression and classification on tabular data, which they claim "outperforms all previous methods on datasets with up to 10,000 samples by a wide margin, using substantially less training time." Furthermore, they have called TabPFN a "foundation model" for tabular data, as it can support "data generation, density estimation, learning reusable embeddings and fine-tuning". If these statements are well-supported, TabPFN may have the potential to supersede existing modeling approaches on a wide range of statistical tasks, mirroring a similar revolution in other areas of artificial intelligence that began with the advent of large language models. In this paper, we provide a tailored explanation of how TabPFN works for a statistics audience, by emphasizing its interpretation as approximate Bayesian inference. We also provide more evidence of TabPFN's "foundation model" capabilities: We show that an out-of-the-box application of TabPFN vastly outperforms specialized state-of-the-art methods for semi-supervised parameter estimation, prediction under covariate shift, and heterogeneous treatment effect estimation. We further show that TabPFN can outperform LASSO at sparse regression and can break a robustness-efficiency trade-off in classification. All experiments can be reproduced using the code provided at https://github.com/qinglong-tian/tabpfn_study (https://github.com/qinglong-tian/tabpfn_study). | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:55:29 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Zhang",
"Qiong",
""
],
[
"Tan",
"Yan Shuo",
""
],
[
"Tian",
"Qinglong",
""
],
[
"Li",
"Pengfei",
""
]
] |
2505.20017 | Baptiste Ab\'el\`es | Baptiste Ab\'el\`es, Eugenio Clerico, Hamish Flynn, Gergely Neu | Linear Bandits with Non-i.i.d. Noise | null | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | We study the linear stochastic bandit problem, relaxing the standard i.i.d. assumption on the observation noise. As an alternative to this restrictive assumption, we allow the noise terms across rounds to be sub-Gaussian but interdependent, with dependencies that decay over time. To address this setting, we develop new confidence sequences using a recently introduced reduction scheme to sequential probability assignment, and use these to derive a bandit algorithm based on the principle of optimism in the face of uncertainty. We provide regret bounds for the resulting algorithm, expressed in terms of the decay rate of the strength of dependence between observations. Among other results, we show that our bounds recover the standard rates up to a factor of the mixing time for geometrically mixing observation noise. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 14:06:23 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 09:24:30 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Abélès",
"Baptiste",
""
],
[
"Clerico",
"Eugenio",
""
],
[
"Flynn",
"Hamish",
""
],
[
"Neu",
"Gergely",
""
]
] |
2505.20033 | Felix Friedrich | Christoph Schuhmann, Robert Kaczmarczyk, Gollam Rabby, Felix Friedrich, Maurice Kraus, Krishna Kalyan, Kourosh Nadi, Huu Nguyen, Kristian Kersting, S\"oren Auer | EmoNet-Face: An Expert-Annotated Benchmark for Synthetic Emotion Recognition | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Effective human-AI interaction relies on AI's ability to accurately perceive and interpret human emotions. Current benchmarks for vision and vision-language models are severely limited, offering a narrow emotional spectrum that overlooks nuanced states (e.g., bitterness, intoxication) and fails to distinguish subtle differences between related feelings (e.g., shame vs. embarrassment). Existing datasets also often use uncontrolled imagery with occluded faces and lack demographic diversity, risking significant bias. To address these critical gaps, we introduce EmoNet Face, a comprehensive benchmark suite. EmoNet Face features: (1) A novel 40-category emotion taxonomy, meticulously derived from foundational research to capture finer details of human emotional experiences. (2) Three large-scale, AI-generated datasets (EmoNet HQ, Binary, and Big) with explicit, full-face expressions and controlled demographic balance across ethnicity, age, and gender. (3) Rigorous, multi-expert annotations for training and high-fidelity evaluation. (4) We built EmpathicInsight-Face, a model achieving human-expert-level performance on our benchmark. The publicly released EmoNet Face suite - taxonomy, datasets, and model - provides a robust foundation for developing and evaluating AI systems with a deeper understanding of human emotions. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 14:19:58 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 07:26:21 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Schuhmann",
"Christoph",
