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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2505.22974 | Yuntao Ma | Yuntao Ma, Andrei Cramariuc, Farbod Farshidian, Marco Hutter | Learning coordinated badminton skills for legged manipulators | Science Robotics DOI: 10.1126/scirobotics.adu3922 | Sci. Robot.10,eadu3922(2025) | 10.1126/scirobotics.adu3922 | null | cs.RO cs.LG | http://creativecommons.org/licenses/by/4.0/ | Coordinating the motion between lower and upper limbs and aligning limb control with perception are substantial challenges in robotics, particularly in dynamic environments. To this end, we introduce an approach for enabling legged mobile manipulators to play badminton, a task that requires precise coordination of perception, locomotion, and arm swinging. We propose a unified reinforcement learning-based control policy for whole-body visuomotor skills involving all degrees of freedom to achieve effective shuttlecock tracking and striking. This policy is informed by a perception noise model that utilizes real-world camera data, allowing for consistent perception error levels between simulation and deployment and encouraging learned active perception behaviors. Our method includes a shuttlecock prediction model, constrained reinforcement learning for robust motion control, and integrated system identification techniques to enhance deployment readiness. Extensive experimental results in a variety of environments validate the robot's capability to predict shuttlecock trajectories, navigate the service area effectively, and execute precise strikes against human players, demonstrating the feasibility of using legged mobile manipulators in complex and dynamic sports scenarios. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:26:30 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Ma",
"Yuntao",
""
],
[
"Cramariuc",
"Andrei",
""
],
[
"Farshidian",
"Farbod",
""
],
[
"Hutter",
"Marco",
""
]
] |
2505.22975 | Piotr Haj{\l}asz | Pawe{\l} Goldstein and Piotr Haj{\l}asz | $\mathbf{C^2}$-Lusin approximation of convex functions: one variable case | null | null | null | null | math.CA | http://creativecommons.org/licenses/by/4.0/ | We prove that if $f:(a,b)\to\mathbb{R}$ is convex, then for any $\varepsilon>0$ there is a convex function $g\in C^2(a,b)$ such that $|\{f\neq g\}|<\varepsilon$ and $\Vert f-g\Vert_\infty<\varepsilon$. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:28:28 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Goldstein",
"Paweł",
""
],
[
"Hajłasz",
"Piotr",
""
]
] |
2505.22976 | Kevin Lian | Kewei Lian, Shaofei Cai, Yilun Du, Yitao Liang | Toward Memory-Aided World Models: Benchmarking via Spatial Consistency | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | The ability to simulate the world in a spatially consistent manner is a crucial requirements for effective world models. Such a model enables high-quality visual generation, and also ensures the reliability of world models for downstream tasks such as simulation and planning. Designing a memory module is a crucial component for addressing spatial consistency: such a model must not only retain long-horizon observational information, but also enables the construction of explicit or implicit internal spatial representations. However, there are no dataset designed to promote the development of memory modules by explicitly enforcing spatial consistency constraints. Furthermore, most existing benchmarks primarily emphasize visual coherence or generation quality, neglecting the requirement of long-range spatial consistency. To bridge this gap, we construct a dataset and corresponding benchmark by sampling 150 distinct locations within the open-world environment of Minecraft, collecting about 250 hours (20 million frames) of loop-based navigation videos with actions. Our dataset follows a curriculum design of sequence lengths, allowing models to learn spatial consistency on increasingly complex navigation trajectories. Furthermore, our data collection pipeline is easily extensible to new Minecraft environments and modules. Four representative world model baselines are evaluated on our benchmark. Dataset, benchmark, and code are open-sourced to support future research. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:28:57 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Lian",
"Kewei",
""
],
[
"Cai",
"Shaofei",
""
],
[
"Du",
"Yilun",
""
],
[
"Liang",
"Yitao",
""
]
] |
2505.22977 | Shuolin Xu | Shuolin Xu, Siming Zheng, Ziyi Wang, HC Yu, Jinwei Chen, Huaqi Zhang, Bo Li, Peng-Tao Jiang | HyperMotion: DiT-Based Pose-Guided Human Image Animation of Complex Motions | 17 pages, 7 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in diffusion models have significantly improved conditional video generation, particularly in the pose-guided human image animation task. Although existing methods are capable of generating high-fidelity and time-consistent animation sequences in regular motions and static scenes, there are still obvious limitations when facing complex human body motions (Hypermotion) that contain highly dynamic, non-standard motions, and the lack of a high-quality benchmark for evaluation of complex human motion animations. To address this challenge, we introduce the \textbf{Open-HyperMotionX Dataset} and \textbf{HyperMotionX Bench}, which provide high-quality human pose annotations and curated video clips for evaluating and improving pose-guided human image animation models under complex human motion conditions. Furthermore, we propose a simple yet powerful DiT-based video generation baseline and design spatial low-frequency enhanced RoPE, a novel module that selectively enhances low-frequency spatial feature modeling by introducing learnable frequency scaling. Our method significantly improves structural stability and appearance consistency in highly dynamic human motion sequences. Extensive experiments demonstrate the effectiveness of our dataset and proposed approach in advancing the generation quality of complex human motion image animations. Code and dataset will be made publicly available. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:30:46 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Xu",
"Shuolin",
""
],
[
"Zheng",
"Siming",
""
],
[
"Wang",
"Ziyi",
""
],
[
"Yu",
"HC",
""
],
[
"Chen",
"Jinwei",
""
],
[
"Zhang",
"Huaqi",
""
],
[
"Li",
"Bo",
""
],
[
"Jiang",
"Peng-Tao",
""
]
] |
2505.22978 | Youngju Na | Youngju Na, Taeyeon Kim, Jumin Lee, Kyu Beom Han, Woo Jae Kim, Sung-eui Yoon | Pose-free 3D Gaussian splatting via shape-ray estimation | ICIP 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | While generalizable 3D Gaussian splatting enables efficient, high-quality rendering of unseen scenes, it heavily depends on precise camera poses for accurate geometry. In real-world scenarios, obtaining accurate poses is challenging, leading to noisy pose estimates and geometric misalignments. To address this, we introduce SHARE, a pose-free, feed-forward Gaussian splatting framework that overcomes these ambiguities by joint shape and camera rays estimation. Instead of relying on explicit 3D transformations, SHARE builds a pose-aware canonical volume representation that seamlessly integrates multi-view information, reducing misalignment caused by inaccurate pose estimates. Additionally, anchor-aligned Gaussian prediction enhances scene reconstruction by refining local geometry around coarse anchors, allowing for more precise Gaussian placement. Extensive experiments on diverse real-world datasets show that our method achieves robust performance in pose-free generalizable Gaussian splatting. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:34:40 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Na",
"Youngju",
""
],
[
"Kim",
"Taeyeon",
""
],
[
"Lee",
"Jumin",
""
],
[
"Han",
"Kyu Beom",
""
],
[
"Kim",
"Woo Jae",
""
],
[
"Yoon",
"Sung-eui",
""
]
] |
2505.22979 | Bengisu G\"uresti | Bengisu Guresti, Chongjie Zhang, Yevgeniy Vorobeychik | Learning Recommender Mechanisms for Bayesian Stochastic Games | null | null | null | null | cs.GT cs.MA | http://creativecommons.org/licenses/by/4.0/ | An important challenge in non-cooperative game theory is coordinating on a single (approximate) equilibrium from many possibilities - a challenge that becomes even more complex when players hold private information. Recommender mechanisms tackle this problem by recommending strategies to players based on their reported type profiles. A key consideration in such mechanisms is to ensure that players are incentivized to participate, report their private information truthfully, and follow the recommendations. While previous work has focused on designing recommender mechanisms for one-shot and extensive-form games, these approaches cannot be effectively applied to stochastic games, particularly if we constrain recommendations to be Markov stationary policies. To bridge this gap, we introduce a novel bi-level reinforcement learning approach for automatically designing recommender mechanisms in Bayesian stochastic games. Our method produces a mechanism represented by a parametric function (such as a neural network), and is therefore highly efficient at execution time. Experimental results on two repeated and two stochastic games demonstrate that our approach achieves social welfare levels competitive with cooperative multi-agent reinforcement learning baselines, while also providing significantly improved incentive properties. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:34:54 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Guresti",
"Bengisu",
""
],
[
"Zhang",
"Chongjie",
""
],
[
"Vorobeychik",
"Yevgeniy",
""
]
] |
2505.22980 | Aimon Rahman | Aimon Rahman, Jiang Liu, Ze Wang, Ximeng Sun, Jialian Wu, Xiaodong Yu, Yusheng Su, Vishal M. Patel, Zicheng Liu, Emad Barsoum | MOVi: Training-free Text-conditioned Multi-Object Video Generation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advances in diffusion-based text-to-video (T2V) models have demonstrated remarkable progress, but these models still face challenges in generating videos with multiple objects. Most models struggle with accurately capturing complex object interactions, often treating some objects as static background elements and limiting their movement. In addition, they often fail to generate multiple distinct objects as specified in the prompt, resulting in incorrect generations or mixed features across objects. In this paper, we present a novel training-free approach for multi-object video generation that leverages the open world knowledge of diffusion models and large language models (LLMs). We use an LLM as the ``director'' of object trajectories, and apply the trajectories through noise re-initialization to achieve precise control of realistic movements. We further refine the generation process by manipulating the attention mechanism to better capture object-specific features and motion patterns, and prevent cross-object feature interference. Extensive experiments validate the effectiveness of our training free approach in significantly enhancing the multi-object generation capabilities of existing video diffusion models, resulting in 42% absolute improvement in motion dynamics and object generation accuracy, while also maintaining high fidelity and motion smoothness. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:41:10 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Rahman",
"Aimon",
""
],
[
"Liu",
"Jiang",
""
],
[
"Wang",
"Ze",
""
],
[
"Sun",
"Ximeng",
""
],
[
"Wu",
"Jialian",
""
],
[
"Yu",
"Xiaodong",
""
],
[
"Su",
"Yusheng",
""
],
[
"Patel",
"Vishal M.",
""
],
[
"Liu",
"Zicheng",
""
],
[
"Barsoum",
"Emad",
""
]
] |
2505.22981 | Siyang Liu | Siyang Liu, Sahand Sabour, Xiaoyang Wang, Rada Mihalcea | Free Lunch for User Experience: Crowdsourcing Agents for Scalable User Studies | null | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We demonstrate the potential of anthropomorphized language agents to generate budget-friendly, moderate-fidelity, yet sufficiently insightful user experiences at scale, supporting fast, early-stage prototyping. We explore this through the case of prototyping Large Language Model-driven non-player characters (NPCs). We present Agentic H-CI, a framework that mirrors traditional user research processes-surveying, screening, experiencing, and collecting feedback and insights-with simulated agents. Using this approach, we easily construct a team of 240 player agents with a balanced range of player types and personality traits, at extremely low cost (\$0.28/player) and minimal time commitment (6.9 minutes/player). Content analysis shows that agent-based players behave in ways aligned with their simulated backgrounds, achieving 82.5\% alignment with designated profiles. From their interactions, we distill 11 user insights and 6 design implications to guide further development. To evaluate practical value, we conduct parallel user studies with human participants recruited locally and via crowdsourcing. Ratings from three professional game developers show that the agentic player team offers a Pareto-optimal and well-balanced trade-off across fidelity, cost, time efficiency, and insight helpfulness. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:44:27 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Liu",
"Siyang",
""
],
[
"Sabour",
"Sahand",
""
],
[
"Wang",
"Xiaoyang",
""
],
[
"Mihalcea",
"Rada",
""
]
] |
2505.22982 | Ralph Hoch | Christoph Luckeneder, Ralph Hoch and Hermann Kaindl | Structural Abstraction and Selective Refinement for Formal Verification | 10 pages | null | null | null | cs.RO cs.SE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Safety verification of robot applications is extremely challenging due to the complexity of the environment that a robot typically operates in. Formal verification with model-checking provides guarantees but it may often take too long or even fail for complex models of the environment. A usual solution approach is abstraction, more precisely behavioral abstraction. Our new approach introduces structural abstraction instead, which we investigated in the context of voxel representation of the robot environment. This kind of abstraction leads to abstract voxels. We also propose a complete and automated verification workflow, which is based on an already existing methodology for robot applications, and inspired by the key ideas behind counterexample-guided abstraction refinement (CEGAR) - performing an initial abstraction and successively introducing refinements based on counterexamples, intertwined with model-checker runs. Hence, our approach uses selective refinement of structural abstractions to improve the runtime efficiency of model-checking. A fully-automated implementation of our approach showed its feasibility, since counterexamples have been found for a realistic scenario with a fairly high (maximal) resolution in a few minutes, while direct model-checker runs led to a crash after a couple of days. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:44:47 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Luckeneder",
"Christoph",
""
],
[
"Hoch",
"Ralph",
""
],
[
"Kaindl",
"Hermann",
""
]
] |
2505.22983 | Bhanuka Manesha Samarasekara Vitharana Gamage | Bhanuka Gamage, Leona Holloway, Nicola McDowell, Thanh-Toan Do, Nicholas Price, Arthur Lowery, and Kim Marriott | Vision-Based Assistive Technologies for People with Cerebral Visual Impairment: A Review and Focus Study | Author's accepted version of a paper published at ASSETS 2024 (October, 2024). arXiv admin note: text overlap with arXiv:2505.21875 | null | 10.1145/3663548.3675637 | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Over the past decade, considerable research has investigated Vision-Based Assistive Technologies (VBAT) to support people with vision impairments to understand and interact with their immediate environment using machine learning, computer vision, image enhancement, and/or augmented/virtual reality. However, this has almost totally overlooked a growing demographic: people with Cerebral Visual Impairment (CVI). Unlike ocular vision impairments, CVI arises from damage to the brain's visual processing centres. Through a scoping review, this paper reveals a significant research gap in addressing the needs of this demographic. Three focus studies involving 7 participants with CVI explored the challenges, current strategies, and opportunities for VBAT. We also discussed the assistive technology needs of people with CVI compared with ocular low vision. Our findings highlight the opportunity for the Human-Computer Interaction and Assistive Technologies research community to explore and address this underrepresented domain, thereby enhancing the quality of life for people with CVI. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:45:14 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Gamage",
"Bhanuka",
""
],
[
"Holloway",
"Leona",
""
],
[
"McDowell",
"Nicola",
""
],
[
"Do",
"Thanh-Toan",
""
],
[
"Price",
"Nicholas",
""
],
[
"Lowery",
"Arthur",
""
],
[
"Marriott",
"Kim",
""
]
] |
2505.22984 | Donghui Yan | Guancheng Zhou and Haiping Xu and Hongkang Xu and Chenyu Li and Donghui Yan | A Computational Approach to Improving Fairness in K-means Clustering | 14 pages, 5 figures | null | null | null | cs.LG cs.CY | http://creativecommons.org/licenses/by/4.0/ | The popular K-means clustering algorithm potentially suffers from a major weakness for further analysis or interpretation. Some cluster may have disproportionately more (or fewer) points from one of the subpopulations in terms of some sensitive variable, e.g., gender or race. Such a fairness issue may cause bias and unexpected social consequences. This work attempts to improve the fairness of K-means clustering with a two-stage optimization formulation--clustering first and then adjust cluster membership of a small subset of selected data points. Two computationally efficient algorithms are proposed in identifying those data points that are expensive for fairness, with one focusing on nearest data points outside of a cluster and the other on highly 'mixed' data points. Experiments on benchmark datasets show substantial improvement on fairness with a minimal impact to clustering quality. The proposed algorithms can be easily extended to a broad class of clustering algorithms or fairness metrics. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:48:12 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Zhou",
"Guancheng",
""
],
[
"Xu",
"Haiping",
""
],
[
"Xu",
"Hongkang",
""
],
[
"Li",
"Chenyu",
""
],
[
"Yan",
"Donghui",
""
]
] |
2505.22985 | Tsuyoshi Okita | Masaharu Kagiyama and Tsuyoshi Okita | Knowledge Distillation for Reservoir-based Classifier: Human Activity Recognition | 23 pages,5 figures | null | null | null | cs.LG cs.AI cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper aims to develop an energy-efficient classifier for time-series data by introducing PatchEchoClassifier, a novel model that leverages a reservoir-based mechanism known as the Echo State Network (ESN). The model is designed for human activity recognition (HAR) using one-dimensional sensor signals and incorporates a tokenizer to extract patch-level representations. To train the model efficiently, we propose a knowledge distillation framework that transfers knowledge from a high-capacity MLP-Mixer teacher to the lightweight reservoir-based student model. Experimental evaluations on multiple HAR datasets demonstrate that our model achieves over 80 percent accuracy while significantly reducing computational cost. Notably, PatchEchoClassifier requires only about one-sixth of the floating point operations (FLOPS) compared to DeepConvLSTM, a widely used convolutional baseline. These results suggest that PatchEchoClassifier is a promising solution for real-time and energy-efficient human activity recognition in edge computing environments. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:48:36 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Kagiyama",
"Masaharu",
""
],
[
"Okita",
"Tsuyoshi",
""
]
] |
2505.22986 | Seungjun Ahn | Seungjun Ahn, Eun Jeong Oh | A Network-Guided Penalized Regression with Application to Proteomics Data | 18 pages, 2 figures | null | null | null | stat.ME q-bio.QM stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Network theory has proven invaluable in unraveling complex protein interactions. Previous studies have employed statistical methods rooted in network theory, including the Gaussian graphical model, to infer networks among proteins, identifying hub proteins based on key structural properties of networks such as degree centrality. However, there has been limited research examining a prognostic role of hub proteins on outcomes, while adjusting for clinical covariates in the context of high-dimensional data. To address this gap, we propose a network-guided penalized regression method. First, we construct a network using the Gaussian graphical model to identify hub proteins. Next, we preserve these identified hub proteins along with clinically relevant factors, while applying adaptive Lasso to non-hub proteins for variable selection. Our network-guided estimators are shown to have variable selection consistency and asymptotic normality. Simulation results suggest that our method produces better results compared to existing methods and demonstrates promise for advancing biomarker identification in proteomics research. Lastly, we apply our method to the Clinical Proteomic Tumor Analysis Consortium (CPTAC) data and identified hub proteins that may serve as prognostic biomarkers for various diseases, including rare genetic disorders and immune checkpoint for cancer immunotherapy. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:50:17 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Ahn",
"Seungjun",
""
],
[
"Oh",
"Eun Jeong",
""
]
] |
2505.22987 | Nick Byrd | Nick Byrd | Strategic Reflectivism In Intelligent Systems | An earlier version of this paper was presented at the 2025 ACM Workshop on Human-AI Interaction for Augmented Reasoning (CHI25-WS-AUGMENTED-REASONING). Permission to copy for educational use is granted, provided copies are not for sale or profit and include this notice and full citation on the first page. Other uses require the author permission | null | null | CHI25-WS-AUGMENTED-REASONING | cs.AI cs.HC econ.TH | http://creativecommons.org/licenses/by-nc-nd/4.0/ | By late 20th century, the rationality wars had launched debates about the nature and norms of intuitive and reflective thinking. Those debates drew from mid-20th century ideas such as bounded rationality, which challenged more idealized notions of rationality observed since the 19th century. Now that 21st century cognitive scientists are applying the resulting dual process theories to artificial intelligence, it is time to dust off some lessons from this history. So this paper synthesizes old ideas with recent results from experiments on humans and machines. The result is Strategic Reflectivism, which takes the position that one key to intelligent systems (human or artificial) is pragmatic switching between intuitive and reflective inference to optimally fulfill competing goals. Strategic Reflectivism builds on American Pragmatism, transcends superficial indicators of reflective thinking such as model size or chains of thought, and becomes increasingly actionable as we learn more about the value of intuition and reflection. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:51:20 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Byrd",
"Nick",
""
]
] |
2505.22988 | Albert Tseng | Albert Tseng and Zhaofeng Sun and Christopher De Sa | Model-Preserving Adaptive Rounding | Preprint | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | The main goal of post-training quantization (PTQ) is to produced a compressed model whose output distribution is as close to the original model's as possible. To do this tractably, almost all LLM PTQ algorithms quantize linear layers by independently minimizing the immediate activation error. However, this localized objective ignores the effect of subsequent layers, so reducing it does not necessarily give a closer model. In this work, we introduce Yet Another Quantization Algorithm (YAQA), an adaptive rounding algorithm that uses Kronecker-factored approximations of each linear layer's Hessian with respect to the \textit{full model} KL divergence. YAQA consists of two components: Kronecker-factored sketches of the full layerwise Hessian that can be tractably computed for hundred-billion parameter LLMs, and a quantizer-independent rounding algorithm that uses these sketches and comes with theoretical guarantees. Across a wide range of models and quantizers, YAQA empirically reduces the KL divergence to the original model by $\approx 30\%$ while achieving state of the art performance on downstream tasks. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:53:00 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Tseng",
"Albert",
""
],
[
"Sun",
"Zhaofeng",
""
],
[
"De Sa",
"Christopher",
""
]
] |
2505.22989 | Brian Seong | Brian Seong, Paul Gebheim | Chainless Apps: A Modular Framework for Building Apps with Web2 Capability and Web3 Trust | 10 pages, 5 figures | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | Modern blockchain applications are often constrained by a trade-off between user experience and trust. Chainless Apps present a new paradigm of application architecture that separates execution, trust, bridging, and settlement into distinct compostable layers. This enables app-specific sequencing, verifiable off-chain computation, chain-agnostic asset and message routing via Agglayer, and finality on Ethereum - resulting in fast Web2-like UX with Web3-grade verifiability. Although consensus mechanisms have historically underpinned verifiable computation, the advent of zkVMs and decentralized validation services opens up new trust models for developers. Chainless Apps leverage this evolution to offer modular, scalable applications that maintain interoperability with the broader blockchain ecosystem while allowing domain-specific trade-offs. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:55:17 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Seong",
"Brian",
""
],
[
"Gebheim",
"Paul",
""
]
] |
2505.22990 | PinHan Chen | Pin-Han Chen, Yu-Sheng Lin, Wei-Cheng Lee, Tin-Yu Leu, Po-Hsiang Hsu, Anjana Dissanayake, Sungjin Oh, Chinq-Shiun Chiu | MenTeR: A fully-automated Multi-agenT workflow for end-to-end RF/Analog Circuits Netlist Design | 9 pages, 7 figures, accepted by IEEE ICLAD 2025 | null | null | null | cs.AI cs.ET cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | RF/Analog design is essential for bridging digital technologies with real-world signals, ensuring the functionality and reliability of a wide range of electronic systems. However, analog design procedures are often intricate, time-consuming and reliant on expert intuition, and hinder the time and cost efficiency of circuit development. To overcome the limitations of the manual circuit design, we introduce MenTeR - a multiagent workflow integrated into an end-to-end analog design framework. By employing multiple specialized AI agents that collaboratively address different aspects of the design process, such as specification understanding, circuit optimization, and test bench validation, MenTeR reduces the dependency on frequent trial-and-error-style intervention. MenTeR not only accelerates the design cycle time but also facilitates a broader exploration of the design space, demonstrating robust capabilities in handling real-world analog systems. We believe that MenTeR lays the groundwork for future "RF/Analog Copilots" that can collaborate seamlessly with human designers. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:58:08 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Chen",
"Pin-Han",
""
],
[
"Lin",
"Yu-Sheng",
""
],
[
"Lee",
"Wei-Cheng",
""
],
[
"Leu",
"Tin-Yu",
""
],
[
"Hsu",
"Po-Hsiang",
""
],
[
"Dissanayake",
"Anjana",
""
],
[
"Oh",
"Sungjin",
""
],
[
"Chiu",
"Chinq-Shiun",
""
]
] |
2505.22991 | Behzad Kamgar-Parsi | Behzad Kamgar-Parsi and Behrooz Kamgar-Parsi | Number of Clusters in a Dataset: A Regularized K-means Approach | 19 pages, 14 figures. arXiv admin note: substantial text overlap with arXiv:1911.06741 | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Finding the number of meaningful clusters in an unlabeled dataset is important in many applications. Regularized k-means algorithm is a possible approach frequently used to find the correct number of distinct clusters in datasets. The most common formulation of the regularization function is the additive linear term $\lambda k$, where $k$ is the number of clusters and $\lambda$ a positive coefficient. Currently, there are no principled guidelines for setting a value for the critical hyperparameter $\lambda$. In this paper, we derive rigorous bounds for $\lambda$ assuming clusters are {\em ideal}. Ideal clusters (defined as $d$-dimensional spheres with identical radii) are close proxies for k-means clusters ($d$-dimensional spherically symmetric distributions with identical standard deviations). Experiments show that the k-means algorithm with additive regularizer often yields multiple solutions. Thus, we also analyze k-means algorithm with multiplicative regularizer. The consensus among k-means solutions with additive and multiplicative regularizations reduces the ambiguity of multiple solutions in certain cases. We also present selected experiments that demonstrate performance of the regularized k-means algorithms as clusters deviate from the ideal assumption. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 01:58:44 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Kamgar-Parsi",
"Behzad",
""
],
[
"Kamgar-Parsi",
"Behrooz",
""
]
] |
2505.22992 | Anirban Dutta | Anirban Dutta, Andrew Fullard, Wolfgang Kerzendorf, J. T. O'Brien, Cecelia Powers, Stuart A Sim, Andreas Fl\"ors, Or Graur | Effect of positronium on the $\gamma$-ray spectra and energy deposition in Type Ia supernovae | Submitted to the Astrophysical Journal. Comments welcome! | null | null | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Type Ia supernovae (SNe Ia) are powered by the radioactive decay of isotopes such as $^{56}$Ni and $^{56}$Co, making their $\gamma$-ray spectra useful probes of the explosion mechanism and ejecta structure. Accurate interpretation of $\gamma$-ray observables, including line ratios and continuum fluxes, requires a detailed understanding of the microphysical processes that shape the spectra. One such process is positronium formation during electron-positron annihilation, which can redistribute flux from the 511 keV line into the surrounding continuum. To assess the impact of positronium on the emergent spectra, we developed a new open-source module TARDIS-HE, for time-dependent three-dimensional $\gamma$-ray transport, integrated into the radiative transfer code TARDIS. The code simulates $\gamma$-ray spectra and light curves from one-dimensional supernova ejecta models and allows for flexible incorporation of decay chains and opacity treatments. Using TARDIS-HE, we explore the effect of positronium formation by varying the positronium fraction from 0 % to 100 %, and assuming an extreme case where 75 % of positronium decays result in three-photon emission. We find that full positronium formation can reduce the 511 keV line flux by approximately 70 % and modestly enhance energy deposition by up to 2 % at around 100 days post-explosion, compared to models without positronium. These results demonstrate that while the effect is not dominant, positronium formation introduces measurable changes to $\gamma$-ray observables. Future observations with missions such as the Compton Spectrometer and Imager (COSI) may offer constraints on positronium formation in SNe Ia and help refine models of their radioactive energy transport. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:02:14 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Dutta",
"Anirban",
""
],
[
"Fullard",
"Andrew",
""
],
[
"Kerzendorf",
"Wolfgang",
""
],
[
"O'Brien",
"J. T.",
""
],
[
"Powers",
"Cecelia",
""
],
[
"Sim",
"Stuart A",
""
],
[
"Flörs",
"Andreas",
""
],
[
"Graur",
"Or",
""
]
] |
2505.22993 | Hoang Pham | Hoang Pham, Thanh-Do Nguyen, Khac-Hoai Nam Bui | Verify-in-the-Graph: Entity Disambiguation Enhancement for Complex Claim Verification with Interactive Graph Representation | Published at NAACL 2025 Main Conference | https://aclanthology.org/2025.naacl-long.268 | null | null | cs.CL cs.AI cs.DB cs.IR | http://creativecommons.org/licenses/by/4.0/ | Claim verification is a long-standing and challenging task that demands not only high accuracy but also explainability of the verification process. This task becomes an emerging research issue in the era of large language models (LLMs) since real-world claims are often complex, featuring intricate semantic structures or obfuscated entities. Traditional approaches typically address this by decomposing claims into sub-claims and querying a knowledge base to resolve hidden or ambiguous entities. However, the absence of effective disambiguation strategies for these entities can compromise the entire verification process. To address these challenges, we propose Verify-in-the-Graph (VeGraph), a novel framework leveraging the reasoning and comprehension abilities of LLM agents. VeGraph operates in three phases: (1) Graph Representation - an input claim is decomposed into structured triplets, forming a graph-based representation that integrates both structured and unstructured information; (2) Entity Disambiguation -VeGraph iteratively interacts with the knowledge base to resolve ambiguous entities within the graph for deeper sub-claim verification; and (3) Verification - remaining triplets are verified to complete the fact-checking process. Experiments using Meta-Llama-3-70B (instruct version) show that VeGraph achieves competitive performance compared to baselines on two benchmarks HoVer and FEVEROUS, effectively addressing claim verification challenges. Our source code and data are available for further exploitation. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:02:55 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Pham",
"Hoang",
""
],
[
"Nguyen",
"Thanh-Do",
""
],
[
"Bui",
"Khac-Hoai Nam",
""
]
] |
2505.22994 | Kyle Daruwalla | Ari S. Benjamin, Kyle Daruwalla, Christian Pehle, Anthony M. Zador | Walking the Weight Manifold: a Topological Approach to Conditioning Inspired by Neuromodulation | 17 pages, 4 figures | null | null | null | cs.LG cs.NE | http://creativecommons.org/licenses/by/4.0/ | One frequently wishes to learn a range of similar tasks as efficiently as possible, re-using knowledge across tasks. In artificial neural networks, this is typically accomplished by conditioning a network upon task context by injecting context as input. Brains have a different strategy: the parameters themselves are modulated as a function of various neuromodulators such as serotonin. Here, we take inspiration from neuromodulation and propose to learn weights which are smoothly parameterized functions of task context variables. Rather than optimize a weight vector, i.e. a single point in weight space, we optimize a smooth manifold in weight space with a predefined topology. To accomplish this, we derive a formal treatment of optimization of manifolds as the minimization of a loss functional subject to a constraint on volumetric movement, analogous to gradient descent. During inference, conditioning selects a single point on this manifold which serves as the effective weight matrix for a particular sub-task. This strategy for conditioning has two main advantages. First, the topology of the manifold (whether a line, circle, or torus) is a convenient lever for inductive biases about the relationship between tasks. Second, learning in one state smoothly affects the entire manifold, encouraging generalization across states. To verify this, we train manifolds with several topologies, including straight lines in weight space (for conditioning on e.g. noise level in input data) and ellipses (for rotated images). Despite their simplicity, these parameterizations outperform conditioning identical networks by input concatenation and better generalize to out-of-distribution samples. These results suggest that modulating weights over low-dimensional manifolds offers a principled and effective alternative to traditional conditioning. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:03:29 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Benjamin",
"Ari S.",
""
],
[
"Daruwalla",
"Kyle",
""
],
[
"Pehle",
"Christian",
""
],
[
"Zador",
"Anthony M.",
""
]
] |
2505.22995 | Quan Wang | Pai Zhu, Quan Wang, Dhruuv Agarwal, Kurt Partridge | LLM-Synth4KWS: Scalable Automatic Generation and Synthesis of Confusable Data for Custom Keyword Spotting | null | null | null | null | eess.AS cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Custom keyword spotting (KWS) allows detecting user-defined spoken keywords from streaming audio. This is achieved by comparing the embeddings from voice enrollments and input audio. State-of-the-art custom KWS models are typically trained contrastively using utterances whose keywords are randomly sampled from training dataset. These KWS models often struggle with confusing keywords, such as "blue" versus "glue". This paper introduces an effective way to augment the training with confusable utterances where keywords are generated and grouped from large language models (LLMs), and speech signals are synthesized with diverse speaking styles from text-to-speech (TTS) engines. To better measure user experience on confusable KWS, we define a new northstar metric using the average area under DET curve from confusable groups (c-AUC). Featuring high scalability and zero labor cost, the proposed method improves AUC by 3.7% and c-AUC by 11.3% on the Speech Commands testing set. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:05:26 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Zhu",
"Pai",
""
],
[
"Wang",
"Quan",
""
],
[
"Agarwal",
"Dhruuv",
""
],
[
"Partridge",
"Kurt",
""
]
] |
2505.22996 | Joshua Peters Mr | Cecilia Gonz\'alez-Tokman and Joshua Peters | Jumping for diffusion in random metastable systems | arXiv admin note: text overlap with arXiv:2406.09672 | null | null | null | math.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Random metastability occurs when an externally forced or noisy system possesses more than one state of apparent equilibrium. This work investigates fluctuations in a class of random dynamical systems, arising from randomly perturbing a piecewise smooth expanding interval map with more than one invariant subinterval. Upon perturbation, this invariance is destroyed, allowing trajectories to switch between subintervals, giving rise to metastable behaviour. We show that the distributions of jumps of a time-homogeneous Markov chain approximate the distributions of jumps for random metastable systems. Additionally, we demonstrate that this approximation extends to the diffusion coefficient for (random) observables of such systems. As an example, our results are applied to Horan's random paired tent maps. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:06:24 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"González-Tokman",
"Cecilia",
""
],
[
"Peters",
"Joshua",
""
]
] |
2505.22997 | Agnideep Aich | Agnideep Aich, Ashit Baran Aich, Bruce Wade | Theoretical Foundations of the Deep Copula Classifier: A Generative Approach to Modeling Dependent Features | Submitted | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | Traditional classifiers often assume feature independence or rely on overly simplistic relationships, leading to poor performance in settings where real-world dependencies matter. We introduce the Deep Copula Classifier (DCC), a generative model that separates the learning of each feature's marginal distribution from the modeling of their joint dependence structure via neural network-parameterized copulas. For each class, lightweight neural networks are used to flexibly and adaptively capture feature interactions, making DCC particularly effective when classification is driven by complex dependencies. We establish that DCC converges to the Bayes-optimal classifier under standard conditions and provide explicit convergence rates of O(n^{-r/(2r + d)}) for r-smooth copula densities. Beyond theoretical guarantees, we outline several practical extensions, including high-dimensional scalability through vine and factor copula architectures, semi-supervised learning via entropy regularization, and online adaptation using streaming gradient methods. By unifying statistical rigor with the representational power of neural networks, DCC offers a mathematically grounded and interpretable framework for dependency-aware classification. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:07:26 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Aich",
"Agnideep",
""
],
[
"Aich",
"Ashit Baran",
""
],
[
"Wade",
"Bruce",
""
]
] |
2505.22998 | Jihwan Oh | Jihwan Oh, Murad Aghazada, Se-Young Yun, Taehyeon Kim | LLM Agents for Bargaining with Utility-based Feedback | Preprint | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bargaining, a critical aspect of real-world interactions, presents challenges for large language models (LLMs) due to limitations in strategic depth and adaptation to complex human factors. Existing benchmarks often fail to capture this real-world complexity. To address this and enhance LLM capabilities in realistic bargaining, we introduce a comprehensive framework centered on utility-based feedback. Our contributions are threefold: (1) BargainArena, a novel benchmark dataset with six intricate scenarios (e.g., deceptive practices, monopolies) to facilitate diverse strategy modeling; (2) human-aligned, economically-grounded evaluation metrics inspired by utility theory, incorporating agent utility and negotiation power, which implicitly reflect and promote opponent-aware reasoning (OAR); and (3) a structured feedback mechanism enabling LLMs to iteratively refine their bargaining strategies. This mechanism can positively collaborate with in-context learning (ICL) prompts, including those explicitly designed to foster OAR. Experimental results show that LLMs often exhibit negotiation strategies misaligned with human preferences, and that our structured feedback mechanism significantly improves their performance, yielding deeper strategic and opponent-aware reasoning. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:07:27 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Oh",
"Jihwan",
""
],
[
"Aghazada",
"Murad",
""
],
[
"Yun",
"Se-Young",
""
],
[
"Kim",
"Taehyeon",
""
]
] |
2505.22999 | Yihua Xu | Yihua Xu, S\"uleyman Kerimov, Sebastian Perez-Salazar | Online Selection with Uncertain Disruption | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In numerous online selection problems, decision-makers (DMs) must allocate on the fly limited resources to customers with uncertain values. The DM faces the tension between allocating resources to currently observed values and saving them for potentially better, unobserved values in the future. Addressing this tension becomes more demanding if an uncertain disruption occurs while serving customers. Without any disruption, the DM gets access to the capacity information to serve customers throughout the time horizon. However, with uncertain disruption, the DM must act more cautiously due to risk of running out of capacity abruptly or misusing the resources. Motivated by this tension, we introduce the Online Selection with Uncertain Disruption (OS-UD) problem. In OS-UD, a DM sequentially observes n non-negative values drawn from a common distribution and must commit to select or reject each value in real time, without revisiting past values. The disruption is modeled as a Bernoulli random variable with probability p each time DM selects a value. We aim to design an online algorithm that maximizes the expected sum of selected values before a disruption occurs, if any.