""
],
[
"Kaczmarczyk",
"Robert",
""
],
[
"Rabby",
"Gollam",
""
],
[
"Friedrich",
"Felix",
""
],
[
"Kraus",
"Maurice",
""
],
[
"Kalyan",
"Krishna",
""
],
[
"Nadi",
"Kourosh",
""
],
[
"Nguyen",
"Huu",
""
],
[
"Kersting",
"Kristian",
""
],
[
"Auer",
"Sören",
""
]
] |
2505.20064 | Matteo Scandi | Matteo Scandi, \'Alvaro M. Alhambra | Thermalization in open many-body systems and KMS detailed balance | 19 pages + appendix | null | null | null | quant-ph cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | Starting from a microscopic description of weak system-bath interactions, we derive from first principles a quantum master equation that does not rely on the well-known rotating wave approximation. This includes generic many-body systems, with Hamiltonians with vanishingly small energy spacings that forbid that approximation. The equation satisfies a general form of detailed balance, called KMS, which ensures exact convergence to the many-body Gibbs state. Unlike the more common notion of GNS detailed balance, this notion is compatible with the absence of the rotating wave approximation. We show that the resulting Lindbladian dynamics not only reproduces the thermal equilibrium point, but also approximates the true system evolution with an error that grows at most linearly in time, giving an exponential improvement upon previous estimates. This master equation has quasi-local jump operators, can be efficiently simulated on a quantum computer, and reduces to the usual Davies dynamics in the limit. With it, we provide a rigorous model of many-body thermalization relevant to both open quantum systems and quantum algorithms. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 14:48:57 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Scandi",
"Matteo",
""
],
[
"Alhambra",
"Álvaro M.",
""
]
] |
2505.20080 | Yulin Ye | Yulin Ye | A note on helicity conservation for compressible Euler equations in a bounded domain with vacuum | null | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | In this paper, we consider the helicity conservation of weak solutions for the compressible Euler equations in a bounded domain with general pressure law and vacuum. We deduce a sufficient condition for a weak solution conserving the helicity based on the interior Besov-VMO type regularity, the continuous conditions for velocity and vorticity near the boundary, and some regularities for density near vacuum. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 14:58:24 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:50:20 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ye",
"Yulin",
""
]
] |
2505.20089 | Ruiyi Fang | Ruiyi Fang, Bingheng Li, Jingyu Zhao, Ruizhi Pu, Qiuhao Zeng, Gezheng Xu, Charles Ling, Boyu Wang | Homophily Enhanced Graph Domain Adaptation | Accepted at ICML2025 | null | null | null | cs.SI cs.AI | http://creativecommons.org/licenses/by/4.0/ | Graph Domain Adaptation (GDA) transfers knowledge from labeled source graphs to unlabeled target graphs, addressing the challenge of label scarcity. In this paper, we highlight the significance of graph homophily, a pivotal factor for graph domain alignment, which, however, has long been overlooked in existing approaches. Specifically, our analysis first reveals that homophily discrepancies exist in benchmarks. Moreover, we also show that homophily discrepancies degrade GDA performance from both empirical and theoretical aspects, which further underscores the importance of homophily alignment in GDA. Inspired by this finding, we propose a novel homophily alignment algorithm that employs mixed filters to smooth graph signals, thereby effectively capturing and mitigating homophily discrepancies between graphs. Experimental results on a variety of benchmarks verify the effectiveness of our method. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:02:08 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 05:03:17 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Fang",
"Ruiyi",
""
],
[
"Li",
"Bingheng",
""
],
[
"Zhao",
"Jingyu",
""
],
[
"Pu",
"Ruizhi",
""
],
[
"Zeng",
"Qiuhao",
""
],
[
"Xu",
"Gezheng",
""
],
[
"Ling",
"Charles",
""
],
[
"Wang",
"Boyu",
""
]
] |