We evaluate online algorithms using the competitive ratio. Using a quantile-based approach, we devise a non-adaptive single-threshold algorithm that attains a competitive ratio of at least 1-1/e, and an adaptive threshold algorithm characterized by a sequence of non-increasing thresholds that attains an asymptotic competitive ratio of at least 0.745. Both of these results are worst-case optimal within their corresponding class of algorithms. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:09:15 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Xu",
"Yihua",
""
],
[
"Kerimov",
"Süleyman",
""
],
[
"Perez-Salazar",
"Sebastian",
""
]
] |
2505.23000 | Yan Peng | Guohua Liu and Yan Peng | Bounds on the minimum orbital periods of non-singular Hayward and Bardeen black holes | 9 pages, 2 figures | null | null | null | gr-qc hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Based on previous studies, universal bounds $4\pi M \leqslant T_{min} \leqslant 6\sqrt{3}\pi M$ were conjectured to be characteristic properties of black hole spacetimes, where $M$ represents the mass of black holes and $T_{min}$ is the minimum orbital periods around black holes. In this work, we explore the minimum orbital periods of objects around Hayward and Bardeen black holes without central singularities. By combining analytical and numerical methods, we show that both Hayward and Bardeen black holes conform to these bounds. Our results imply that such bounds may be connected to the presence of the black hole horizon rather than the singularity. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:19:04 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Liu",
"Guohua",
""
],
[
"Peng",
"Yan",
""
]
] |
2505.23001 | Yize Cheng | Yize Cheng, Wenxiao Wang, Mazda Moayeri, Soheil Feizi | DyePack: Provably Flagging Test Set Contamination in LLMs Using Backdoors | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Open benchmarks are essential for evaluating and advancing large language models, offering reproducibility and transparency. However, their accessibility makes them likely targets of test set contamination. In this work, we introduce DyePack, a framework that leverages backdoor attacks to identify models that used benchmark test sets during training, without requiring access to the loss, logits, or any internal details of the model. Like how banks mix dye packs with their money to mark robbers, DyePack mixes backdoor samples with the test data to flag models that trained on it. We propose a principled design incorporating multiple backdoors with stochastic targets, enabling exact false positive rate (FPR) computation when flagging every model. This provably prevents false accusations while providing strong evidence for every detected case of contamination. We evaluate DyePack on five models across three datasets, covering both multiple-choice and open-ended generation tasks. For multiple-choice questions, it successfully detects all contaminated models with guaranteed FPRs as low as 0.000073% on MMLU-Pro and 0.000017% on Big-Bench-Hard using eight backdoors. For open-ended generation tasks, it generalizes well and identifies all contaminated models on Alpaca with a guaranteed false positive rate of just 0.127% using six backdoors. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:22:14 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Cheng",
"Yize",
""
],
[
"Wang",
"Wenxiao",
""
],
[
"Moayeri",
"Mazda",
""
],
[
"Feizi",
"Soheil",
""
]
] |
2505.23002 | Qiao Zhu | Q. Zhu, D. Chaikovskii, B. Jin, Y. Zhang | Deep asymptotic expansion method for solving singularly perturbed time-dependent reaction-advection-diffusion equations | 24 pages | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Physics-informed neural network (PINN) has shown great potential in solving differential equations. However, it faces challenges when dealing with problems involving steep gradients. For singularly perturbed time-dependent reaction-advection-diffusion equations, which exhibit internal transition layers with sharp gradients, we propose a deep asymptotic expansion (DAE) method that leverages deep learning to obtain explicit smooth approximate solutions. Inspired by asymptotic analysis, we first derive the governing equations for transition layers and then solve them using PINN. Numerical experiments show that DAE outperforms PINN, gPINN and PINN with adaptive sampling. We also show its robustness with respect to training point distributions, network architectures, and random seeds. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:22:38 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Zhu",
"Q.",
""
],
[
"Chaikovskii",
"D.",
""
],
[
"Jin",
"B.",
""
],
[
"Zhang",
"Y.",
""
]
] |
2505.23003 | Linh Le Pham Van Mr | Linh Le Pham Van, Minh Hoang Nguyen, Hung Le, Hung The Tran, Sunil Gupta | Hybrid Cross-domain Robust Reinforcement Learning | Accepted at ECML PKDD 2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Robust reinforcement learning (RL) aims to learn policies that remain effective despite uncertainties in its environment, which frequently arise in real-world applications due to variations in environment dynamics. The robust RL methods learn a robust policy by maximizing value under the worst-case models within a predefined uncertainty set. Offline robust RL algorithms are particularly promising in scenarios where only a fixed dataset is available and new data cannot be collected. However, these approaches often require extensive offline data, and gathering such datasets for specific tasks in specific environments can be both costly and time-consuming. Using an imperfect simulator offers a faster, cheaper, and safer way to collect data for training, but it can suffer from dynamics mismatch. In this paper, we introduce HYDRO, the first Hybrid Cross-Domain Robust RL framework designed to address these challenges. HYDRO utilizes an online simulator to complement the limited amount of offline datasets in the non-trivial context of robust RL. By measuring and minimizing performance gaps between the simulator and the worst-case models in the uncertainty set, HYDRO employs novel uncertainty filtering and prioritized sampling to select the most relevant and reliable simulator samples. Our extensive experiments demonstrate HYDRO's superior performance over existing methods across various tasks, underscoring its potential to improve sample efficiency in offline robust RL. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:25:13 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Van",
"Linh Le Pham",
""
],
[
"Nguyen",
"Minh Hoang",
""
],
[
"Le",
"Hung",
""
],
[
"Tran",
"Hung The",
""
],
[
"Gupta",
"Sunil",
""
]
] |
2505.23004 | Kyle Chickering | Kyle R. Chickering and Bangzheng Li and Muhao Chen | QLIP: A Dynamic Quadtree Vision Prior Enhances MLLM Performance Without Retraining | 22 pages, 19 figures | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Multimodal Large Language Models (MLLMs) encode images into visual tokens, aligning visual and textual signals within a shared latent space to facilitate crossmodal representation learning. The CLIP model is a widely adopted foundational vision language model whose vision encoder has played a critical role in the development of MLLMs such as LLaVA. However, the CLIP vision encoder suffers from notable limitations including being constrained to only handling fixed input resolutions and a failure to produce separated embeddings for dissimilar images. Replacing the vision encoder of an existing model typically incurs substantial computational costs because such a change often necessitates retraining the entire model pipeline.
In this work, we identify two factors which underlie the limitations of the CLIP vision encoder: mesoscopic bias and interpolation bias. To address these issues, we propose QLIP, a drop-in replacement for CLIP that can be seamlessly integrated with existing MLLMs with only a few lines of code and can enhance both coarse-grained and fine-grained visual understanding, without re-training. QLIP is designed around an image quadtree which replaces the standard uniform grid patches with a novel content aware patchification. Our experimental results demonstrate that QLIP improves the general visual question answering accuracy of the LLaVA v1.5 model series across various model sizes--without requiring retraining or fine-tuning of the full MLLM. Notably, QLIP boosts detailed understanding performance on the challenging $V^{\ast}$ benchmark by up to 13.6 percent. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:26:34 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Chickering",
"Kyle R.",
""
],
[
"Li",
"Bangzheng",
""
],
[
"Chen",
"Muhao",
""
]
] |
2505.23005 | Harshda Saxena | H. Saxena, J. Sayers, A. Gavidia, J.B. Melin, E.T. Lau, J. Kim, L. Chappuis, D. Eckert, S. Ettori, M. Gaspari, F. Gastaldello, S. Kay, L. Lovisari, F. Oppizzi, M.D. Petris, G.W. Pratt, E. Pointecouteau, E. Rasia, M. Rossetti, M. Sereno | CHEX-MATE: The Impact of Triaxiality and Orientation on Planck SZ Cluster Selection and Weak Lensing Mass Measurements | null | null | null | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Galaxy cluster abundance measurements are a valuable tool for constraining cosmological parameters like the mass density ($\Omega_m$) and density fluctuation amplitude ($\sigma_8$). Wide area surveys detect clusters based on observables, such as the total integrated Sunyaev-Zel'dovich effect signal ($Y_{SZ}$) in the case of Planck. Quantifying the survey selection function is necessary for a cosmological analysis, with completeness representing the probability of detecting a cluster as a function of its intrinsic properties. Employing a Monte-Carlo method, we inject triaxial cluster profiles into random positions within the Planck all-sky maps, and subsequently determine the completeness of the Planck-selected CHEXMATE sample as a function of both geometry and SZ brightness. This is then used to generate 1000 mock CHEX-MATE cluster catalogs, and the distribution of shapes and orientations of the detected clusters, along with any associated bias in weak lensing-derived mass ($M_{WL}$) due to this orientation-dependent selection, denoted as $1 - b_{\chi}$, is obtained. We show that cluster orientation impacts completeness, with a higher probability of detecting clusters elongated along the line of sight (LOS). This leads to $1 - b_{\chi}$ values of $0-4\%$ for CHEXMATE clusters relative to a random population. The largest increase in $M_{WL}$ is observed in the lowest mass objects, which are most impacted by orientation-related selection bias. This bias is relevant for upcoming SZ surveys like CMB-S4, and should be considered for surveys utilizing other probes for cluster detection, such as Euclid. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:29:34 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Saxena",
"H.",
""
],
[
"Sayers",
"J.",
""
],
[
"Gavidia",
"A.",
""
],
[
"Melin",
"J. B.",
""
],
[
"Lau",
"E. T.",
""
],
[
"Kim",
"J.",
""
],
[
"Chappuis",
"L.",
""
],
[
"Eckert",
"D.",
""
],
[
"Ettori",
"S.",
""
],
[
"Gaspari",
"M.",
""
],
[
"Gastaldello",
"F.",
""
],
[
"Kay",
"S.",
""
],
[
"Lovisari",
"L.",
""
],
[
"Oppizzi",
"F.",
""
],
[
"Petris",
"M. D.",
""
],
[
"Pratt",
"G. W.",
""
],
[
"Pointecouteau",
"E.",
""
],
[
"Rasia",
"E.",
""
],
[
"Rossetti",
"M.",
""
],
[
"Sereno",
"M.",
""
]
] |
2505.23006 | Daeryong Kim | Chiwan Park, Wonjun Jang, Daeryong Kim, Aelim Ahn, Kichang Yang, Woosung Hwang, Jihyeon Roh, Hyerin Park, Hyosun Wang, Min Seok Kim, Jihoon Kang | A Practical Approach for Building Production-Grade Conversational Agents with Workflow Graphs | Accepted to ACL 2025 Industry Track. 12 pages, 5 figures | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The advancement of Large Language Models (LLMs) has led to significant improvements in various service domains, including search, recommendation, and chatbot applications. However, applying state-of-the-art (SOTA) research to industrial settings presents challenges, as it requires maintaining flexible conversational abilities while also strictly complying with service-specific constraints. This can be seen as two conflicting requirements due to the probabilistic nature of LLMs. In this paper, we propose our approach to addressing this challenge and detail the strategies we employed to overcome their inherent limitations in real-world applications. We conduct a practical case study of a conversational agent designed for the e-commerce domain, detailing our implementation workflow and optimizations. Our findings provide insights into bridging the gap between academic research and real-world application, introducing a framework for developing scalable, controllable, and reliable AI-driven agents. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:30:27 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Park",
"Chiwan",
""
],
[
"Jang",
"Wonjun",
""
],
[
"Kim",
"Daeryong",
""
],
[
"Ahn",
"Aelim",
""
],
[
"Yang",
"Kichang",
""
],
[
"Hwang",
"Woosung",
""
],
[
"Roh",
"Jihyeon",
""
],
[
"Park",
"Hyerin",
""
],
[
"Wang",
"Hyosun",
""
],
[
"Kim",
"Min Seok",
""
],
[
"Kang",
"Jihoon",
""
]
] |
2505.23007 | Masanori Tanaka | Qing-Hong Cao, Masanori Tanaka, Jun-Chen Wang, Ke-Pan Xie, Jing-Jun Zhang | Can we live in a baby universe formed by a delayed first-order phase transition? | 5 pages, 2 figures | null | null | CPTNP-2025-014 | hep-ph gr-qc | http://creativecommons.org/licenses/by/4.0/ | We examine the idea that our universe began as a baby universe and show that this is feasible in a $U(1)_{B-L}$ extension of the Standard Model with the classically conformal principle. The framework is consistent with current cosmological data and predicts a heavy neutral gauge boson, which could be detected at colliders. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:32:15 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Cao",
"Qing-Hong",
""
],
[
"Tanaka",
"Masanori",
""
],
[
"Wang",
"Jun-Chen",
""
],
[
"Xie",
"Ke-Pan",
""
],
[
"Zhang",
"Jing-Jun",
""
]
] |
2505.23008 | Jonathan Li | Jonathan Li, Zoltan Csaki, Nidhi Hiremath, Etash Guha, Fenglu Hong, Edward Ma, Urmish Thakker | Synthetic Document Question Answering in Hungarian | null | null | null | null | cs.CV cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Modern VLMs have achieved near-saturation accuracy in English document visual question-answering (VQA). However, this task remains challenging in lower resource languages due to a dearth of suitable training and evaluation data. In this paper we present scalable methods for curating such datasets by focusing on Hungarian, approximately the 17th highest resource language on the internet. Specifically, we present HuDocVQA and HuDocVQA-manual, document VQA datasets that modern VLMs significantly underperform on compared to English DocVQA. HuDocVQA-manual is a small manually curated dataset based on Hungarian documents from Common Crawl, while HuDocVQA is a larger synthetically generated VQA data set from the same source. We apply multiple rounds of quality filtering and deduplication to HuDocVQA in order to match human-level quality in this dataset. We also present HuCCPDF, a dataset of 117k pages from Hungarian Common Crawl PDFs along with their transcriptions, which can be used for training a model for Hungarian OCR. To validate the quality of our datasets, we show how finetuning on a mixture of these datasets can improve accuracy on HuDocVQA for Llama 3.2 11B Instruct by +7.2%. Our datasets and code will be released to the public to foster further research in multilingual DocVQA. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:34:36 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Li",
"Jonathan",
""
],
[
"Csaki",
"Zoltan",
""
],
[
"Hiremath",
"Nidhi",
""
],
[
"Guha",
"Etash",
""
],
[
"Hong",
"Fenglu",
""
],
[
"Ma",
"Edward",
""
],
[
"Thakker",
"Urmish",
""
]
] |
2505.23009 | Ruskin Raj Manku | Ruskin Raj Manku, Yuzhi Tang, Xingjian Shi, Mu Li, Alex Smola | EmergentTTS-Eval: Evaluating TTS Models on Complex Prosodic, Expressiveness, and Linguistic Challenges Using Model-as-a-Judge | null | null | null | null | cs.LG cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | Text-to-Speech (TTS) benchmarks often fail to capture how well models handle nuanced and semantically complex text. Building on $\textit{EmergentTTS}$, we introduce $\textit{EmergentTTS-Eval}$, a comprehensive benchmark covering six challenging TTS scenarios: emotions, paralinguistics, foreign words, syntactic complexity, complex pronunciation (e.g. URLs, formulas), and questions. Crucially, our framework automates both test-case generation and evaluation, making the benchmark easily extensible. Starting from a small set of human-written seed prompts, we iteratively extend them using LLMs to target specific structural, phonetic and prosodic challenges, resulting in 1,645 diverse test cases. Moreover, we employ a model-as-a-judge approach, using a Large Audio Language Model (LALM) to assess the speech across multiple dimensions such as expressed emotion, prosodic, intonational, and pronunciation accuracy. We evaluate state-of-the-art open-source and proprietary TTS systems, such as 11Labs, Deepgram, and OpenAI's 4o-mini-TTS, on EmergentTTS-Eval, demonstrating its ability to reveal fine-grained performance differences. Results show that the model-as-a-judge approach offers robust TTS assessment and a high correlation with human preferences. We open source the evaluation $\href{https://github.com/boson-ai/EmergentTTS-Eval-public}{code}$ and the $\href{https://huggingface.co/datasets/bosonai/EmergentTTS-Eval}{dataset}$. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:36:24 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Manku",
"Ruskin Raj",
""
],
[
"Tang",
"Yuzhi",
""
],
[
"Shi",
"Xingjian",
""
],
[
"Li",
"Mu",
""
],
[
"Smola",
"Alex",
""
]
] |
2505.23010 | Bowen Chen | Bowen Chen, Keyan Chen, Mohan Yang, Zhengxia Zou, Zhenwei Shi | SeG-SR: Integrating Semantic Knowledge into Remote Sensing Image Super-Resolution via Vision-Language Model | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | High-resolution (HR) remote sensing imagery plays a vital role in a wide range of applications, including urban planning and environmental monitoring. However, due to limitations in sensors and data transmission links, the images acquired in practice often suffer from resolution degradation. Remote Sensing Image Super-Resolution (RSISR) aims to reconstruct HR images from low-resolution (LR) inputs, providing a cost-effective and efficient alternative to direct HR image acquisition. Existing RSISR methods primarily focus on low-level characteristics in pixel space, while neglecting the high-level understanding of remote sensing scenes. This may lead to semantically inconsistent artifacts in the reconstructed results. Motivated by this observation, our work aims to explore the role of high-level semantic knowledge in improving RSISR performance. We propose a Semantic-Guided Super-Resolution framework, SeG-SR, which leverages Vision-Language Models (VLMs) to extract semantic knowledge from input images and uses it to guide the super resolution (SR) process. Specifically, we first design a Semantic Feature Extraction Module (SFEM) that utilizes a pretrained VLM to extract semantic knowledge from remote sensing images. Next, we propose a Semantic Localization Module (SLM), which derives a series of semantic guidance from the extracted semantic knowledge. Finally, we develop a Learnable Modulation Module (LMM) that uses semantic guidance to modulate the features extracted by the SR network, effectively incorporating high-level scene understanding into the SR pipeline. We validate the effectiveness and generalizability of SeG-SR through extensive experiments: SeG-SR achieves state-of-the-art performance on two datasets and consistently delivers performance improvements across various SR architectures. Codes can be found at https://github.com/Mr-Bamboo/SeG-SR. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:38:34 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Chen",
"Bowen",
""
],
[
"Chen",
"Keyan",
""
],
[
"Yang",
"Mohan",
""
],
[
"Zou",
"Zhengxia",
""
],
[
"Shi",
"Zhenwei",
""
]
] |
2505.23011 | Oscar Dahlsten Dr | Oscar C.O. Dahlsten | Page Curve of average subsystem entropy | Chapter in "Black Hole Information Paradox", editors C. Bambini and A. Akil, Springer, 2025 | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Page curve is a curve of average subsystem entropy as a function of subsystem size. The curve starts from 0, rises, peaks near the maximal possible entropy when the subsystem makes up half of the total system, and then falls back to 0. We here describe subsystem entropy, averaging over quantum states, and why the curve rises and falls in that manner. We also discuss the connection between the curve and the black hole information paradox. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:40:07 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Dahlsten",
"Oscar C. O.",
""
]
] |
2505.23012 | Shanaka Ramesh | Shanaka Ramesh Gunasekara, Wanqing Li, Philip Ogunbona, Jack Yang | Spatio-Temporal Joint Density Driven Learning for Skeleton-Based Action Recognition | null | IEEE Transactions on Biometrics, Behavior, and Identity Science (2025) | 10.1109/TBIOM.2025.3566212 | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Traditional approaches in unsupervised or self supervised learning for skeleton-based action classification have concentrated predominantly on the dynamic aspects of skeletal sequences. Yet, the intricate interaction between the moving and static elements of the skeleton presents a rarely tapped discriminative potential for action classification. This paper introduces a novel measurement, referred to as spatial-temporal joint density (STJD), to quantify such interaction. Tracking the evolution of this density throughout an action can effectively identify a subset of discriminative moving and/or static joints termed "prime joints" to steer self-supervised learning. A new contrastive learning strategy named STJD-CL is proposed to align the representation of a skeleton sequence with that of its prime joints while simultaneously contrasting the representations of prime and nonprime joints. In addition, a method called STJD-MP is developed by integrating it with a reconstruction-based framework for more effective learning. Experimental evaluations on the NTU RGB+D 60, NTU RGB+D 120, and PKUMMD datasets in various downstream tasks demonstrate that the proposed STJD-CL and STJD-MP improved performance, particularly by 3.5 and 3.6 percentage points over the state-of-the-art contrastive methods on the NTU RGB+D 120 dataset using X-sub and X-set evaluations, respectively. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:40:47 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Gunasekara",
"Shanaka Ramesh",
""
],
[
"Li",
"Wanqing",
""
],
[
"Ogunbona",
"Philip",
""
],
[
"Yang",
"Jack",
""
]
] |