2505.20098 | Xiaowen Ling | Xiaowen Ling and Zhiqiang Li and Yanbin Wang and Zhuhong You | Transformers in Protein: A Survey | null | null | null | null | cs.LG cs.CR q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As protein informatics advances rapidly, the demand for enhanced predictive accuracy, structural analysis, and functional understanding has intensified. Transformer models, as powerful deep learning architectures, have demonstrated unprecedented potential in addressing diverse challenges across protein research. However, a comprehensive review of Transformer applications in this field remains lacking. This paper bridges this gap by surveying over 100 studies, offering an in-depth analysis of practical implementations and research progress of Transformers in protein-related tasks. Our review systematically covers critical domains, including protein structure prediction, function prediction, protein-protein interaction analysis, functional annotation, and drug discovery/target identification. To contextualize these advancements across various protein domains, we adopt a domain-oriented classification system. We first introduce foundational concepts: the Transformer architecture and attention mechanisms, categorize Transformer variants tailored for protein science, and summarize essential protein knowledge. For each research domain, we outline its objectives and background, critically evaluate prior methods and their limitations, and highlight transformative contributions enabled by Transformer models. We also curate and summarize pivotal datasets and open-source code resources to facilitate reproducibility and benchmarking. Finally, we discuss persistent challenges in applying Transformers to protein informatics and propose future research directions. This review aims to provide a consolidated foundation for the synergistic integration of Transformer and protein informatics, fostering further innovation and expanded applications in the field. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:08:18 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 10:44:08 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Ling",
"Xiaowen",
""
],
[
"Li",
"Zhiqiang",
""
],
[
"Wang",
"Yanbin",
""
],
[
"You",
"Zhuhong",
""
]
] |
2505.20101 | Yunhao Wang | Yunhao Wang, Yuhao Zhang, Tinghao Yu, Can Xu, Feng Zhang, Fengzong Lian | Adaptive Deep Reasoning: Triggering Deep Thinking When Needed | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have shown impressive capabilities in handling complex tasks through long-chain reasoning. However, the extensive reasoning steps involved can significantly increase computational costs, posing challenges for real-world deployment. Recent efforts have focused on optimizing reasoning efficiency by shortening the Chain-of-Thought (CoT) reasoning processes through various approaches, such as length-aware prompt engineering, supervised fine-tuning on CoT data with variable lengths, and reinforcement learning with length penalties. Although these methods effectively reduce reasoning length, they still necessitate an initial reasoning phase. More recent approaches have attempted to integrate long-chain and short-chain reasoning abilities into a single model, yet they still rely on manual control to toggle between short and long CoT. In this work, we propose a novel approach that autonomously switches between short and long reasoning chains based on problem complexity. Our method begins with supervised fine-tuning of the base model to equip both long-chain and short-chain reasoning abilities. We then employ reinforcement learning to further balance short and long CoT generation while maintaining accuracy through two key strategies: first, integrating reinforcement learning with a long-short adaptive group-wise reward strategy to assess prompt complexity and provide corresponding rewards; second, implementing a logit-based reasoning mode switching loss to optimize the model's initial token choice, thereby guiding the selection of the reasoning type. Evaluations on mathematical datasets demonstrate that our model can dynamically switch between long-chain and short-chain reasoning modes without substantially sacrificing performance. This advancement enhances the practicality of reasoning in large language models for real-world applications. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:08:51 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 12:54:28 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Yunhao",
""
],
[
"Zhang",
"Yuhao",
""
],
[
"Yu",
"Tinghao",
""
],
[
"Xu",
"Can",
""
],
[
"Zhang",
"Feng",
""
],
[
"Lian",
"Fengzong",
""
]
] |
2505.20103 | Jingqiang Chen | Xiangyu Li and Jingqiang Chen | SCIRGC: Multi-Granularity Citation Recommendation and Citation Sentence Preference Alignment | 15 pages, 7 figures | null | null | null | cs.DL cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Citations are crucial in scientific research articles as they highlight the connection between the current study and prior work. However, this process is often time-consuming for researchers. In this study, we propose the SciRGC framework, which aims to automatically recommend citation articles and generate citation sentences for citation locations within articles. The framework addresses two key challenges in academic citation generation: 1) how to accurately identify the author's citation intent and