2505.23013 | Junjie Yao | Liangkai Hang, Junjie Yao, Zhiwei Bai, Tianyi Chen, Yang Chen, Rongjie Diao, Hezhou Li, Pengxiao Lin, Zhiwei Wang, Cheng Xu, Zhongwang Zhang, Zhangchen Zhou, Zhiyu Li, Zehao Lin, Kai Chen, Feiyu Xiong, Yaoyu Zhang, Weinan E, Hongkang Yang, Zhi-Qin John Xu | Scalable Complexity Control Facilitates Reasoning Ability of LLMs | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The reasoning ability of large language models (LLMs) has been rapidly advancing in recent years, attracting interest in more fundamental approaches that can reliably enhance their generalizability. This work demonstrates that model complexity control, conveniently implementable by adjusting the initialization rate and weight decay coefficient, improves the scaling law of LLMs consistently over varying model sizes and data sizes. This gain is further illustrated by comparing the benchmark performance of 2.4B models pretrained on 1T tokens with different complexity hyperparameters. Instead of fixing the initialization std, we found that a constant initialization rate (the exponent of std) enables the scaling law to descend faster in both model and data sizes. These results indicate that complexity control is a promising direction for the continual advancement of LLMs. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:42:20 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Hang",
"Liangkai",
""
],
[
"Yao",
"Junjie",
""
],
[
"Bai",
"Zhiwei",
""
],
[
"Chen",
"Tianyi",
""
],
[
"Chen",
"Yang",
""
],
[
"Diao",
"Rongjie",
""
],
[
"Li",
"Hezhou",
""
],
[
"Lin",
"Pengxiao",
""
],
[
"Wang",
"Zhiwei",
""
],
[
"Xu",
"Cheng",
""
],
[
"Zhang",
"Zhongwang",
""
],
[
"Zhou",
"Zhangchen",
""
],
[
"Li",
"Zhiyu",
""
],
[
"Lin",
"Zehao",
""
],
[
"Chen",
"Kai",
""
],
[
"Xiong",
"Feiyu",
""
],
[
"Zhang",
"Yaoyu",
""
],
[
"E",
"Weinan",
""
],
[
"Yang",
"Hongkang",
""
],
[
"Xu",
"Zhi-Qin John",
""
]
] |
2505.23014 | Juwei Yue | Juwei Yue, Haikuo Li, Jiawei Sheng, Xiaodong Li, Taoyu Su, Tingwen Liu, Li Guo | Hyperbolic-PDE GNN: Spectral Graph Neural Networks in the Perspective of A System of Hyperbolic Partial Differential Equations | 18 pages, 2 figures, published to ICML 2025 | International Conference on Machine Learning 2025 | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph neural networks (GNNs) leverage message passing mechanisms to learn the topological features of graph data. Traditional GNNs learns node features in a spatial domain unrelated to the topology, which can hardly ensure topological features. In this paper, we formulates message passing as a system of hyperbolic partial differential equations (hyperbolic PDEs), constituting a dynamical system that explicitly maps node representations into a particular solution space. This solution space is spanned by a set of eigenvectors describing the topological structure of graphs. Within this system, for any moment in time, a node features can be decomposed into a superposition of the basis of eigenvectors. This not only enhances the interpretability of message passing but also enables the explicit extraction of fundamental characteristics about the topological structure. Furthermore, by solving this system of hyperbolic partial differential equations, we establish a connection with spectral graph neural networks (spectral GNNs), serving as a message passing enhancement paradigm for spectral GNNs.We further introduce polynomials to approximate arbitrary filter functions. Extensive experiments demonstrate that the paradigm of hyperbolic PDEs not only exhibits strong flexibility but also significantly enhances the performance of various spectral GNNs across diverse graph tasks. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:49:26 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Yue",
"Juwei",
""
],
[
"Li",
"Haikuo",
""
],
[
"Sheng",
"Jiawei",
""
],
[
"Li",
"Xiaodong",
""
],
[
"Su",
"Taoyu",
""
],
[
"Liu",
"Tingwen",
""
],
[
"Guo",
"Li",
""
]
] |
2505.23015 | Jinwen Chen | Jinwen Chen and Hainan Zhang and Fei Sun and Qinnan Zhang and Sijia Wen and Ziwei Wang and Zhiming Zheng | Detecting Stealthy Backdoor Samples based on Intra-class Distance for Large Language Models | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fine-tuning LLMs with datasets containing stealthy backdoors from publishers poses security risks to downstream applications. Mainstream detection methods either identify poisoned samples by analyzing the prediction probability of poisoned classification models or rely on the rewriting model to eliminate the stealthy triggers. However, the former cannot be applied to generation tasks, while the latter may degrade generation performance and introduce new triggers. Therefore, efficiently eliminating stealthy poisoned samples for LLMs remains an urgent problem. We observe that after applying TF-IDF clustering to the sample response, there are notable differences in the intra-class distances between clean and poisoned samples. Poisoned samples tend to cluster closely because of their specific malicious outputs, whereas clean samples are more scattered due to their more varied responses. Thus, in this paper, we propose a stealthy backdoor sample detection method based on Reference-Filtration and Tfidf-Clustering mechanisms (RFTC). Specifically, we first compare the sample response with the reference model's outputs and consider the sample suspicious if there's a significant discrepancy. And then we perform TF-IDF clustering on these suspicious samples to identify the true poisoned samples based on the intra-class distance. Experiments on two machine translation datasets and one QA dataset demonstrate that RFTC outperforms baselines in backdoor detection and model performance. Further analysis of different reference models also confirms the effectiveness of our Reference-Filtration. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:49:29 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Chen",
"Jinwen",
""
],
[
"Zhang",
"Hainan",
""
],
[
"Sun",
"Fei",
""
],
[
"Zhang",
"Qinnan",
""
],
[
"Wen",
"Sijia",
""
],
[
"Wang",
"Ziwei",
""
],
[
"Zheng",
"Zhiming",
""
]
] |
2505.23016 | Aniruddh Mishra | Aniruddh Mishra, Arthur K. Barnes, Jose E. Tabarez, Adam Mate | Sensitivity of DC Network Representation for GIC Analysis | 6 pages, 1 table, 7 figures | Proceedings of the 2025 IEEE Texas Power and Energy Conference (TPEC) | 10.1109/TPEC63981.2025.10907180 | LA-UR-24-31515 | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | Geomagnetic disturbances are a threat to the reliability and security of our national critical energy infrastructures. These events specifically result in geomagnetically induced currents, which can cause damage to transformers due to magnetic saturation. In order to mitigate these effects, blocker devices must be placed in optimal locations. Finding this placement requires a dc representation of the ac transmission lines, which this paper discusses. Different decisions in this process, including the method of representing the blocking devices, result in significant variations to the power loss calculations. To analyze these effects, we conclude the paper by comparing the losses on a sample network with different modeling implementations. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:49:57 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Mishra",
"Aniruddh",
""
],
[
"Barnes",
"Arthur K.",
""
],
[
"Tabarez",
"Jose E.",
""
],
[
"Mate",
"Adam",
""
]
] |
2505.23017 | Xingjian Wu | Xingjian Wu, Xiangfei Qiu, Hongfan Gao, Jilin Hu, Bin Yang, Chenjuan Guo | $K^2$VAE: A Koopman-Kalman Enhanced Variational AutoEncoder for Probabilistic Time Series Forecasting | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Probabilistic Time Series Forecasting (PTSF) plays a crucial role in decision-making across various fields, including economics, energy, and transportation. Most existing methods excell at short-term forecasting, while overlooking the hurdles of Long-term Probabilistic Time Series Forecasting (LPTSF). As the forecast horizon extends, the inherent nonlinear dynamics have a significant adverse effect on prediction accuracy, and make generative models inefficient by increasing the cost of each iteration. To overcome these limitations, we introduce $K^2$VAE, an efficient VAE-based generative model that leverages a KoopmanNet to transform nonlinear time series into a linear dynamical system, and devises a KalmanNet to refine predictions and model uncertainty in such linear system, which reduces error accumulation in long-term forecasting. Extensive experiments demonstrate that $K^2$VAE outperforms state-of-the-art methods in both short- and long-term PTSF, providing a more efficient and accurate solution. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:52:59 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Wu",
"Xingjian",
""
],
[
"Qiu",
"Xiangfei",
""
],
[
"Gao",
"Hongfan",
""
],
[
"Hu",
"Jilin",
""
],
[
"Yang",
"Bin",
""
],
[
"Guo",
"Chenjuan",
""
]
] |
2505.23018 | Haoqin Sun | Haoqin Sun, Xuechen Wang, Jinghua Zhao, Shiwan Zhao, Jiaming Zhou, Hui Wang, Jiabei He, Aobo Kong, Xi Yang, Yequan Wang, Yonghua Lin, Yong Qin | EmotionTalk: An Interactive Chinese Multimodal Emotion Dataset With Rich Annotations | null | null | null | null | cs.MM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In recent years, emotion recognition plays a critical role in applications such as human-computer interaction, mental health monitoring, and sentiment analysis. While datasets for emotion analysis in languages such as English have proliferated, there remains a pressing need for high-quality, comprehensive datasets tailored to the unique linguistic, cultural, and multimodal characteristics of Chinese. In this work, we propose \textbf{EmotionTalk}, an interactive Chinese multimodal emotion dataset with rich annotations. This dataset provides multimodal information from 19 actors participating in dyadic conversational settings, incorporating acoustic, visual, and textual modalities. It includes 23.6 hours of speech (19,250 utterances), annotations for 7 utterance-level emotion categories (happy, surprise, sad, disgust, anger, fear, and neutral), 5-dimensional sentiment labels (negative, weakly negative, neutral, weakly positive, and positive) and 4-dimensional speech captions (speaker, speaking style, emotion and overall). The dataset is well-suited for research on unimodal and multimodal emotion recognition, missing modality challenges, and speech captioning tasks. To our knowledge, it represents the first high-quality and versatile Chinese dialogue multimodal emotion dataset, which is a valuable contribution to research on cross-cultural emotion analysis and recognition. Additionally, we conduct experiments on EmotionTalk to demonstrate the effectiveness and quality of the dataset. It will be open-source and freely available for all academic purposes. The dataset and codes will be made available at: https://github.com/NKU-HLT/EmotionTalk. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 02:56:08 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Sun",
"Haoqin",
""
],
[
"Wang",
"Xuechen",
""
],
[
"Zhao",
"Jinghua",
""
],
[
"Zhao",
"Shiwan",
""
],
[
"Zhou",
"Jiaming",
""
],
[
"Wang",
"Hui",
""
],
[
"He",
"Jiabei",
""
],
[
"Kong",
"Aobo",
""
],
[
"Yang",
"Xi",
""
],
[
"Wang",
"Yequan",
""
],
[
"Lin",
"Yonghua",
""
],
[
"Qin",
"Yong",
""
]
] |
2505.23019 | Zeying Gong | Zeying Gong, Rong Li, Tianshuai Hu, Ronghe Qiu, Lingdong Kong, Lingfeng Zhang, Yiyi Ding, Leying Zhang, Junwei Liang | Stairway to Success: Zero-Shot Floor-Aware Object-Goal Navigation via LLM-Driven Coarse-to-Fine Exploration | 34 pages, 12 figures, 10 tables | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Object-Goal Navigation (OGN) remains challenging in real-world, multi-floor environments and under open-vocabulary object descriptions. We observe that most episodes in widely used benchmarks such as HM3D and MP3D involve multi-floor buildings, with many requiring explicit floor transitions. However, existing methods are often limited to single-floor settings or predefined object categories. To address these limitations, we tackle two key challenges: (1) efficient cross-level planning and (2) zero-shot object-goal navigation (ZS-OGN), where agents must interpret novel object descriptions without prior exposure. We propose ASCENT, a framework that combines a Multi-Floor Spatial Abstraction module for hierarchical semantic mapping and a Coarse-to-Fine Frontier Reasoning module leveraging Large Language Models (LLMs) for context-aware exploration, without requiring additional training on new object semantics or locomotion data. Our method outperforms state-of-the-art ZS-OGN approaches on HM3D and MP3D benchmarks while enabling efficient multi-floor navigation. We further validate its practicality through real-world deployment on a quadruped robot, achieving successful object exploration across unseen floors. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:00:03 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Gong",
"Zeying",
""
],
[
"Li",
"Rong",
""
],
[
"Hu",
"Tianshuai",
""
],
[
"Qiu",
"Ronghe",
""
],
[
"Kong",
"Lingdong",
""
],
[
"Zhang",
"Lingfeng",
""
],
[
"Ding",
"Yiyi",
""
],
[
"Zhang",
"Leying",
""
],
[
"Liang",
"Junwei",
""
]
] |
2505.23020 | Jinchuan Zhang | Jinchuan Zhang, Lu Yin, Yan Zhou, Songlin Hu | AgentAlign: Navigating Safety Alignment in the Shift from Informative to Agentic Large Language Models | Submitted to ACL 2025 | null | null | null | cs.CR cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The acquisition of agentic capabilities has transformed LLMs from "knowledge providers" to "action executors", a trend that while expanding LLMs' capability boundaries, significantly increases their susceptibility to malicious use. Previous work has shown that current LLM-based agents execute numerous malicious tasks even without being attacked, indicating a deficiency in agentic use safety alignment during the post-training phase. To address this gap, we propose AgentAlign, a novel framework that leverages abstract behavior chains as a medium for safety alignment data synthesis. By instantiating these behavior chains in simulated environments with diverse tool instances, our framework enables the generation of highly authentic and executable instructions while capturing complex multi-step dynamics. The framework further ensures model utility by proportionally synthesizing benign instructions through non-malicious interpretations of behavior chains, precisely calibrating the boundary between helpfulness and harmlessness. Evaluation results on AgentHarm demonstrate that fine-tuning three families of open-source models using our method substantially improves their safety (35.8% to 79.5% improvement) while minimally impacting or even positively enhancing their helpfulness, outperforming various prompting methods. The dataset and code have both been open-sourced. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:02:18 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Zhang",
"Jinchuan",
""
],
[
"Yin",
"Lu",
""
],
[
"Zhou",
"Yan",
""
],
[
"Hu",
"Songlin",
""
]
] |
2505.23021 | Victor Mukherjee | Mrutyunjaya Sahoo, Rahul Ghosh, Bandita Das, Shishira Mahunta, Bodhaditya Santra, and Victor Mukherjee | Generating discrete time crystals through optimal control | 5 pages, 4 figures | null | null | null | quant-ph cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | In this work we use optimal control to generate Discrete Time Crystals (DTC) in generic many-body quantum systems. We define appropriate cost functions, which, when optimized, result in the formation of DTCs. This hitherto unexplored method represents DTCs as an optimization problem, and allows us to find non-trivial realistic periodic control pulses and parameter regimes which result in spontaneous breaking of time-translational symmetry in quantum systems. We exemplify our approach using many-body quantum systems in the presence, as well as absence of dissipation. We also discuss possible experimental realization of the control protocol for generating DTCs. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:02:54 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Sahoo",
"Mrutyunjaya",
""
],
[
"Ghosh",
"Rahul",
""
],
[
"Das",
"Bandita",
""
],
[
"Mahunta",
"Shishira",
""
],
[
"Santra",
"Bodhaditya",
""
],
[
"Mukherjee",
"Victor",
""
]
] |
2505.23022 | Yinghao Tang | Yinghao Tang, Tingfeng Lan, Xiuqi Huang, Hui Lu, Wei Chen | SCORPIO: Serving the Right Requests at the Right Time for Heterogeneous SLOs in LLM Inference | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Existing Large Language Model (LLM) serving systems prioritize maximum throughput. They often neglect Service Level Objectives (SLOs) such as Time to First Token (TTFT) and Time Per Output Token (TPOT), which leads to suboptimal SLO attainment. This paper introduces SCORPIO, an SLO-oriented LLM serving system designed to maximize system goodput and SLO attainment for workloads with heterogeneous SLOs. Our core insight is to exploit SLO heterogeneity for adaptive scheduling across admission control, queue management, and batch selection. SCORPIO features a TTFT Guard, which employs least-deadline-first reordering and rejects unattainable requests, and a TPOT Guard, which utilizes a VBS-based admission control and a novel credit-based batching mechanism. Both guards are supported by a predictive module. Evaluations demonstrate that SCORPIO improves system goodput by up to 14.4X and SLO adherence by up to 46.5% compared to state-of-the-art baselines. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:03:09 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Tang",
"Yinghao",
""
],
[
"Lan",
"Tingfeng",
""
],
[
"Huang",
"Xiuqi",
""
],
[
"Lu",
"Hui",
""
],
[
"Chen",
"Wei",
""
]
] |
2505.23023 | Jack Kendrick | Jack Kendrick | Density Estimation on Rectifiable Sets | null | null | null | null | math.ST math.CA stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Kernel density estimation is a popular method for estimating unseen probability distributions. However, the convergence of these classical estimators to the true density slows down in high dimensions. Moreover, they do not define meaningful probability distributions when the intrinsic dimension of data is much smaller than its ambient dimension. We build on previous work on density estimation on manifolds to show that a modified kernel density estimator converges to the true density on $d-$rectifiable sets. As a special case, we consider algebraic varieties and semi-algebraic sets and prove a convergence rate in this setting. We conclude the paper with a numerical experiment illustrating the convergence of this estimator on sparse data. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:06:40 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Kendrick",
"Jack",
""
]
] |
2505.23024 | Zhihao Wang | Zhihao Wang, Wenke Huang, Tian Chen, Zekun Shi, Guancheng Wan, Yu Qiao, Bin Yang, Jian Wang, Bing Li, Mang Ye | An Empirical Study of Federated Prompt Learning for Vision Language Model | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Vision Language Model (VLM) excels in aligning vision and language representations, and prompt learning has emerged as a key technique for adapting such models to downstream tasks. However, the application of prompt learning with VLM in federated learning (\fl{}) scenarios remains underexplored. This paper systematically investigates the behavioral differences between language prompt learning (LPT) and vision prompt learning (VPT) under data heterogeneity challenges, including label skew and domain shift. We conduct extensive experiments to evaluate the impact of various \fl{} and prompt configurations, such as client scale, aggregation strategies, and prompt length, to assess the robustness of Federated Prompt Learning (FPL). Furthermore, we explore strategies for enhancing prompt learning in complex scenarios where label skew and domain shift coexist, including leveraging both prompt types when computational resources allow. Our findings offer practical insights into optimizing prompt learning in federated settings, contributing to the broader deployment of VLMs in privacy-preserving environments. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:09:15 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Wang",
"Zhihao",
""
],
[
"Huang",
"Wenke",
""
],
[
"Chen",
"Tian",
""
],
[
"Shi",
"Zekun",
""
],
[
"Wan",
"Guancheng",
""
],
[
"Qiao",
"Yu",
""
],
[
"Yang",
"Bin",
""
],
[
"Wang",
"Jian",
""
],
[
"Li",
"Bing",
""
],
[
"Ye",
"Mang",
""
]
] |
2505.23025 | Geyue Sun | Geyue Sun, Xiao Liu, Tomas Williams, Roberto Samaniego | Learning to Regulate: A New Event-Level Dataset of Capital Control Measures | Work in Progress | null | null | null | econ.GN q-fin.EC | http://creativecommons.org/licenses/by/4.0/ | We construct a novel event-level Capital Control Measures (CCM) dataset covering 196 countries from 1999 to 2023 by leveraging prompt-based large language models (LLMs). The dataset enables event study analysis and cross-country comparisons based on rich policy attributes, including action type, intensity, direction, implementing entity, and other multidimensional characteristics. Using a two-step prompt framework with GPT-4.1, we extract structured information from the IMF's Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER), resulting in 5,198 capital control events with 27 annotated fields and corresponding model reasoning. Secondly, to facilitate real-time classification and extension to external sources, we fine-tune an open-source Meta Llama 3.1-8B model, named CCM-Llama, trained on AREAER change logs and final status reports. The model achieves 90.09\% accuracy in category classification and 99.55\% in status prediction. Finally, we apply the CCM dataset in an empirical application: an event study on China, Australia, and the US. The results show that inward capital control measures significantly reduce fund inflows within one month, and restrictive policies tend to have stronger effects than liberalizing ones, with notable heterogeneity across countries. Our work contributes to the growing literature on the use of LLMs in economics by providing both a novel high-frequency policy dataset and a replicable framework for automated classification of capital control events from diverse and evolving information sources. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:11:46 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Sun",
"Geyue",
""
],
[
"Liu",
"Xiao",
""
],
[
"Williams",
"Tomas",
""
],
[
"Samaniego",
"Roberto",
""
]
] |
2505.23026 | Gyubin Choi | Haewon Park, Gyubin Choi, Minjun Kim, Yohan Jo | Context Robust Knowledge Editing for Language Models | ACL 2025 Findings. Our code and datasets are available at (https://github.com/holi-lab/CoRE) | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Knowledge editing (KE) methods offer an efficient way to modify knowledge in large language models. Current KE evaluations typically assess editing success by considering only the edited knowledge without any preceding contexts. In real-world applications, however, preceding contexts often trigger the retrieval of the original knowledge and undermine the intended edit. To address this issue, we develop CHED -- a benchmark designed to evaluate the context robustness of KE methods. Evaluations on CHED show that they often fail when preceding contexts are present. To mitigate this shortcoming, we introduce CoRE, a KE method designed to strengthen context robustness by minimizing context-sensitive variance in hidden states of the model for edited knowledge. This method not only improves the editing success rate in situations where a preceding context is present but also preserves the overall capabilities of the model. We provide an in-depth analysis of the differing impacts of preceding contexts when introduced as user utterances versus assistant responses, and we dissect attention-score patterns to assess how specific tokens influence editing success. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:11:53 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Park",