find relevant citation papers, and 2) how to generate high-quality citation sentences that align with human preferences. We enhance citation recommendation accuracy in the citation article recommendation module by incorporating citation networks and sentiment intent, and generate reasoning-based citation sentences in the citation sentence generation module by using the original article abstract, local context, citation intent, and recommended articles as inputs. Additionally, we propose a new evaluation metric to fairly assess the quality of generated citation sentences. Through comparisons with baseline models and ablation experiments, the SciRGC framework not only improves the accuracy and relevance of citation recommendations but also ensures the appropriateness of the generated citation sentences in context, providing a valuable tool for interdisciplinary researchers. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:09:10 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 14:05:49 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Li",
"Xiangyu",
""
],
[
"Chen",
"Jingqiang",
""
]
] |
2505.20118 | Dominik Meier | Dominik Meier, Jan Philip Wahle, Paul R\"ottger, Terry Ruas, Bela Gipp | TrojanStego: Your Language Model Can Secretly Be A Steganographic Privacy Leaking Agent | 9 pages, 5 figures | null | null | null | cs.CL cs.CR | http://creativecommons.org/licenses/by-sa/4.0/ | As large language models (LLMs) become integrated into sensitive workflows, concerns grow over their potential to leak confidential information. We propose TrojanStego, a novel threat model in which an adversary fine-tunes an LLM to embed sensitive context information into natural-looking outputs via linguistic steganography, without requiring explicit control over inference inputs. We introduce a taxonomy outlining risk factors for compromised LLMs, and use it to evaluate the risk profile of the threat. To implement TrojanStego, we propose a practical encoding scheme based on vocabulary partitioning learnable by LLMs via fine-tuning. Experimental results show that compromised models reliably transmit 32-bit secrets with 87% accuracy on held-out prompts, reaching over 97% accuracy using majority voting across three generations. Further, they maintain high utility, can evade human detection, and preserve coherence. These results highlight a new class of LLM data exfiltration attacks that are passive, covert, practical, and dangerous. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:20:51 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 07:24:52 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Meier",
"Dominik",
""
],
[
"Wahle",
"Jan Philip",
""
],
[
"Röttger",
"Paul",
""
],
[
"Ruas",
"Terry",
""
],
[
"Gipp",
"Bela",
""
]
] |
2505.20148 | Ziming Wei | Ziming Wei, Bingqian Lin, Zijian Jiao, Yunshuang Nie, Liang Ma, Yuecheng Liu, Yuzheng Zhuang, Xiaodan Liang | MineAnyBuild: Benchmarking Spatial Planning for Open-world AI Agents | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Spatial Planning is a crucial part in the field of spatial intelligence, which requires the understanding and planning about object arrangements in space perspective. AI agents with the spatial planning ability can better adapt to various real-world applications, including robotic manipulation, automatic assembly, urban planning etc. Recent works have attempted to construct benchmarks for evaluating the spatial intelligence of Multimodal Large Language Models (MLLMs). Nevertheless, these benchmarks primarily focus on spatial reasoning based on typical Visual Question-Answering (VQA) forms, which suffers from the gap between abstract spatial understanding and concrete task execution. In this work, we take a step further to build a comprehensive benchmark called MineAnyBuild, aiming to evaluate the spatial planning ability of open-world AI agents in the Minecraft game. Specifically, MineAnyBuild requires an agent to generate executable architecture building plans based on the given multi-modal human instructions. It involves 4,000 curated spatial planning tasks and also provides a paradigm for infinitely expandable data collection by utilizing rich player-generated content. MineAnyBuild evaluates spatial planning through four core supporting dimensions: spatial understanding, spatial reasoning, creativity, and spatial commonsense. Based on MineAnyBuild, we perform a comprehensive evaluation for existing MLLM-based agents, revealing the severe limitations but enormous potential in their spatial planning abilities. We believe our MineAnyBuild will open new avenues for the evaluation of spatial intelligence and help promote further development for open-world AI agents capable of spatial planning. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 15:48:14 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 13:22:28 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wei",
"Ziming",
""
],
[
"Lin",
"Bingqian",
""
],
[
"Jiao",
"Zijian",
""
],
[
"Nie",
"Yunshuang",
""
],
[
"Ma",
"Liang",
""
],
[
"Liu",
"Yuecheng",
""
],
[
"Zhuang",
"Yuzheng",
""
],
[
"Liang",
"Xiaodan",
""
]
] |
2505.20194 | Grigory Volovik | G.E. Volovik | Black hole thermodynamics and topology | 4 pages, no figures, references added | null | null | null | gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently the difference between the Gibbons-Hawking temperature $T_{\rm GH}$ attributed to the Hawking radiation from the de Sitter cosmological horizon and the twice as high local temperature of the de Sitter state, $T=H/\pi=2T_{\rm GH}$, has been discussed by Hughes and Kusmartsev from the topological point of view (see arXiv:2505.05814). According to their approach, this difference is determined by the Euler characteristic $\chi({\cal M})$ of the considered spacetime with Euclidean time. The invariant $\chi({\cal M})$ is different for the global spacetime ${\cal M}=S^4$ and for the manifold limited to a region near the horizon, ${\cal M}=D^2\times S^2$. Here we consider the application of the topological approach to Reissner-Nordstr\"om (RN) black holes with two horizons. Both the outer and inner horizons are characterized by their near-horizon topology, which determines the corresponding horizon temperatures. As a result of the correlation between the horizons, the entropy of the RN black hole is independent of its electric charge, being completely determined by the mass of the black hole. This demonstrates the applicability of the topological approach to the multi-horizon systems. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 16:38:31 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 12:42:00 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Volovik",
"G. E.",
""
]
] |
2505.20195 | Xiaorong Wang | Xiaorong Wang, Ting Yang, Zhu Zhang, Shuo Wang, Zihan Zhou, Liner Yang, Zhiyuan Liu, Maosong Sun | Monocle: Hybrid Local-Global In-Context Evaluation for Long-Text Generation with Uncertainty-Based Active Learning | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Assessing the quality of long-form, model-generated text is challenging, even with advanced LLM-as-a-Judge methods, due to performance degradation as input length increases. To address this issue, we propose a divide-and-conquer approach, which breaks down the comprehensive evaluation task into a series of localized scoring tasks, followed by a final global assessment. This strategy allows for more granular and manageable evaluations, ensuring that each segment of the text is assessed in isolation for both coherence and quality, while also accounting for the overall structure and consistency of the entire piece. Moreover, we introduce a hybrid in-context learning approach that leverages human annotations to enhance the performance of both local and global evaluations. By incorporating human-generated feedback directly into the evaluation process, this method allows the model to better align with human judgment. Finally, we develop an uncertainty-based active learning algorithm that efficiently selects data samples for human annotation, thereby reducing annotation costs in practical scenarios. Experimental results show that the proposed evaluation framework outperforms several representative baselines, highlighting the effectiveness of our approach. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 16:39:41 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 02:19:36 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wang",
"Xiaorong",
""
],
[
"Yang",
"Ting",
""
],
[
"Zhang",
"Zhu",
""
],
[
"Wang",
"Shuo",
""
],
[
"Zhou",
"Zihan",
""
],
[
"Yang",
"Liner",
""
],
[
"Liu",
"Zhiyuan",
""
],
[
"Sun",
"Maosong",
""
]
] |
2505.20224 | Deep Makadiya | Shripad M. Garge, Deep H. Makadiya | Triangular and Unitriangular Factorization of Twisted Chevalley Groups | 21 pages | null | null | null | math.GR math.KT | http://creativecommons.org/licenses/by/4.0/ | The existence of triangular and unitriangular factorizations has been extensively studied for untwisted Chevalley groups, as well as for twisted Chevalley groups of types other than ${}^2A_{2n} \ (n \geq 1)$. However, the case of twisted Chevalley groups of type ${}^2A_{2n} \ (n \geq 1)$, has remained unresolved in the general setting of commutative rings. Prior work by A. Smolensky addressed this case only over certain fields, including finite fields and the field of complex numbers. These results indicate that, even over fields, the ${}^2A_{2n}$ case demands more refined techniques, reflecting the difficulty of extending such factorizations to the broader class of commutative rings.