"Haewon",
""
],
[
"Choi",
"Gyubin",
""
],
[
"Kim",
"Minjun",
""
],
[
"Jo",
"Yohan",
""
]
] |
2505.23027 | Nguyen Nhat Minh To | Minh Nguyen Nhat To, Paul F RWilson, Viet Nguyen, Mohamed Harmanani, Michael Cooper, Fahimeh Fooladgar, Purang Abolmaesumi, Parvin Mousavi, Rahul G. Krishnan | Diverse Prototypical Ensembles Improve Robustness to Subpopulation Shift | ICML 2025 Paper | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | The subpopulationtion shift, characterized by a disparity in subpopulation distributibetween theween the training and target datasets, can significantly degrade the performance of machine learning models. Current solutions to subpopulation shift involve modifying empirical risk minimization with re-weighting strategies to improve generalization. This strategy relies on assumptions about the number and nature of subpopulations and annotations on group membership, which are unavailable for many real-world datasets. Instead, we propose using an ensemble of diverse classifiers to adaptively capture risk associated with subpopulations. Given a feature extractor network, we replace its standard linear classification layer with a mixture of prototypical classifiers, where each member is trained to classify the data while focusing on different features and samples from other members. In empirical evaluation on nine real-world datasets, covering diverse domains and kinds of subpopulation shift, our method of Diverse Prototypical Ensembles (DPEs) often outperforms the prior state-of-the-art in worst-group accuracy. The code is available at https://github.com/minhto2802/dpe4subpop | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:12:56 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"To",
"Minh Nguyen Nhat",
""
],
[
"RWilson",
"Paul F",
""
],
[
"Nguyen",
"Viet",
""
],
[
"Harmanani",
"Mohamed",
""
],
[
"Cooper",
"Michael",
""
],
[
"Fooladgar",
"Fahimeh",
""
],
[
"Abolmaesumi",
"Purang",
""
],
[
"Mousavi",
"Parvin",
""
],
[
"Krishnan",
"Rahul G.",
""
]
] |
2505.23028 | Anqi Mu | Anqi Mu, Nathan Ng, Andrew J. Millis, David R. Reichman | Beyond mean-field dynamics of the Dicke model with non-Markovian dephasing | 7 pages, 3 figures plus supplementary material | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a density matrix-based time dependent projection operator formalism to calculate the beyond mean-field dynamics of systems with non-Markovian local baths and one-to-all interactions. Such models encapsulate the physics of condensed phase systems immersed in optical cavities. We use this method, combined with tensor network influence functionals, to study the dynamics of the Dicke model coupled to non-Markovian local dephasing baths at zero temperature, which has a superradiant phase transition in the mean-field limit. The method corrects a spurious initial state dependence found in the mean-field dynamics and describes the emergence of new time scales which are absent in the mean-field dynamics. Our formalism, based on density matrices, is applicable to other quantum optical systems with one-to-all interactions at finite temperatures. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:13:47 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Mu",
"Anqi",
""
],
[
"Ng",
"Nathan",
""
],
[
"Millis",
"Andrew J.",
""
],
[
"Reichman",
"David R.",
""
]
] |
2505.23029 | Si Wu | Si Wu, Sebastian Bruch | Uncovering Visual-Semantic Psycholinguistic Properties from the Distributional Structure of Text Embedding Spac | Accepted for ACL 2025. This is the camera-ready version. Will be presenting in July 2025 in Vienna | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Imageability (potential of text to evoke a mental image) and concreteness (perceptibility of text) are two psycholinguistic properties that link visual and semantic spaces. It is little surprise that computational methods that estimate them do so using parallel visual and semantic spaces, such as collections of image-caption pairs or multi-modal models. In this paper, we work on the supposition that text itself in an image-caption dataset offers sufficient signals to accurately estimate these properties. We hypothesize, in particular, that the peakedness of the neighborhood of a word in the semantic embedding space reflects its degree of imageability and concreteness. We then propose an unsupervised, distribution-free measure, which we call Neighborhood Stability Measure (NSM), that quantifies the sharpness of peaks. Extensive experiments show that NSM correlates more strongly with ground-truth ratings than existing unsupervised methods, and is a strong predictor of these properties for classification. Our code and data are available on GitHub (https://github.com/Artificial-Memory-Lab/imageability). | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:14:11 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Wu",
"Si",
""
],
[
"Bruch",
"Sebastian",
""
]
] |
2505.23030 | Shruti Hegde | Shruti Hegde, Mabon Manoj Ninan, Jonathan R. Dillman, Shireen Hayatghaibi, Lynn Babcock, Elanchezhian Somasundaram | Can Modern NLP Systems Reliably Annotate Chest Radiography Exams? A Pre-Purchase Evaluation and Comparative Study of Solutions from AWS, Google, Azure, John Snow Labs, and Open-Source Models on an Independent Pediatric Dataset | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | General-purpose clinical natural language processing (NLP) tools are increasingly used for the automatic labeling of clinical reports. However, independent evaluations for specific tasks, such as pediatric chest radiograph (CXR) report labeling, are limited. This study compares four commercial clinical NLP systems - Amazon Comprehend Medical (AWS), Google Healthcare NLP (GC), Azure Clinical NLP (AZ), and SparkNLP (SP) - for entity extraction and assertion detection in pediatric CXR reports. Additionally, CheXpert and CheXbert, two dedicated chest radiograph report labelers, were evaluated on the same task using CheXpert-defined labels. We analyzed 95,008 pediatric CXR reports from a large academic pediatric hospital. Entities and assertion statuses (positive, negative, uncertain) from the findings and impression sections were extracted by the NLP systems, with impression section entities mapped to 12 disease categories and a No Findings category. CheXpert and CheXbert extracted the same 13 categories. Outputs were compared using Fleiss Kappa and accuracy against a consensus pseudo-ground truth. Significant differences were found in the number of extracted entities and assertion distributions across NLP systems. SP extracted 49,688 unique entities, GC 16,477, AZ 31,543, and AWS 27,216. Assertion accuracy across models averaged around 62%, with SP highest (76%) and AWS lowest (50%). CheXpert and CheXbert achieved 56% accuracy. Considerable variability in performance highlights the need for careful validation and review before deploying NLP tools for clinical report labeling. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:16:18 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Hegde",
"Shruti",
""
],
[
"Ninan",
"Mabon Manoj",
""
],
[
"Dillman",
"Jonathan R.",
""
],
[
"Hayatghaibi",
"Shireen",
""
],
[
"Babcock",
"Lynn",
""
],
[
"Somasundaram",
"Elanchezhian",
""
]
] |
2505.23031 | Jinyi Chang | Jinyi Chang, Dongliang Chang, Lei Chen, Bingyao Yu, Zhanyu Ma | Towards Privacy-Preserving Fine-Grained Visual Classification via Hierarchical Learning from Label Proportions | 10 pages, 5 figures, 5 tables | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | In recent years, Fine-Grained Visual Classification (FGVC) has achieved impressive recognition accuracy, despite minimal inter-class variations. However, existing methods heavily rely on instance-level labels, making them impractical in privacy-sensitive scenarios such as medical image analysis. This paper aims to enable accurate fine-grained recognition without direct access to instance labels. To achieve this, we leverage the Learning from Label Proportions (LLP) paradigm, which requires only bag-level labels for efficient training. Unlike existing LLP-based methods, our framework explicitly exploits the hierarchical nature of fine-grained datasets, enabling progressive feature granularity refinement and improving classification accuracy. We propose Learning from Hierarchical Fine-Grained Label Proportions (LHFGLP), a framework that incorporates Unrolled Hierarchical Fine-Grained Sparse Dictionary Learning, transforming handcrafted iterative approximation into learnable network optimization. Additionally, our proposed Hierarchical Proportion Loss provides hierarchical supervision, further enhancing classification performance. Experiments on three widely-used fine-grained datasets, structured in a bag-based manner, demonstrate that our framework consistently outperforms existing LLP-based methods. We will release our code and datasets to foster further research in privacy-preserving fine-grained classification. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:18:25 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Chang",
"Jinyi",
""
],
[
"Chang",
"Dongliang",
""
],
[
"Chen",
"Lei",
""
],
[
"Yu",
"Bingyao",
""
],
[
"Ma",
"Zhanyu",
""
]
] |
2505.23032 | Dongwoo Lee | Dongwoo Lee, Dong Bok Lee, Steven Adriaensen, Juho Lee, Sung Ju Hwang, Frank Hutter, Seon Joo Kim, Hae Beom Lee | Bayesian Neural Scaling Laws Extrapolation with Prior-Fitted Networks | Accepted to ICML 2025 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Scaling has been a major driver of recent advancements in deep learning. Numerous empirical studies have found that scaling laws often follow the power-law and proposed several variants of power-law functions to predict the scaling behavior at larger scales. However, existing methods mostly rely on point estimation and do not quantify uncertainty, which is crucial for real-world applications involving decision-making problems such as determining the expected performance improvements achievable by investing additional computational resources. In this work, we explore a Bayesian framework based on Prior-data Fitted Networks (PFNs) for neural scaling law extrapolation. Specifically, we design a prior distribution that enables the sampling of infinitely many synthetic functions resembling real-world neural scaling laws, allowing our PFN to meta-learn the extrapolation. We validate the effectiveness of our approach on real-world neural scaling laws, comparing it against both the existing point estimation methods and Bayesian approaches. Our method demonstrates superior performance, particularly in data-limited scenarios such as Bayesian active learning, underscoring its potential for reliable, uncertainty-aware extrapolation in practical applications. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:19:17 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Lee",
"Dongwoo",
""
],
[
"Lee",
"Dong Bok",
""
],
[
"Adriaensen",
"Steven",
""
],
[
"Lee",
"Juho",
""
],
[
"Hwang",
"Sung Ju",
""
],
[
"Hutter",
"Frank",
""
],
[
"Kim",
"Seon Joo",
""
],
[
"Lee",
"Hae Beom",
""
]
] |
2505.23033 | Sanchita Ghosh | Sanchita Ghosh, Tanushree Roy | Detecting Switching Attacks On Traffic Flow Regulation For Changing Driving Patterns | 17 pages, 5 figures, 3 tables | null | null | null | eess.SY cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern traffic management systems increasingly adopt hierarchical control strategies for improved efficiency and scalability, where a local traffic controller mode is chosen by a supervisory controller based on the changing large-scale driving patterns. Unfortunately, such local metering controllers are also vulnerable to cyberattacks that can disrupt the controller switching, leading to undesired, inefficient, and even unsafe traffic operations. Additionally, the detection of such attacks becomes challenging when the operational mode of the traffic is uncertain and the operational mode identification is delayed. Thus, in this work, we propose a cyberattack detection scheme to detect the compromised controller switching in ramp metering for an uncertain, multimodal macroscopic traffic operation of a freeway segment. In particular, we propose a bank of detectors corresponding to each admissible traffic mode that can compensate for the uncertain traffic mode of the freeway. Furthermore, we utilize backstepping tools along with Lyapunov function theory to achieve analytical performance guarantees for the detector, such as nominal exponential stability, anomaly/uncertainty-to-residual stability, robustness, and sensitivity. Finally, we demonstrate the efficacy of the proposed detection scheme through simulations of free traffic under realistic traffic parameters, uncertainties, and commonly occurring attack scenarios. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:19:39 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Ghosh",
"Sanchita",
""
],
[
"Roy",
"Tanushree",
""
]
] |
2505.23034 | Guangyi Liu | Guangyi Liu, Yongqi Zhang, Xunyuan Liu, Quanming Yao | Case-Based Reasoning Enhances the Predictive Power of LLMs in Drug-Drug Interaction | null | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Drug-drug interaction (DDI) prediction is critical for treatment safety. While large language models (LLMs) show promise in pharmaceutical tasks, their effectiveness in DDI prediction remains challenging. Inspired by the well-established clinical practice where physicians routinely reference similar historical cases to guide their decisions through case-based reasoning (CBR), we propose CBR-DDI, a novel framework that distills pharmacological principles from historical cases to improve LLM reasoning for DDI tasks. CBR-DDI constructs a knowledge repository by leveraging LLMs to extract pharmacological insights and graph neural networks (GNNs) to model drug associations. A hybrid retrieval mechanism and dual-layer knowledge-enhanced prompting allow LLMs to effectively retrieve and reuse relevant cases. We further introduce a representative sampling strategy for dynamic case refinement. Extensive experiments demonstrate that CBR-DDI achieves state-of-the-art performance, with a significant 28.7% accuracy improvement over both popular LLMs and CBR baseline, while maintaining high interpretability and flexibility. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:20:53 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Liu",
"Guangyi",
""
],
[
"Zhang",
"Yongqi",
""
],
[
"Liu",
"Xunyuan",
""
],
[
"Yao",
"Quanming",
""
]
] |
2505.23035 | Hyunwoo Kim | Hyunwoo Kim, Hanau Yi | Machine-Facing English: Defining a Hybrid Register Shaped by Human-AI Discourse | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Machine-Facing English (MFE) is an emergent register shaped by the adaptation of everyday language to the expanding presence of AI interlocutors. Drawing on register theory (Halliday 1985, 2006), enregisterment (Agha 2003), audience design (Bell 1984), and interactional pragmatics (Giles & Ogay 2007), this study traces how sustained human-AI interaction normalizes syntactic rigidity, pragmatic simplification, and hyper-explicit phrasing - features that enhance machine parseability at the expense of natural fluency. Our analysis is grounded in qualitative observations from bilingual (Korean/English) voice- and text-based product testing sessions, with reflexive drafting conducted using Natural Language Declarative Prompting (NLD-P) under human curation. Thematic analysis identifies five recurrent traits - redundant clarity, directive syntax, controlled vocabulary, flattened prosody, and single-intent structuring - that improve execution accuracy but compress expressive range. MFE's evolution highlights a persistent tension between communicative efficiency and linguistic richness, raising design challenges for conversational interfaces and pedagogical considerations for multilingual users. We conclude by underscoring the need for comprehensive methodological exposition and future empirical validation. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:22:39 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Kim",
"Hyunwoo",
""
],
[
"Yi",
"Hanau",
""
]
] |
2505.23036 | Dai Yuhang | Yuhang Dai, He Wang, Xingchen Li, Zihan Zhang, Shuiyuan Wang, Lei Xie, Xin Xu, Hongxiao Guo, Shaoji Zhang, Hui Bu, Wei Chen | AISHELL-5: The First Open-Source In-Car Multi-Channel Multi-Speaker Speech Dataset for Automatic Speech Diarization and Recognition | 5 pages, 1 figures, 3 tables, accepted by InterSpeech 2025 | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper delineates AISHELL-5, the first open-source in-car multi-channel multi-speaker Mandarin automatic speech recognition (ASR) dataset. AISHLL-5 includes two parts: (1) over 100 hours of multi-channel speech data recorded in an electric vehicle across more than 60 real driving scenarios. This audio data consists of four far-field speech signals captured by microphones located on each car door, as well as near-field signals obtained from high-fidelity headset microphones worn by each speaker. (2) a collection of 40 hours of real-world environmental noise recordings, which supports the in-car speech data simulation. Moreover, we also provide an open-access, reproducible baseline system based on this dataset. This system features a speech frontend model that employs speech source separation to extract each speaker's clean speech from the far-field signals, along with a speech recognition module that accurately transcribes the content of each individual speaker. Experimental results demonstrate the challenges faced by various mainstream ASR models when evaluated on the AISHELL-5. We firmly believe the AISHELL-5 dataset will significantly advance the research on ASR systems under complex driving scenarios by establishing the first publicly available in-car ASR benchmark. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:22:59 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Dai",
"Yuhang",
""
],
[
"Wang",
"He",
""
],
[
"Li",
"Xingchen",
""
],
[
"Zhang",
"Zihan",
""
],
[
"Wang",
"Shuiyuan",
""
],
[
"Xie",
"Lei",
""
],
[
"Xu",
"Xin",
""
],
[
"Guo",
"Hongxiao",
""
],
[
"Zhang",
"Shaoji",
""
],
[
"Bu",
"Hui",
""
],
[
"Chen",
"Wei",
""
]
] |
2505.23037 | Longyin Zhang | Longyin Zhang and Bowei Zou and Ai Ti Aw | Improving Multilingual Social Media Insights: Aspect-based Comment Analysis | The paper was peer-reviewed | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The inherent nature of social media posts, characterized by the freedom of language use with a disjointed array of diverse opinions and topics, poses significant challenges to downstream NLP tasks such as comment clustering, comment summarization, and social media opinion analysis. To address this, we propose a granular level of identifying and generating aspect terms from individual comments to guide model attention. Specifically, we leverage multilingual large language models with supervised fine-tuning for comment aspect term generation (CAT-G), further aligning the model's predictions with human expectations through DPO. We demonstrate the effectiveness of our method in enhancing the comprehension of social media discourse on two NLP tasks. Moreover, this paper contributes the first multilingual CAT-G test set on English, Chinese, Malay, and Bahasa Indonesian. As LLM capabilities vary among languages, this test set allows for a comparative analysis of performance across languages with varying levels of LLM proficiency. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:24:39 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Zhang",
"Longyin",
""
],
[
"Zou",
"Bowei",
""
],
[
"Aw",
"Ai Ti",
""
]
] |
2505.23038 | Yuzhen Xiao | Yuzhen Xiao, Jiahe Song, Yongxin Xu, Ruizhe Zhang, Yiqi Xiao, Xin Lu, Runchuan Zhu, Bowen Jiang, Junfeng Zhao | EL4NER: Ensemble Learning for Named Entity Recognition via Multiple Small-Parameter Large Language Models | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | In-Context Learning (ICL) technique based on Large Language Models (LLMs) has gained prominence in Named Entity Recognition (NER) tasks for its lower computing resource consumption, less manual labeling overhead, and stronger generalizability. Nevertheless, most ICL-based NER methods depend on large-parameter LLMs: the open-source models demand substantial computational resources for deployment and inference, while the closed-source ones incur high API costs, raise data-privacy concerns, and hinder community collaboration. To address this question, we propose an Ensemble Learning Method for Named Entity Recognition (EL4NER), which aims at aggregating the ICL outputs of multiple open-source, small-parameter LLMs to enhance overall performance in NER tasks at less deployment and inference cost. Specifically, our method comprises three key components. First, we design a task decomposition-based pipeline that facilitates deep, multi-stage ensemble learning. Second, we introduce a novel span-level sentence similarity algorithm to establish an ICL demonstration retrieval mechanism better suited for NER tasks. Third, we incorporate a self-validation mechanism to mitigate the noise introduced during the ensemble process. We evaluated EL4NER on multiple widely adopted NER datasets from diverse domains. Our experimental results indicate that EL4NER surpasses most closed-source, large-parameter LLM-based methods at a lower parameter cost and even attains state-of-the-art (SOTA) performance among ICL-based methods on certain datasets. These results show the parameter efficiency of EL4NER and underscore the feasibility of employing open-source, small-parameter LLMs within the ICL paradigm for NER tasks. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:25:14 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Xiao",
"Yuzhen",
""
],
[
"Song",
"Jiahe",
""
],
[
"Xu",
"Yongxin",
""
],
[
"Zhang",
"Ruizhe",
""
],
[
"Xiao",
"Yiqi",
""
],
[
"Lu",
"Xin",
""
],
[
"Zhu",
"Runchuan",
""
],
[
"Jiang",
"Bowen",
""
],
[
"Zhao",
"Junfeng",
""
]
] |
2505.23039 | Jialin Ding | Kapil Vaidya, Jialin Ding, Sebastian Kosak, David Kernert, Chuan Lei, Xiao Qin, Abhinav Tripathy, Ramesh Balan, Balakrishnan Narayanaswamy, Tim Kraska | TailorSQL: An NL2SQL System Tailored to Your Query Workload | null | null | null | null | cs.DB cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | NL2SQL (natural language to SQL) translates natural language questions into SQL queries, thereby making structured data accessible to non-technical users, serving as the foundation for intelligent data applications. State-of-the-art NL2SQL techniques typically perform translation by retrieving database-specific information, such as the database schema, and invoking a pre-trained large language model (LLM) using the question and retrieved information to generate the SQL query.