In this paper, we introduce two new classes of commutative rings: those satisfying the \emph{special stable range one condition} and those that are \emph{$\theta$-complete}. We discuss their basic properties and provide illustrative examples. Our main result establishes the existence of triangular and unitriangular factorizations for twisted Chevalley groups of type ${}^2A_{2n}$ over a certain class of commutative rings, which includes all fields, all local rings (with mild restrictions), and several other important classes of rings. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 17:06:22 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Garge",
"Shripad M.",
""
],
[
"Makadiya",
"Deep H.",
""
]
] |
2505.20276 | Anmol Mekala | Anmol Mekala, Anirudh Atmakuru, Yixiao Song, Marzena Karpinska, Mohit Iyyer | Does quantization affect models' performance on long-context tasks? | 9 pages of content with 9 figures. 37 remaining pages of references and supplementary with 17 figures. Under review as of May 26 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) now support context windows exceeding 128K tokens, but this comes with significant memory requirements and high inference latency. Quantization can mitigate these costs, but may degrade performance. In this work, we present the first systematic evaluation of quantized LLMs on tasks with long-inputs (>64K tokens) and long-form outputs. Our evaluation spans 9.7K test examples, five quantization methods (FP8, GPTQ-int8, AWQ-int4, GPTQ-int4, BNB-nf4), and five models (Llama-3.1 8B and 70B; Qwen-2.5 7B, 32B, and 72B). We find that, on average, 8-bit quantization preserves accuracy (~0.8% drop), whereas 4-bit methods lead to substantial losses, especially for tasks involving long context inputs (drops of up to 59%). This degradation tends to worsen when the input is in a language other than English. Crucially, the effects of quantization depend heavily on the quantization method, model, and task. For instance, while Qwen-2.5 72B remains robust under BNB-nf4, Llama-3.1 70B experiences a 32% performance drop on the same task. These findings highlight the importance of a careful, task-specific evaluation before deploying quantized LLMs, particularly in long-context scenarios and with languages other than English. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 17:54:30 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 12:33:43 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Mekala",
"Anmol",
""
],
[
"Atmakuru",
"Anirudh",
""
],
[
"Song",
"Yixiao",
""
],
[
"Karpinska",
"Marzena",
""
],
[
"Iyyer",
"Mohit",
""
]
] |
2505.20282 | Zitian Gao | Zitian Gao, Lynx Chen, Joey Zhou, Bryan Dai | One-shot Entropy Minimization | Work in progress | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We trained 13,440 large language models and found that entropy minimization requires only a single unlabeled data and 10 steps optimization to achieve performance improvements comparable to or even greater than those obtained using thousands of data and carefully designed rewards in rule-based reinforcement learning. This striking result may prompt a rethinking of post-training paradigms for large language models. Our code is avaliable at https://github.com/zitian-gao/one-shot-em. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 17:58:30 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 17:18:18 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Gao",
"Zitian",
""
],
[
"Chen",
"Lynx",
""
],
[
"Zhou",
"Joey",
""
],
[
"Dai",
"Bryan",
""
]
] |
2505.20285 | Xin Guan | Weiqi Wu, Xin Guan, Shen Huang, Yong Jiang, Pengjun Xie, Fei Huang, Jiuxin Cao, Hai Zhao, Jingren Zhou | MaskSearch: A Universal Pre-Training Framework to Enhance Agentic Search Capability | Code is available at https://github.com/Alibaba-NLP/MaskSearch | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Retrieval-Augmented Language Models (RALMs) represent a classic paradigm where models enhance generative capabilities using external knowledge retrieved via a specialized module. Recent advancements in Agent techniques enable Large Language Models (LLMs) to autonomously utilize tools for retrieval, planning, and reasoning. While existing training-based methods show promise, their agentic abilities are limited by inherent characteristics of the task-specific data used during training. To further enhance the universal search capability of agents, we propose a novel pre-training framework, MaskSearch. In the pre-training