However, existing NL2SQL techniques miss a key opportunity which is present in real-world settings: NL2SQL is typically applied on existing databases which have already served many SQL queries in the past. The past query workload implicitly contains information which is helpful for accurate NL2SQL translation and is not apparent from the database schema alone, such as common join paths and the semantics of obscurely-named tables and columns. We introduce TailorSQL, a NL2SQL system that takes advantage of information in the past query workload to improve both the accuracy and latency of translating natural language questions into SQL. By specializing to a given workload, TailorSQL achieves up to 2$\times$ improvement in execution accuracy on standardized benchmarks. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:27:22 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Vaidya",
"Kapil",
""
],
[
"Ding",
"Jialin",
""
],
[
"Kosak",
"Sebastian",
""
],
[
"Kernert",
"David",
""
],
[
"Lei",
"Chuan",
""
],
[
"Qin",
"Xiao",
""
],
[
"Tripathy",
"Abhinav",
""
],
[
"Balan",
"Ramesh",
""
],
[
"Narayanaswamy",
"Balakrishnan",
""
],
[
"Kraska",
"Tim",
""
]
] |
2505.23040 | Yihang Wu | Yihang Wu, Muhammad Owais, Reem Kateb, Ahmad Chaddad | Deep Modeling and Optimization of Medical Image Classification | Accepted in ISBI2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep models, such as convolutional neural networks (CNNs) and vision transformer (ViT), demonstrate remarkable performance in image classification. However, those deep models require large data to fine-tune, which is impractical in the medical domain due to the data privacy issue. Furthermore, despite the feasible performance of contrastive language image pre-training (CLIP) in the natural domain, the potential of CLIP has not been fully investigated in the medical field. To face these challenges, we considered three scenarios: 1) we introduce a novel CLIP variant using four CNNs and eight ViTs as image encoders for the classification of brain cancer and skin cancer, 2) we combine 12 deep models with two federated learning techniques to protect data privacy, and 3) we involve traditional machine learning (ML) methods to improve the generalization ability of those deep models in unseen domain data. The experimental results indicate that maxvit shows the highest averaged (AVG) test metrics (AVG = 87.03\%) in HAM10000 dataset with multimodal learning, while convnext\_l demonstrates remarkable test with an F1-score of 83.98\% compared to swin\_b with 81.33\% in FL model. Furthermore, the use of support vector machine (SVM) can improve the overall test metrics with AVG of $\sim 2\%$ for swin transformer series in ISIC2018. Our codes are available at https://github.com/AIPMLab/SkinCancerSimulation. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:27:51 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Wu",
"Yihang",
""
],
[
"Owais",
"Muhammad",
""
],
[
"Kateb",
"Reem",
""
],
[
"Chaddad",
"Ahmad",
""
]
] |
2505.23041 | Huang Zhiqiang | Zhiqiang Huang and Qing-yu Cai | On the generic increase of entropy in isolated systems | 10 pages, 5 figures | null | null | null | quant-ph cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study establishes a universal mechanism for entropy production in isolated quantum systems governed by the eigenstate thermalization hypothesis (ETH). By developing a resolvent-based framework, we demonstrate that steady-state entropy generically arises from many-body interactions, independent of specific coupling details. Analytical arguments reveal that entropy generation is driven by two universal pathways: interaction-induced energy broadening and temporal coarse-graining over exponentially small energy gaps. Numerical simulations of nonintegrable Ising spin chains confirm logarithmic entropy scaling, consistent with predictions derived from ETH-governed eigenstate mixing. The derived self-consistent equations for energy shift and broadening parameters agree closely with numerical results. These results unify observational entropy concepts with von Neumann entropy dynamics, providing predictive tools for thermodynamic behavior in quantum many-body systems. Our findings resolve longstanding debates about interaction-dependent entropy scaling and offer pathways for entropy control in quantum technologies. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:28:27 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Huang",
"Zhiqiang",
""
],
[
"Cai",
"Qing-yu",
""
]
] |
2505.23042 | Siwen Wang | Siwen Wang, Shitou Zhang, Wan-Lin Chen, Dung Truong, Tzyy-Ping Jung | From Theory to Application: Fine-Tuning Large EEG Model with Real-World Stress Data | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in Large Language Models have inspired the development of foundation models across various domains. In this study, we evaluate the efficacy of Large EEG Models (LEMs) by fine-tuning LaBraM, a state-of-the-art foundation EEG model, on a real-world stress classification dataset collected in a graduate classroom. Unlike previous studies that primarily evaluate LEMs using data from controlled clinical settings, our work assesses their applicability to real-world environments. We train a binary classifier that distinguishes between normal and elevated stress states using resting-state EEG data recorded from 18 graduate students during a class session. The best-performing fine-tuned model achieves a balanced accuracy of 90.47% with a 5-second window, significantly outperforming traditional stress classifiers in both accuracy and inference efficiency. We further evaluate the robustness of the fine-tuned LEM under random data shuffling and reduced channel counts. These results demonstrate the capability of LEMs to effectively process real-world EEG data and highlight their potential to revolutionize brain-computer interface applications by shifting the focus from model-centric to data-centric design. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:40:20 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Wang",
"Siwen",
""
],
[
"Zhang",
"Shitou",
""
],
[
"Chen",
"Wan-Lin",
""
],
[
"Truong",
"Dung",
""
],
[
"Jung",
"Tzyy-Ping",
""
]
] |
2505.23043 | Jihai Zhang | Jihai Zhang, Tianle Li, Linjie Li, Zhengyuan Yang, Yu Cheng | Are Unified Vision-Language Models Necessary: Generalization Across Understanding and Generation | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in unified vision-language models (VLMs), which integrate both visual understanding and generation capabilities, have attracted significant attention. The underlying hypothesis is that a unified architecture with mixed training on both understanding and generation tasks can enable mutual enhancement between understanding and generation. However, this hypothesis remains underexplored in prior works on unified VLMs. To address this gap, this paper systematically investigates the generalization across understanding and generation tasks in unified VLMs. Specifically, we design a dataset closely aligned with real-world scenarios to facilitate extensive experiments and quantitative evaluations. We evaluate multiple unified VLM architectures to validate our findings. Our key findings are as follows. First, unified VLMs trained with mixed data exhibit mutual benefits in understanding and generation tasks across various architectures, and this mutual benefits can scale up with increased data. Second, better alignment between multimodal input and output spaces will lead to better generalization. Third, the knowledge acquired during generation tasks can transfer to understanding tasks, and this cross-task generalization occurs within the base language model, beyond modality adapters. Our findings underscore the critical necessity of unifying understanding and generation in VLMs, offering valuable insights for the design and optimization of unified VLMs. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:40:21 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Zhang",
"Jihai",
""
],
[
"Li",
"Tianle",
""
],
[
"Li",
"Linjie",
""
],
[
"Yang",
"Zhengyuan",
""
],
[
"Cheng",
"Yu",
""
]
] |
2505.23044 | Yu Sheng | Yu Sheng, Jiajun Deng, Xinran Zhang, Yu Zhang, Bei Hua, Yanyong Zhang, Jianmin Ji | SpatialSplat: Efficient Semantic 3D from Sparse Unposed Images | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | A major breakthrough in 3D reconstruction is the feedforward paradigm to generate pixel-wise 3D points or Gaussian primitives from sparse, unposed images. To further incorporate semantics while avoiding the significant memory and storage costs of high-dimensional semantic features, existing methods extend this paradigm by associating each primitive with a compressed semantic feature vector. However, these methods have two major limitations: (a) the naively compressed feature compromises expressiveness, affecting the model's ability to capture fine-grained semantics, and (b) the pixel-wise primitive prediction introduces redundancy in overlapping areas, causing unnecessary memory overhead. To this end, we introduce \textbf{SpatialSplat}, a feedforward framework that produces redundancy-aware Gaussians and capitalizes on a dual-field semantic representation. Particularly, with the insight that primitives within the same instance exhibit high semantic consistency, we decompose the semantic representation into a coarse feature field that encodes uncompressed semantics with minimal primitives, and a fine-grained yet low-dimensional feature field that captures detailed inter-instance relationships. Moreover, we propose a selective Gaussian mechanism, which retains only essential Gaussians in the scene, effectively eliminating redundant primitives. Our proposed Spatialsplat learns accurate semantic information and detailed instances prior with more compact 3D Gaussians, making semantic 3D reconstruction more applicable. We conduct extensive experiments to evaluate our method, demonstrating a remarkable 60\% reduction in scene representation parameters while achieving superior performance over state-of-the-art methods. The code will be made available for future investigation. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:40:52 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Sheng",
"Yu",
""
],
[
"Deng",
"Jiajun",
""
],
[
"Zhang",
"Xinran",
""
],
[
"Zhang",
"Yu",
""
],
[
"Hua",
"Bei",
""
],
[
"Zhang",
"Yanyong",
""
],
[
"Ji",
"Jianmin",
""
]
] |
2505.23045 | Chuanhao Li | Chuanhao Li, Wenbo Ye, Zhen Li, Yuwei Wu, Yunde Jia | Multi-Sourced Compositional Generalization in Visual Question Answering | Accepted by IJCAI 2025 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Compositional generalization is the ability of generalizing novel compositions from seen primitives, and has received much attention in vision-and-language (V\&L) recently. Due to the multi-modal nature of V\&L tasks, the primitives composing compositions source from different modalities, resulting in multi-sourced novel compositions. However, the generalization ability over multi-sourced novel compositions, \textit{i.e.}, multi-sourced compositional generalization (MSCG) remains unexplored. In this paper, we explore MSCG in the context of visual question answering (VQA), and propose a retrieval-augmented training framework to enhance the MSCG ability of VQA models by learning unified representations for primitives from different modalities. Specifically, semantically equivalent primitives are retrieved for each primitive in the training samples, and the retrieved features are aggregated with the original primitive to refine the model. This process helps the model learn consistent representations for the same semantic primitives across different modalities. To evaluate the MSCG ability of VQA models, we construct a new GQA-MSCG dataset based on the GQA dataset, in which samples include three types of novel compositions composed of primitives from different modalities. Experimental results demonstrate the effectiveness of the proposed framework. We release GQA-MSCG at https://github.com/NeverMoreLCH/MSCG. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:41:36 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Li",
"Chuanhao",
""
],
[
"Ye",
"Wenbo",
""
],
[
"Li",
"Zhen",
""
],
[
"Wu",
"Yuwei",
""
],
[
"Jia",
"Yunde",
""
]
] |
2505.23046 | Anru R. Zhang | Runshi Tang and Julien Chhor and Olga Klopp and Anru R. Zhang | Revisit CP Tensor Decomposition: Statistical Optimality and Fast Convergence | null | null | null | null | stat.ME cs.NA math.NA math.ST stat.ML stat.TH | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Canonical Polyadic (CP) tensor decomposition is a fundamental technique for analyzing high-dimensional tensor data. While the Alternating Least Squares (ALS) algorithm is widely used for computing CP decomposition due to its simplicity and empirical success, its theoretical foundation, particularly regarding statistical optimality and convergence behavior, remain underdeveloped, especially in noisy, non-orthogonal, and higher-rank settings.
In this work, we revisit CP tensor decomposition from a statistical perspective and provide a comprehensive theoretical analysis of ALS under a signal-plus-noise model. We establish non-asymptotic, minimax-optimal error bounds for tensors of general order, dimensions, and rank, assuming suitable initialization. To enable such initialization, we propose Tucker-based Approximation with Simultaneous Diagonalization (TASD), a robust method that improves stability and accuracy in noisy regimes. Combined with ALS, TASD yields a statistically consistent estimator. We further analyze the convergence dynamics of ALS, identifying a two-phase pattern-initial quadratic convergence followed by linear refinement. We further show that in the rank-one setting, ALS with an appropriately chosen initialization attains optimal error within just one or two iterations. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:42:03 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Tang",
"Runshi",
""
],
[
"Chhor",
"Julien",
""
],
[
"Klopp",
"Olga",
""
],
[
"Zhang",
"Anru R.",
""
]
] |
2505.23047 | John Ball | John L. Ball, Shon Mackie, Jacob G. van de Lindt, Willow Morrissey, Artur Perevalov, Zachary Short, Nicholas Schwartz, Timothy W. Koeth, Brian L. Beaudoin, Carlos A. Romero-Talamas, John Rice, R. Alex Tinguely | Measurements of Fusion Yield on the Centrifugal Mirror Fusion Experiment | Submitted to Nuclear Fusion | null | null | null | physics.plasm-ph | http://creativecommons.org/licenses/by-sa/4.0/ | The Centrifugal Mirror Fusion Experiment (CMFX) at the University of Maryland, College Park is a rotating mirror device that utilizes a central cathode to generate a radial electric field which induces a strongly sheared azimuthal $E\times B$ flow to improve plasma confinement and stability. The fusion yield of CMFX plasmas is assessed by diagnosis of neutron emission for the first time. The total neutron yield is measured with two xylene (EJ-301) and deuterated-xylene (EJ-301D) liquid scintillator detectors absolutely calibrated with an in silico method. A larger xylene scintillator was cross-calibrated and used to measure the time dynamics of the fusion rate under various experimental conditions. A permanently installed $^3$He gas tube detector was independently calibrated with a Cf-252 neutron source to make total yield measurements and provide an independent validation of the scintillator calibration. An interpretive modeling framework was developed using the 0D code MCTrans++ (Schwartz et al 2024 JPP) to infer undiagnosed plasma parameters such as density, temperature, and confinement time. A peak neutron emission rate of 8.4$\times 10^{6}$ $\pm$ 7.0$\times 10^{5}$ was measured (neglecting modeling uncertainties), with an inferred triple product of 1.9~$\times~10^{17}$ $\mathrm{m^{-3}}$ keV s from 0D modeling. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:42:51 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Ball",
"John L.",
""
],
[
"Mackie",
"Shon",
""
],
[
"van de Lindt",
"Jacob G.",
""
],
[
"Morrissey",
"Willow",
""
],
[
"Perevalov",
"Artur",
""
],
[
"Short",
"Zachary",
""
],
[
"Schwartz",
"Nicholas",
""
],
[
"Koeth",
"Timothy W.",
""
],
[
"Beaudoin",
"Brian L.",
""
],
[
"Romero-Talamas",
"Carlos A.",
""
],
[
"Rice",
"John",
""
],
[
"Tinguely",
"R. Alex",
""
]
] |
2505.23048 | Tianci Bu | Tianci Bu, Le Zhou, Wenchuan Yang, Jianhong Mou, Kang Yang, Suoyi Tan, Feng Yao, Jingyuan Wang, Xin Lu | ProDiff: Prototype-Guided Diffusion for Minimal Information Trajectory Imputation | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Trajectory data is crucial for various applications but often suffers from incompleteness due to device limitations and diverse collection scenarios. Existing imputation methods rely on sparse trajectory or travel information, such as velocity, to infer missing points. However, these approaches assume that sparse trajectories retain essential behavioral patterns, which place significant demands on data acquisition and overlook the potential of large-scale human trajectory embeddings. To address this, we propose ProDiff, a trajectory imputation framework that uses only two endpoints as minimal information. It integrates prototype learning to embed human movement patterns and a denoising diffusion probabilistic model for robust spatiotemporal reconstruction. Joint training with a tailored loss function ensures effective imputation. ProDiff outperforms state-of-the-art methods, improving accuracy by 6.28\% on FourSquare and 2.52\% on WuXi. Further analysis shows a 0.927 correlation between generated and real trajectories, demonstrating the effectiveness of our approach. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:43:16 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Bu",
"Tianci",
""
],
[
"Zhou",
"Le",
""
],
[
"Yang",
"Wenchuan",
""
],
[
"Mou",
"Jianhong",
""
],
[
"Yang",
"Kang",
""
],
[
"Tan",
"Suoyi",
""
],
[
"Yao",
"Feng",
""
],
[
"Wang",
"Jingyuan",
""
],
[
"Lu",
"Xin",
""
]
] |
2505.23049 | Tianteng Gu | Tianteng Gu, Bei Liu, Bo Xiao, Ke Zeng, Jiacheng Liu, Yanmin Qian | DenoiseRotator: Enhance Pruning Robustness for LLMs via Importance Concentration | null | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | Pruning is a widely used technique to compress large language models (LLMs) by removing unimportant weights, but it often suffers from significant performance degradation - especially under semi-structured sparsity constraints. Existing pruning methods primarily focus on estimating the importance of individual weights, which limits their ability to preserve critical capabilities of the model. In this work, we propose a new perspective: rather than merely selecting which weights to prune, we first redistribute parameter importance to make the model inherently more amenable to pruning. By minimizing the information entropy of normalized importance scores, our approach concentrates importance onto a smaller subset of weights, thereby enhancing pruning robustness. We instantiate this idea through DenoiseRotator, which applies learnable orthogonal transformations to the model's weight matrices. Our method is model-agnostic and can be seamlessly integrated with existing pruning techniques such as Magnitude, SparseGPT, and Wanda. Evaluated on LLaMA3, Qwen2.5, and Mistral models under 50% unstructured and 2:4 semi-structured sparsity, DenoiseRotator consistently improves perplexity and zero-shot accuracy. For instance, on LLaMA3-70B pruned with SparseGPT at 2:4 semi-structured sparsity, DenoiseRotator reduces the perplexity gap to the dense model by 58%, narrowing the degradation from 8.1 to 3.4 points. Codes are available at https://github.com/Axel-gu/DenoiseRotator. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:44:09 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Gu",
"Tianteng",
""
],
[
"Liu",
"Bei",
""
],
[
"Xiao",
"Bo",
""
],
[
"Zeng",
"Ke",
""
],
[
"Liu",
"Jiacheng",
""
],
[
"Qian",
"Yanmin",
""
]
] |
2505.23050 | Dexing Miao | Dexing Miao, Zijun Xu, Zhiyu Xiang, Pingcheng Liu, Giovanni Ambrosi, Mattia Barbanera, Mengke Cai, Xudong Cai, Hsin-Yi Chou, Matteo Duranti, Valerio Formato, Maria Ionica, Yaozu Jiang, Liangchenglong Jin, Vladimir Koutsenko, Qinze Li, Cong Liu, Xingjian Lv, Alberto Oliva, Wenxi Peng, Rui Qiao, Gianluigi Silvestre, Zibing Wu, Xuhao Yuan, Hongyu Zhang, Xiyuan Zhang, and Jianchun Wang | A Silicon Microstrip Detector for Power-Limited and Large Sensitive Area Applications | 9 pages, 13 figures | null | null | null | physics.ins-det astro-ph.IM hep-ex | http://creativecommons.org/licenses/by/4.0/ | A silicon microstrip detector (SSD) has been developed to have state of the art spatial resolution and a large sensitive area under stringent power constraints. The design incorporates three floating strips with their bias resistors inserted between two aluminum readout strips. Beam test measurements with the single sensor confirmed that this configuration achieves a total detection efficiency of $99.8 \, \%$ and spatial resolution $7.6 \, \mathrm{\mu m}$ for MIPs. A double-$\eta$ algorithm was developed to optimize hit position reconstruction for this SSD. The design can be adapted for large area silicon detectors. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:44:14 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Miao",