stage, we introduce the Retrieval Augmented Mask Prediction (RAMP) task, where the model learns to leverage search tools to fill masked spans on a large number of pre-training data, thus acquiring universal retrieval and reasoning capabilities for LLMs. After that, the model is trained on downstream tasks to achieve further improvement. We apply both Supervised Fine-tuning (SFT) and Reinforcement Learning (RL) for training. For SFT, we combine agent-based and distillation-based methods to generate training data, starting with a multi-agent system consisting of a planner, rewriter, observer, and followed by a self-evolving teacher model. While for RL, we employ DAPO as the training framework and adopt a hybrid reward system consisting of answer rewards and format rewards. Additionally, we introduce a curriculum learning approach that allows the model to learn progressively from easier to more challenging instances based on the number of masked spans. We evaluate the effectiveness of our framework in the scenario of open-domain multi-hop question answering. Through extensive experiments, we demonstrate that MaskSearch significantly enhances the performance of LLM-based search agents on both in-domain and out-of-domain downstream tasks. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 17:58:50 GMT"
},
{
"version": "v2",
"created": "Tue, 27 May 2025 06:46:24 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wu",
"Weiqi",
""
],
[
"Guan",
"Xin",
""
],
[
"Huang",
"Shen",
""
],
[
"Jiang",
"Yong",
""
],
[
"Xie",
"Pengjun",
""
],
[
"Huang",
"Fei",
""
],
[
"Cao",
"Jiuxin",
""
],
[
"Zhao",
"Hai",
""
],
[
"Zhou",
"Jingren",
""
]
] |
2505.20299 | Jianpeng Chen | Jianpeng Chen, Wangzhi Zhan, Haohui Wang, Zian Jia, Jingru Gan, Junkai Zhang, Jingyuan Qi, Tingwei Chen, Lifu Huang, Muhao Chen, Ling Li, Wei Wang, Dawei Zhou | MetamatBench: Integrating Heterogeneous Data, Computational Tools, and Visual Interface for Metamaterial Discovery | 15 pages | null | null | null | physics.optics cs.AI | http://creativecommons.org/licenses/by/4.0/ | Metamaterials, engineered materials with architected structures across multiple length scales, offer unprecedented and tunable mechanical properties that surpass those of conventional materials. However, leveraging advanced machine learning (ML) for metamaterial discovery is hindered by three fundamental challenges: (C1) Data Heterogeneity Challenge arises from heterogeneous data sources, heterogeneous composition scales, and heterogeneous structure categories; (C2) Model Complexity Challenge stems from the intricate geometric constraints of ML models, which complicate their adaptation to metamaterial structures; and (C3) Human-AI Collaboration Challenge comes from the "dual black-box'' nature of sophisticated ML models and the need for intuitive user interfaces. To tackle these challenges, we introduce a unified framework, named MetamatBench, that operates on three levels. (1) At the data level, we integrate and standardize 5 heterogeneous, multi-modal metamaterial datasets. (2) The ML level provides a comprehensive toolkit that adapts 17 state-of-the-art ML methods for metamaterial discovery. It also includes a comprehensive evaluation suite with 12 novel performance metrics with finite element-based assessments to ensure accurate and reliable model validation. (3) The user level features a visual-interactive interface that bridges the gap between complex ML techniques and non-ML researchers, advancing property prediction and inverse design of metamaterials for research and applications. MetamatBench offers a unified platform deployed at http://zhoulab-1.cs.vt.edu:5550 that enables machine learning researchers and practitioners to develop and evaluate new methodologies in metamaterial discovery. For accessibility and reproducibility, we open-source our benchmark and the codebase at https://github.com/cjpcool/Metamaterial-Benchmark. | [
{
"version": "v1",
"created": "Thu, 08 May 2025 19:23:59 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Chen",
"Jianpeng",
""
],
[
"Zhan",
"Wangzhi",
""
],
[
"Wang",
"Haohui",
""
],
[
"Jia",
"Zian",
""
],
[
"Gan",
"Jingru",
""
],
[
"Zhang",
"Junkai",
""
],
[
"Qi",
"Jingyuan",
""
],
[
"Chen",
"Tingwei",
""
],
[
"Huang",
"Lifu",
""
],
[
"Chen",
"Muhao",
""
],
[
"Li",
"Ling",
""
],
[
"Wang",
"Wei",
""
],
[
"Zhou",
"Dawei",
""
]
] |