"Dexing",
""
],
[
"Xu",
"Zijun",
""
],
[
"Xiang",
"Zhiyu",
""
],
[
"Liu",
"Pingcheng",
""
],
[
"Ambrosi",
"Giovanni",
""
],
[
"Barbanera",
"Mattia",
""
],
[
"Cai",
"Mengke",
""
],
[
"Cai",
"Xudong",
""
],
[
"Chou",
"Hsin-Yi",
""
],
[
"Duranti",
"Matteo",
""
],
[
"Formato",
"Valerio",
""
],
[
"Ionica",
"Maria",
""
],
[
"Jiang",
"Yaozu",
""
],
[
"Jin",
"Liangchenglong",
""
],
[
"Koutsenko",
"Vladimir",
""
],
[
"Li",
"Qinze",
""
],
[
"Liu",
"Cong",
""
],
[
"Lv",
"Xingjian",
""
],
[
"Oliva",
"Alberto",
""
],
[
"Peng",
"Wenxi",
""
],
[
"Qiao",
"Rui",
""
],
[
"Silvestre",
"Gianluigi",
""
],
[
"Wu",
"Zibing",
""
],
[
"Yuan",
"Xuhao",
""
],
[
"Zhang",
"Hongyu",
""
],
[
"Zhang",
"Xiyuan",
""
],
[
"Wang",
"Jianchun",
""
]
] |
2505.23051 | MD Ramiz Aktar | Ramiz Aktar, Kuo-Chuan Pan and Toru Okuda | Global resistive MHD accretion flows around spinning AGNs: impact of resistivity on MAD state | 17 pages, 9 figures, 2 tables; Accepted for publication in ApJ | null | null | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | In this study, we investigate the effect of resistivity on the dynamics of global magnetohydrodynamic accretion flows (Res-MHD) around a spinning supermassive black hole. We perform a comparative study of 2D and 3D resistive models around black holes. We examine accretion flow dynamics considering globally uniform resistivity values, ranging from $\sim 0$ to 0.1. During the simulation time of $t \lesssim 1000~t_g$, we find that the mass accretion rate is comparable for both the 2D and 3D models. However, as the flow becomes increasingly turbulent, non-axisymmetric effects begin to dominate, resulting in significant differences in the mass accretion rates between the 3D and 2D. All the resistive models in a highly magnetized flow belong to the Magnetically Arrested Disk (MAD) state. We propose an efficient and physically motivated approach to examine the magnetic state by estimating the spatial average plasma beta parameter across the computational domain. We find that when the average plasma beta is close to or below unity $( \beta_{\text{ave}} \lesssim 1 )$, the accretion flow enters the MAD state. Additionally, we find that high-resistivity flow reduces magnetorotational instability (MRI) turbulence in the accretion flow, while the turbulence structures remain qualitatively similar in low-resistivity flows. Moreover, we observe indications of plasmoid formations in low-resistivity flow compared to high-resistivity flow. Furthermore, we do not find a clear relationship between the variability of the accretion rate, magnetic flux, and resistivity. Lastly, our findings suggest that low-resistivity models produce higher power jets than those with higher resistivity. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:44:23 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Aktar",
"Ramiz",
""
],
[
"Pan",
"Kuo-Chuan",
""
],
[
"Okuda",
"Toru",
""
]
] |
2505.23052 | Jiarui Zhang | Jiarui Zhang, Xiangyu Liu, Yong Hu, Chaoyue Niu, Fan Wu, Guihai Chen | Query Routing for Retrieval-Augmented Language Models | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Retrieval-Augmented Generation (RAG) significantly improves the performance of Large Language Models (LLMs) on knowledge-intensive tasks. However, varying response quality across LLMs under RAG necessitates intelligent routing mechanisms, which select the most suitable model for each query from multiple retrieval-augmented LLMs via a dedicated router model. We observe that external documents dynamically affect LLMs' ability to answer queries, while existing routing methods, which rely on static parametric knowledge representations, exhibit suboptimal performance in RAG scenarios. To address this, we formally define the new retrieval-augmented LLM routing problem, incorporating the influence of retrieved documents into the routing framework. We propose RAGRouter, a RAG-aware routing design, which leverages document embeddings and RAG capability embeddings with contrastive learning to capture knowledge representation shifts and enable informed routing decisions. Extensive experiments on diverse knowledge-intensive tasks and retrieval settings show that RAGRouter outperforms the best individual LLM by 3.61% on average and existing routing methods by 3.29%-9.33%. With an extended score-threshold-based mechanism, it also achieves strong performance-efficiency trade-offs under low-latency constraints. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:44:56 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Zhang",
"Jiarui",
""
],
[
"Liu",
"Xiangyu",
""
],
[
"Hu",
"Yong",
""
],
[
"Niu",
"Chaoyue",
""
],
[
"Wu",
"Fan",
""
],
[
"Chen",
"Guihai",
""
]
] |
2505.23053 | Wei-Hsiang Huang | Wei-Hsiang Huang, Chen-Wei Ke, Wei-Ning Chiu, Yu-Xuan Su, Chun-Chun Yang, Chieh-Yuan Cheng, Yun-Nung Chen, Pu-Jen Cheng | Augment or Not? A Comparative Study of Pure and Augmented Large Language Model Recommenders | null | null | null | null | cs.IR cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have introduced new paradigms for recommender systems by enabling richer semantic understanding and incorporating implicit world knowledge. In this study, we propose a systematic taxonomy that classifies existing approaches into two categories: (1) Pure LLM Recommenders, which rely solely on LLMs, and (2) Augmented LLM Recommenders, which integrate additional non-LLM techniques to enhance performance. This taxonomy provides a novel lens through which to examine the evolving landscape of LLM-based recommendation. To support fair comparison, we introduce a unified evaluation platform that benchmarks representative models under consistent experimental settings, highlighting key design choices that impact effectiveness. We conclude by discussing open challenges and outlining promising directions for future research. This work offers both a comprehensive overview and practical guidance for advancing next-generation LLM-powered recommender. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:50:24 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Huang",
"Wei-Hsiang",
""
],
[
"Ke",
"Chen-Wei",
""
],
[
"Chiu",
"Wei-Ning",
""
],
[
"Su",
"Yu-Xuan",
""
],
[
"Yang",
"Chun-Chun",
""
],
[
"Cheng",
"Chieh-Yuan",
""
],
[
"Chen",
"Yun-Nung",
""
],
[
"Cheng",
"Pu-Jen",
""
]
] |
2505.23054 | Yuxuan Lin | Yuxuan Lin, Ruihang Chu, Zhenyu Chen, Xiao Tang, Lei Ke, Haoling Li, Yingji Zhong, Zhihao Li, Shiyong Liu, Xiaofei Wu, Jianzhuang Liu and Yujiu Yang | Zero-P-to-3: Zero-Shot Partial-View Images to 3D Object | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Generative 3D reconstruction shows strong potential in incomplete observations. While sparse-view and single-image reconstruction are well-researched, partial observation remains underexplored. In this context, dense views are accessible only from a specific angular range, with other perspectives remaining inaccessible. This task presents two main challenges: (i) limited View Range: observations confined to a narrow angular scope prevent effective traditional interpolation techniques that require evenly distributed perspectives. (ii) inconsistent Generation: views created for invisible regions often lack coherence with both visible regions and each other, compromising reconstruction consistency. To address these challenges, we propose \method, a novel training-free approach that integrates the local dense observations and multi-source priors for reconstruction. Our method introduces a fusion-based strategy to effectively align these priors in DDIM sampling, thereby generating multi-view consistent images to supervise invisible views. We further design an iterative refinement strategy, which uses the geometric structures of the object to enhance reconstruction quality. Extensive experiments on multiple datasets show the superiority of our method over SOTAs, especially in invisible regions. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:51:37 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Lin",
"Yuxuan",
""
],
[
"Chu",
"Ruihang",
""
],
[
"Chen",
"Zhenyu",
""
],
[
"Tang",
"Xiao",
""
],
[
"Ke",
"Lei",
""
],
[
"Li",
"Haoling",
""
],
[
"Zhong",
"Yingji",
""
],
[
"Li",
"Zhihao",
""
],
[
"Liu",
"Shiyong",
""
],
[
"Wu",
"Xiaofei",
""
],
[
"Liu",
"Jianzhuang",
""
],
[
"Yang",
"Yujiu",
""
]
] |
2505.23055 | Zhen Xiang | Zhen Xiang, Aliyah R. Hsu, Austin V. Zane, Aaron E. Kornblith, Margaret J. Lin-Martore, Jasmanpreet C. Kaur, Vasuda M. Dokiparthi, Bo Li, Bin Yu | CDR-Agent: Intelligent Selection and Execution of Clinical Decision Rules Using Large Language Model Agents | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Clinical decision-making is inherently complex and fast-paced, particularly in emergency departments (EDs) where critical, rapid and high-stakes decisions are made. Clinical Decision Rules (CDRs) are standardized evidence-based tools that combine signs, symptoms, and clinical variables into decision trees to make consistent and accurate diagnoses. CDR usage is often hindered by the clinician's cognitive load, limiting their ability to quickly recall and apply the appropriate rules. We introduce CDR-Agent, a novel LLM-based system designed to enhance ED decision-making by autonomously identifying and applying the most appropriate CDRs based on unstructured clinical notes. To validate CDR-Agent, we curated two novel ED datasets: synthetic and CDR-Bench, although CDR-Agent is applicable to non ED clinics. CDR-Agent achieves a 56.3\% (synthetic) and 8.7\% (CDR-Bench) accuracy gain relative to the standalone LLM baseline in CDR selection. Moreover, CDR-Agent significantly reduces computational overhead. Using these datasets, we demonstrated that CDR-Agent not only selects relevant CDRs efficiently, but makes cautious yet effective imaging decisions by minimizing unnecessary interventions while successfully identifying most positively diagnosed cases, outperforming traditional LLM prompting approaches. Code for our work can be found at: https://github.com/zhenxianglance/medagent-cdr-agent | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:51:51 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Xiang",
"Zhen",
""
],
[
"Hsu",
"Aliyah R.",
""
],
[
"Zane",
"Austin V.",
""
],
[
"Kornblith",
"Aaron E.",
""
],
[
"Lin-Martore",
"Margaret J.",
""
],
[
"Kaur",
"Jasmanpreet C.",
""
],
[
"Dokiparthi",
"Vasuda M.",
""
],
[
"Li",
"Bo",
""
],
[
"Yu",
"Bin",
""
]
] |
2505.23056 | Zijian Liu | Zijian Liu, Zhengyuan Zhou | Improved Last-Iterate Convergence of Shuffling Gradient Methods for Nonsmooth Convex Optimization | ICML 2025 | null | null | null | math.OC cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the convergence of the shuffling gradient method, a popular algorithm employed to minimize the finite-sum function with regularization, in which functions are passed to apply (Proximal) Gradient Descent (GD) one by one whose order is determined by a permutation on the indices of functions. In contrast to its easy implementation and effective performance in practice, the theoretical understanding remains limited. A recent advance by (Liu & Zhou, 2024b) establishes the first last-iterate convergence results under various settings, especially proving the optimal rates for smooth (strongly) convex optimization. However, their bounds for nonsmooth (strongly) convex functions are only as fast as Proximal GD. In this work, we provide the first improved last-iterate analysis for the nonsmooth case demonstrating that the widely used Random Reshuffle ($\textsf{RR}$) and Single Shuffle ($\textsf{SS}$) strategies are both provably faster than Proximal GD, reflecting the benefit of randomness. As an important implication, we give the first (nearly) optimal convergence result for the suffix average under the $\textsf{RR}$ sampling scheme in the general convex case, matching the lower bound shown by (Koren et al., 2022). | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:53:45 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Liu",
"Zijian",
""
],
[
"Zhou",
"Zhengyuan",
""
]
] |
2505.23057 | Jun Kigami | Jun Kigami and Yuka Ota | Conductive homogeneity of locally symmetric polygon-based self-similar sets | null | null | null | null | math.MG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We provide a rich family of self-similar sets, called locally symmetric polygon-based self-similar sets, as examples of metric spaces having conductive homogeneity, which was introduced as a sufficient condition for the construction of counterparts of "Sobolev spaces" on compact metric spaces. In particular, our results imply the existence of "Brownian motions" on our family of self-similar sets at the same time. Unlike the known examples like the Sierpinski carpet by Barlow-Bass, unconstrained carpet by Cao and Qiu and the Octa-carpet by Andrews, our examples may have no global symmetries, i.e. the group of isometries is trivial. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 03:55:39 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Kigami",
"Jun",
""
],
[
"Ota",
"Yuka",
""
]
] |
2505.23058 | Yutong Xie | Yutong Xie, Zhuoheng Li, Xiyuan Wang, Yijun Pan, Qijia Liu, Xingzhi Cui, Kuang-Yu Lo, Ruoyi Gao, Xingjian Zhang, Jin Huang, Walter Yuan, Matthew O. Jackson, Qiaozhu Mei | Be.FM: Open Foundation Models for Human Behavior | null | null | null | null | cs.AI cs.CE cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Despite their success in numerous fields, the potential of foundation models for modeling and understanding human behavior remains largely unexplored. We introduce Be.FM, one of the first open foundation models designed for human behavior modeling. Built upon open-source large language models and fine-tuned on a diverse range of behavioral data, Be.FM can be used to understand and predict human decision-making. We construct a comprehensive set of benchmark tasks for testing the capabilities of behavioral foundation models. Our results demonstrate that Be.FM can predict behaviors, infer characteristics of individuals and populations, generate insights about contexts, and apply behavioral science knowledge. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:03:51 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Xie",
"Yutong",
""
],
[
"Li",
"Zhuoheng",
""
],
[
"Wang",
"Xiyuan",
""
],
[
"Pan",
"Yijun",
""
],
[
"Liu",
"Qijia",
""
],
[
"Cui",
"Xingzhi",
""
],
[
"Lo",
"Kuang-Yu",
""
],
[
"Gao",
"Ruoyi",
""
],
[
"Zhang",
"Xingjian",
""
],
[
"Huang",
"Jin",
""
],
[
"Yuan",
"Walter",
""
],
[
"Jackson",
"Matthew O.",
""
],
[
"Mei",
"Qiaozhu",
""
]
] |
2505.23059 | Dohyeon Lee | Dohyeon Lee, Yeonseok Jeong, Seung-won Hwang | From Token to Action: State Machine Reasoning to Mitigate Overthinking in Information Retrieval | null | null | null | null | cs.IR cs.AI | http://creativecommons.org/licenses/by/4.0/ | Chain-of-Thought (CoT) prompting enables complex reasoning in large language models (LLMs), including applications in information retrieval (IR). However, it often leads to overthinking, where models produce excessively long and semantically redundant traces with little or no benefit. We identify two key challenges in IR: redundant trajectories that revisit similar states and misguided reasoning that diverges from user intent. To address these, we propose State Machine Reasoning (SMR), a transition-based reasoning framework composed of discrete actions (Refine, Rerank, Stop) that support early stopping and fine-grained control. Experiments on the BEIR and BRIGHT benchmarks show that SMR improves retrieval performance (nDCG@10) by 3.4% while reducing token usage by 74.4%. It generalizes across LLMs and retrievers without requiring task-specific tuning, offering a practical alternative to conventional CoT reasoning. The code and details are available at https://github.com/ldilab/SMR. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:04:25 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Lee",
"Dohyeon",
""
],
[
"Jeong",
"Yeonseok",
""
],
[
"Hwang",
"Seung-won",
""
]
] |
2505.23060 | Jeonghun Cho | Jeonghun Cho, Deokhyung Kang, Hyounghun Kim, Gary Geunbae Lee | Self-Correcting Code Generation Using Small Language Models | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Self-correction has demonstrated potential in code generation by allowing language models to revise and improve their outputs through successive refinement. Recent studies have explored prompting-based strategies that incorporate verification or feedback loops using proprietary models, as well as training-based methods that leverage their strong reasoning capabilities. However, whether smaller models possess the capacity to effectively guide their outputs through self-reflection remains unexplored. Our findings reveal that smaller models struggle to exhibit reflective revision behavior across both self-correction paradigms. In response, we introduce CoCoS, an approach designed to enhance the ability of small language models for multi-turn code correction. Specifically, we propose an online reinforcement learning objective that trains the model to confidently maintain correct outputs while progressively correcting incorrect outputs as turns proceed. Our approach features an accumulated reward function that aggregates rewards across the entire trajectory and a fine-grained reward better suited to multi-turn correction scenarios. This facilitates the model in enhancing initial response quality while achieving substantial improvements through self-correction. With 1B-scale models, CoCoS achieves improvements of 35.8% on the MBPP and 27.7% on HumanEval compared to the baselines. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:04:44 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Cho",
"Jeonghun",
""
],
[
"Kang",
"Deokhyung",
""
],
[
"Kim",
"Hyounghun",
""
],
[
"Lee",
"Gary Geunbae",
""
]
] |
2505.23061 | Debangshu Banerjee | Tarun Suresh, Debangshu Banerjee, Shubham Ugare, Sasa Misailovic, Gagandeep Singh | DINGO: Constrained Inference for Diffusion LLMs | DINGO an algorithm to provably apply constraints to diffusion LLM generations | null | null | null | cs.LG cs.PL cs.SE | http://creativecommons.org/licenses/by/4.0/ | Diffusion LLMs have emerged as a promising alternative to conventional autoregressive LLMs, offering significant potential for improved runtime efficiency. However, existing diffusion models lack the ability to provably enforce user-specified formal constraints, such as regular expressions, which makes them unreliable for tasks that require structured outputs, such as fixed-schema JSON generation. Unlike autoregressive models that generate tokens sequentially, diffusion LLMs predict a block of tokens in parallel. This parallelism makes traditional constrained decoding algorithms, which are designed for sequential token prediction, ineffective at preserving the true output distribution. To address this limitation, we propose DINGO, a dynamic programming-based constrained decoding strategy that is both efficient and provably distribution-preserving. DINGO enables sampling of output strings with the highest probability under the model's predicted distribution, while strictly satisfying any user-specified regular expression. On standard symbolic math and JSON generation benchmarks, DINGO achieves up to a 68 percentage point improvement over unconstrained inference | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:04:54 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Suresh",
"Tarun",
""
],
[
"Banerjee",
"Debangshu",
""
],
[
"Ugare",
"Shubham",
""
],
[
"Misailovic",
"Sasa",
""
],
[
"Singh",
"Gagandeep",
""
]
] |
2505.23062 | Lingkai Kong | Lingkai Kong, Haichuan Wang, Tonghan Wang, Guojun Xiong, Milind Tambe | Composite Flow Matching for Reinforcement Learning with Shifted-Dynamics Data | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Incorporating pre-collected offline data from a source environment can significantly improve the sample efficiency of reinforcement learning (RL), but this benefit is often challenged by discrepancies between the transition dynamics of the source and target environments. Existing methods typically address this issue by penalizing or filtering out source transitions in high dynamics-gap regions. However, their estimation of the dynamics gap often relies on KL divergence or mutual information, which can be ill-defined when the source and target dynamics have disjoint support. To overcome these limitations, we propose CompFlow, a method grounded in the theoretical connection between flow matching and optimal transport. Specifically, we model the target dynamics as a conditional flow built upon the output distribution of the source-domain flow, rather than learning it directly from a Gaussian prior. This composite structure offers two key advantages: (1) improved generalization for learning target dynamics, and (2) a principled estimation of the dynamics gap via the Wasserstein distance between source and target transitions. Leveraging our principled estimation of the dynamics gap, we further introduce an optimistic active data collection strategy that prioritizes exploration in regions of high dynamics gap, and theoretically prove that it reduces the performance disparity with the optimal policy. Empirically, CompFlow outperforms strong baselines across several RL benchmarks with shifted dynamics. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:09:19 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Kong",
"Lingkai",
""
],
[
"Wang",
"Haichuan",
""
],
[
"Wang",
"Tonghan",
""
],
[
"Xiong",
"Guojun",
""
],
[
"Tambe",
"Milind",
""
]
] |
2505.23063 | Denis Mamba Kabala | Denis Mamba Kabala, Adel Hafiane, Laurent Bobelin and Raphael Canals | Loss-Guided Model Sharing and Local Learning Correction in Decentralized Federated Learning for Crop Disease Classification | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Crop disease detection and classification is a critical challenge in agriculture, with major implications for productivity, food security, and environmental sustainability. While deep learning models such as CNN and ViT have shown excellent performance in classifying plant diseases from images, their large-scale deployment is often limited by data privacy concerns. Federated Learning (FL) addresses this issue, but centralized FL remains vulnerable to single-point failures and scalability limits. In this paper, we introduce a novel Decentralized Federated Learning (DFL) framework that uses validation loss (Loss_val) both to guide model sharing between peers and to correct local training via an adaptive loss function controlled by weighting parameter. We conduct extensive experiments using PlantVillage datasets with three deep learning architectures (ResNet50, VGG16, and ViT_B16), analyzing the impact of weighting parameter, the number of shared models, the number of clients, and the use of Loss_val versus Loss_train of other clients. Results demonstrate that our DFL approach not only improves accuracy and convergence speed, but also ensures better generalization and robustness across heterogeneous data environments making it particularly well-suited for privacy-preserving agricultural applications. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:12:53 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Kabala",
"Denis Mamba",
""
],
[
"Hafiane",
"Adel",
""
],
[
"Bobelin",
"Laurent",
""
],
[
"Canals",
"Raphael",
""
]
] |
2505.23064 | Boran Ma | Bradley Lamb, Saroj Upreti, Yunfei Wang, Daniel Struble, Chenhui Zhu, Guillaume Freychet, Xiaodan Gu, and Boran Ma | Machine Learning Framework for Characterizing Processing-Structure Relationship in Block Copolymer Thin Films | null | null | null | null | cond-mat.mtrl-sci cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The morphology of block copolymers (BCPs) critically influences material properties and applications. This work introduces a machine learning (ML)-enabled, high-throughput framework for analyzing grazing incidence small-angle X-ray scattering (GISAXS) data and atomic force microscopy (AFM) images to characterize BCP thin film morphology. A convolutional neural network was trained to classify AFM images by morphology type, achieving 97% testing accuracy. Classified images were then analyzed to extract 2D grain size measurements from the samples in a high-throughput manner. ML models were developed to predict morphological features based on processing parameters such as solvent ratio, additive type, and additive ratio. GISAXS-based properties were predicted with strong performances ($R^2$ > 0.75), while AFM-based property predictions were less accurate ($R^2$ < 0.60), likely due to the localized nature of AFM measurements compared to the bulk information captured by GISAXS. Beyond model performance, interpretability was addressed using Shapley Additive exPlanations (SHAP). SHAP analysis revealed that the additive ratio had the largest impact on morphological predictions, where additive provides the BCP chains with increased volume to rearrange into thermodynamically favorable morphologies. This interpretability helps validate model predictions and offers insight into parameter importance. Altogether, the presented framework combining high-throughput characterization and interpretable ML offers an approach to exploring and optimizing BCP thin film morphology across a broad processing landscape. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:14:42 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Lamb",
"Bradley",
""
],
[
"Upreti",
"Saroj",
""
],
[
"Wang",
"Yunfei",
""
],
[
"Struble",
"Daniel",
""
],
[
"Zhu",
"Chenhui",
""
],
[
"Freychet",
"Guillaume",
""
],
[
"Gu",
"Xiaodan",
""
],
[
"Ma",
"Boran",
""
]
] |
2505.23065 | Hongcheng Guo | Hongcheng Guo, Zheyong Xie, Shaosheng Cao, Boyang Wang, Weiting Liu, Anjie Le, Lei Li, Zhoujun Li | SNS-Bench-VL: Benchmarking Multimodal Large Language Models in Social Networking Services | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the increasing integration of visual and textual content in Social Networking Services (SNS), evaluating the multimodal capabilities of Large Language Models (LLMs) is crucial for enhancing user experience, content understanding, and platform intelligence. Existing benchmarks primarily focus on text-centric tasks, lacking coverage of the multimodal contexts prevalent in modern SNS ecosystems. In this paper, we introduce SNS-Bench-VL, a comprehensive multimodal benchmark designed to assess the performance of Vision-Language LLMs in real-world social media scenarios. SNS-Bench-VL incorporates images and text across 8 multimodal tasks, including note comprehension, user engagement analysis, information retrieval, and personalized recommendation. It comprises 4,001 carefully curated multimodal question-answer pairs, covering single-choice, multiple-choice, and open-ended tasks. We evaluate over 25 state-of-the-art multimodal LLMs, analyzing their performance across tasks. Our findings highlight persistent challenges in multimodal social context comprehension. We hope SNS-Bench-VL will inspire future research towards robust, context-aware, and human-aligned multimodal intelligence for next-generation social networking services. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:16:24 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Guo",
"Hongcheng",
""
],
[
"Xie",
"Zheyong",
""
],
[
"Cao",
"Shaosheng",
""
],
[
"Wang",
"Boyang",
""
],
[
"Liu",
"Weiting",
""
],
[
"Le",
"Anjie",
""
],
[
"Li",
"Lei",
""
],
[
"Li",
"Zhoujun",
""
]
] |
2505.23066 | Jeremiah Deng | Shuyin Xia and Xiaojiang Tian and Suzhen Yuan and Jeremiah D. Deng | Efficient Quantum Approximate $k$NN Algorithm via Granular-Ball Computing | 8 pages; 7 figure; accepted by IJCAI 2025 | null | null | null | quant-ph cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | High time complexity is one of the biggest challenges faced by $k$-Nearest Neighbors ($k$NN). Although current classical and quantum $k$NN algorithms have made some improvements, they still have a speed bottleneck when facing large amounts of data. To address this issue, we propose an innovative algorithm called Granular-Ball based Quantum $k$NN(GB-Q$k$NN). This approach achieves higher efficiency by first employing granular-balls, which reduces the data size needed to processed. The search process is then accelerated by adopting a Hierarchical Navigable Small World (HNSW) method. Moreover, we optimize the time-consuming steps, such as distance calculation, of the HNSW via quantization, further reducing the time complexity of the construct and search process. By combining the use of granular-balls and quantization of the HNSW method, our approach manages to take advantage of these treatments and significantly reduces the time complexity of the $k$NN-like algorithms, as revealed by a comprehensive complexity analysis. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:16:29 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Xia",
"Shuyin",
""
],
[
"Tian",
"Xiaojiang",
""
],
[
"Yuan",
"Suzhen",
""
],
[
"Deng",
"Jeremiah D.",
""
]
] |
2505.23067 | Zhizhang Wang | Rongli Huang, Changzheng Qu, Zhizhang Wang, Weifeng Wo | Second boundary value problem for the Hessian curvature flow | null | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | We investigate the evolution of strictly convex hypersurfaces driven by the $k$-Hessian curvature flow, subject to the second boundary condition. We first explore the translating solutions corresponding to this boundary value problem. Next, we establish the long-time existence of the flow and prove that it converges to a translating solution. To overcome the difficulty of driving boundary $C^2$ estimates, we employ an orthogonal invariance technique. Using this method, we extend the results of Schn\"urer-Smoczyk \cite{Schnurer2003} and Schn\"urer \cite{Schnurer2002} from the second boundary value problem of Gauss curvature flow to $k$-Hessian curvature flow. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:16:40 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Huang",
"Rongli",
""
],
[
"Qu",
"Changzheng",
""
],
[
"Wang",
"Zhizhang",
""
],
[
"Wo",
"Weifeng",
""
]
] |
2505.23068 | Rui Xu | Rui Xu, Yuzhen Niu, Yuezhou Li, Huangbiao Xu, Wenxi Liu, Yuzhong Chen | URWKV: Unified RWKV Model with Multi-state Perspective for Low-light Image Restoration | This paper has been accepted to CVPR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing low-light image enhancement (LLIE) and joint LLIE and deblurring (LLIE-deblur) models have made strides in addressing predefined degradations, yet they are often constrained by dynamically coupled degradations. To address these challenges, we introduce a Unified Receptance Weighted Key Value (URWKV) model with multi-state perspective, enabling flexible and effective degradation restoration for low-light images. Specifically, we customize the core URWKV block to perceive and analyze complex degradations by leveraging multiple intra- and inter-stage states. First, inspired by the pupil mechanism in the human visual system, we propose Luminance-adaptive Normalization (LAN) that adjusts normalization parameters based on rich inter-stage states, allowing for adaptive, scene-aware luminance modulation. Second, we aggregate multiple intra-stage states through exponential moving average approach, effectively capturing subtle variations while mitigating information loss inherent in the single-state mechanism. To reduce the degradation effects commonly associated with conventional skip connections, we propose the State-aware Selective Fusion (SSF) module, which dynamically aligns and integrates multi-state features across encoder stages, selectively fusing contextual information. In comparison to state-of-the-art models, our URWKV model achieves superior performance on various benchmarks, while requiring significantly fewer parameters and computational resources. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:17:09 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Xu",
"Rui",
""
],
[
"Niu",
"Yuzhen",
""
],
[
"Li",
"Yuezhou",
""
],
[
"Xu",
"Huangbiao",
""
],
[
"Liu",
"Wenxi",
""
],
[
"Chen",
"Yuzhong",
""
]
] |
2505.23069 | Matteo Baggioli | Arabinda Bera, Ido Regev, Alessio Zaccone, Matteo Baggioli | Burgers rings as topological signatures of Eshelby-like plastic events in glasses | v1: comments welcome | null | null | null | cond-mat.dis-nn cond-mat.mtrl-sci cond-mat.soft cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | Eshelby-like quadrupolar structures serve as the fundamental microscopic units for characterizing plastic instabilities in amorphous solids and play a crucial role in explaining their mechanical failure, including the formation of shear bands. However, identifying Eshelby-like plastic events in glasses remains challenging due to their inherent structural and dynamical complexity. In this work, we show that Eshelby-like structures can be precisely identified and localized using a topological invariant known as the continuous Burgers vector. By combining analytical and simulation techniques, we reveal the emergence of a topological Burgers ring around Eshelby plastic events, enabling the precise identification of their center of mass and capturing their orientation as well. This proposed method offers a clear and unambiguous framework to locate and characterize the plastic rearrangements that govern plasticity in glasses. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:19:48 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Bera",
"Arabinda",
""
],
[
"Regev",
"Ido",
""
],
[
"Zaccone",
"Alessio",
""
],
[
"Baggioli",
"Matteo",
""
]
] |
2505.23070 | Anjana Wijayawardhana Hemasiri Nawarathna | Anjana Wijayawardhana, David Gunawan, Thomas Suesse | Non-Gaussian Simultaneous Autoregressive Models with Missing Data | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Standard simultaneous autoregressive (SAR) models are usually assumed to have normally distributed errors, an assumption that is often violated in real-world datasets, which are frequently found to exhibit non-normal, skewed, and heavy-tailed characteristics. New SAR models are proposed to capture these non-Gaussian features. In this project, the spatial error model (SEM), a widely used SAR-type model, is considered. Three novel SEMs are introduced that extend the standard Gaussian SEM by incorporating Student's $t$-distributed errors after a one-to-one transformation is applied to the response variable. Variational Bayes (VB) estimation methods are developed for these models, and the framework is further extended to handle missing response data. Standard variational Bayes (VB) methods perform well with complete datasets; however, handling missing data requires a Hybrid VB (HVB) approach, which integrates a Markov chain Monte Carlo (MCMC) sampler to generate missing values. The proposed VB methods are evaluated using both simulated and real-world datasets, demonstrating their robustness and effectiveness in dealing with non-normal data and missing data in spatial models. Although the method is demonstrated using SAR models, the proposed model specifications and estimation approaches are widely applicable to various types of models for handling non-Gaussian data with missing values. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:21:26 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Wijayawardhana",
"Anjana",
""
],
[
"Gunawan",
"David",
""
],
[
"Suesse",
"Thomas",
""
]
] |
2505.23071 | Peizheng Guo | Peizheng Guo, Jingyao Wang, Huijie Guo, Jiangmeng Li, Chuxiong Sun, Changwen Zheng, Wenwen Qiang | Multi-Modal Learning with Bayesian-Oriented Gradient Calibration | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-Modal Learning (MML) integrates information from diverse modalities to improve predictive accuracy. However, existing methods mainly aggregate gradients with fixed weights and treat all dimensions equally, overlooking the intrinsic gradient uncertainty of each modality. This may lead to (i) excessive updates in sensitive dimensions, degrading performance, and (ii) insufficient updates in less sensitive dimensions, hindering learning. To address this issue, we propose BOGC-MML, a Bayesian-Oriented Gradient Calibration method for MML to explicitly model the gradient uncertainty and guide the model optimization towards the optimal direction. Specifically, we first model each modality's gradient as a random variable and derive its probability distribution, capturing the full uncertainty in the gradient space. Then, we propose an effective method that converts the precision (inverse variance) of each gradient distribution into a scalar evidence. This evidence quantifies the confidence of each modality in every gradient dimension. Using these evidences, we explicitly quantify per-dimension uncertainties and fuse them via a reduced Dempster-Shafer rule. The resulting uncertainty-weighted aggregation produces a calibrated update direction that balances sensitivity and conservatism across dimensions. Extensive experiments on multiple benchmark datasets demonstrate the effectiveness and advantages of the proposed method. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:23:22 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Guo",
"Peizheng",
""
],
[
"Wang",
"Jingyao",
""
],
[
"Guo",
"Huijie",
""
],
[
"Li",
"Jiangmeng",
""
],
[
"Sun",
"Chuxiong",
""
],
[
"Zheng",
"Changwen",
""
],
[
"Qiang",
"Wenwen",
""
]
] |
2505.23072 | Takeshi Yoshimura | Takeshi Yoshimura and Tatsuhiro Chiba and Manish Sethi and Daniel Waddington and Swaminathan Sundararaman | Speeding up Model Loading with fastsafetensors | 12 pages, 15 figures, IEEE CLOUD 2025 | null | null | null | cs.DC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The rapid increases in model parameter sizes introduces new challenges in pre-trained model loading. Currently, machine learning code often deserializes each parameter as a tensor object in host memory before copying it to device memory. We found that this approach underutilized storage throughput and significantly slowed down loading large models with a widely-used model file formats, safetensors. In this work, we present fastsafetensors, a Python library designed to optimize the deserialization of tensors in safetensors files. Our approach first copies groups of on-disk parameters to device memory, where they are directly instantiated as tensor objects. This design enables further optimization in low-level I/O and high-level tensor preprocessing, including parallelized copying, peer-to-peer DMA, and GPU offloading. Experimental results show performance improvements of 4.8x to 7.5x in loading models such as Llama (7, 13, and 70 billion parameters), Falcon (40 billion parameters), and the Bloom (176 billion parameters). | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:24:56 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Yoshimura",
"Takeshi",
""
],
[
"Chiba",
"Tatsuhiro",
""
],
[
"Sethi",
"Manish",
""
],
[
"Waddington",
"Daniel",
""
],
[
"Sundararaman",
"Swaminathan",
""
]
] |
2505.23073 | Alireza Khadem | Alireza Khadem, Kamalavasan Kamalakkannan, Zhenyan Zhu, Akash Poptani, Yufeng Gu, Jered Benjamin Dominguez-Trujillo, Nishil Talati, Daichi Fujiki, Scott Mahlke, Galen Shipman, Reetuparna Das | DX100: A Programmable Data Access Accelerator for Indirection | Proceedings of the 52nd Annual International Symposium on Computer Architecture (ISCA 2025) | null | 10.1145/3695053.3731015 | null | cs.AR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Indirect memory accesses frequently appear in applications where memory bandwidth is a critical bottleneck. Prior indirect memory access proposals, such as indirect prefetchers, runahead execution, fetchers, and decoupled access/execute architectures, primarily focus on improving memory access latency by loading data ahead of computation but still rely on the DRAM controllers to reorder memory requests and enhance memory bandwidth utilization. DRAM controllers have limited visibility to future memory accesses due to the small capacity of request buffers and the restricted memory-level parallelism of conventional core and memory systems. We introduce DX100, a programmable data access accelerator for indirect memory accesses. DX100 is shared across cores to offload bulk indirect memory accesses and associated address calculation operations. DX100 reorders, interleaves, and coalesces memory requests to improve DRAM row-buffer hit rate and memory bandwidth utilization. DX100 provides a general-purpose ISA to support diverse access types, loop patterns, conditional accesses, and address calculations. To support this accelerator without significant programming efforts, we discuss a set of MLIR compiler passes that automatically transform legacy code to utilize DX100. Experimental evaluations on 12 benchmarks spanning scientific computing, database, and graph applications show that DX100 achieves performance improvements of 2.6x over a multicore baseline and 2.0x over the state-of-the-art indirect prefetcher. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 04:25:10 GMT"
}
] | 2025-05-30T00:00:00 | [
[
"Khadem",
"Alireza",
""
],
[
"Kamalakkannan",
"Kamalavasan",
""
],
[
"Zhu",
"Zhenyan",
""
],
[
"Poptani",
"Akash",
""
],
[
"Gu",
"Yufeng",
""
],
[
"Dominguez-Trujillo",
"Jered Benjamin",
""
],
[
"Talati",
"Nishil",
""
],
[
"Fujiki",
"Daichi",
""
],
[
"Mahlke",
"Scott",
""
],
[
"Shipman",
"Galen",
""
],
[
"Das",
"Reetuparna",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.