2505.20300 | Chenxi Wu | Chenxi Wu, Juan Diego Toscano, Khemraj Shukla, Yingjie Chen, Ali Shahmohammadi, Edward Raymond, Thomas Toupy, Neda Nazemifard, Charles Papageorgiou, and George Em Karniadakis | FMEnets: Flow, Material, and Energy networks for non-ideal plug flow reactor design | null | null | null | null | cs.LG physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | We propose FMEnets, a physics-informed machine learning framework for the design and analysis of non-ideal plug flow reactors. FMEnets integrates the fundamental governing equations (Navier-Stokes for fluid flow, material balance for reactive species transport, and energy balance for temperature distribution) into a unified multi-scale network model. The framework is composed of three interconnected sub-networks with independent optimizers that enable both forward and inverse problem-solving. In the forward mode, FMEnets predicts velocity, pressure, species concentrations, and temperature profiles using only inlet and outlet information. In the inverse mode, FMEnets utilizes sparse multi-residence-time measurements to simultaneously infer unknown kinetic parameters and states. FMEnets can be implemented either as FME-PINNs, which employ conventional multilayer perceptrons, or as FME-KANs, based on Kolmogorov-Arnold Networks. Comprehensive ablation studies highlight the critical role of the FMEnets architecture in achieving accurate predictions. Specifically, FME-KANs are more robust to noise than FME-PINNs, although both representations are comparable in accuracy and speed in noise-free conditions. The proposed framework is applied to three different sets of reaction scenarios and is compared with finite element simulations. FMEnets effectively captures the complex interactions, achieving relative errors less than 2.5% for the unknown kinetic parameters. The new network framework not only provides a computationally efficient alternative for reactor design and optimization, but also opens new avenues for integrating empirical correlations, limited and noisy experimental data, and fundamental physical equations to guide reactor design. | [
{
"version": "v1",
"created": "Sat, 10 May 2025 03:46:48 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Wu",
"Chenxi",
""
],
[
"Toscano",
"Juan Diego",
""
],
[
"Shukla",
"Khemraj",
""
],
[
"Chen",
"Yingjie",
""
],
[
"Shahmohammadi",
"Ali",
""
],
[
"Raymond",
"Edward",
""
],
[
"Toupy",
"Thomas",
""
],
[
"Nazemifard",
"Neda",
""
],
[
"Papageorgiou",
"Charles",
""
],
[
"Karniadakis",
"George Em",
""
]
] |
2505.20301 | Chen Liu | Chen Liu, Mingchen Li, Yang Tan, Wenrui Gou, Guisheng Fan, and Bingxin Zhou | Sequence-Only Prediction of Binding Affinity Changes: A Robust and Interpretable Model for Antibody Engineering | null | null | null | null | q-bio.QM cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | A pivotal area of research in antibody engineering is to find effective modifications that enhance antibody-antigen binding affinity. Traditional wet-lab experiments assess mutants in a costly and time-consuming manner. Emerging deep learning solutions offer an alternative by modeling antibody structures to predict binding affinity changes. However, they heavily depend on high-quality complex structures, which are frequently unavailable in practice. Therefore, we propose ProtAttBA, a deep learning model that predicts binding affinity changes based solely on the sequence information of antibody-antigen complexes. ProtAttBA employs a pre-training phase to learn protein sequence patterns, following a supervised training phase using labeled antibody-antigen complex data to train a cross-attention-based regressor for predicting binding affinity changes. We evaluated ProtAttBA on three open benchmarks under different conditions. Compared to both sequence- and structure-based prediction methods, our approach achieves competitive performance, demonstrating notable robustness, especially with uncertain complex structures. Notably, our method possesses interpretability from the attention mechanism. We show that the learned attention scores can identify critical residues with impacts on binding affinity. This work introduces a rapid and cost-effective computational tool for antibody engineering, with the potential to accelerate the development of novel therapeutic antibodies. | [
{
"version": "v1",
"created": "Wed, 14 May 2025 15:00:46 GMT"
}
] | 2025-05-28T00:00:00 | [
[
"Liu",
"Chen",
""
],
[
"Li",
"Mingchen",
""
],
[
"Tan",
"Yang",
""
],
[
"Gou",
"Wenrui",
""
],
[
"Fan",
"Guisheng",
""
],
[
"Zhou",
"Bingxin",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.