机器学习学术速递[8.26]

点击阅读原文访问arxivdaily.com，涵盖CS|物理|数学|经济|统计|金融|生物|电气领域，更有搜索、收藏等功能！

cs.LG 方向，今日共计277篇

大模型相关(33篇)

【1】Unraveling the cognitive patterns of Large Language Models through module communities
标题：通过模块社区解开大型语言模型的认知模式
链接：https://arxiv.org/abs/2508.18192

作者：j Bhandari, Pin-Yu Chen, Jianxi Gao

【2】AdLoCo: adaptive batching significantly improves communications efficiency and convergence for Large Language Models
标题：AdLoCo：自适应收件箱显着提高了大型语言模型的通信效率和收敛性
链接：https://arxiv.org/abs/2508.18182

作者：utuzov, Makar Baderko, Stepan Kulibaba, Artem Dzhalilov, Daniel Bobrov, Maxim Mashtaler, Alexander Gasnikov

【3】CMPhysBench: A Benchmark for Evaluating Large Language Models in Condensed Matter Physics
标题：CMPspel Bench：评估凝聚物物理中大型语言模型的基准
链接：https://arxiv.org/abs/2508.18124

作者：g, Dongchen Huang, Jiatong Li, Tengchao Yang, Ziyang Zheng, Di Zhang, Dong Han, Benteng Chen, Binzhao Luo, Zhiyu Liu, Kunling Liu, Zhiyuan Gao, Shiqi Geng, Wei Ma, Jiaming Su, Xin Li, Shuchen Pu, Yuhan Shui, Qianjia Cheng, Zhihao Dou, Dongfei Cui, Changyong He, Jin Zeng, Zeke Xie, Mao Su, Dongzhan Zhou, Yuqiang Li, Wanli Ouyang, Lei Bai, Yunqi Cai, Xi Dai, Shufei Zhang, Jinguang Cheng, Zhong Fang, Hongming Weng
备注：29 pages, 7 figures

【4】Detecting and Characterizing Planning in Language Models
标题：语言模型中的规划检测和描述
链接：https://arxiv.org/abs/2508.18098

作者：nani, Sankaran Vaidyanathan, Connor Watts, Andre N. Assis, Alice Rigg
备注：9 pages, 4 figures

【5】How Quantization Shapes Bias in Large Language Models
标题：量化如何消除大型语言模型中的偏差
链接：https://arxiv.org/abs/2508.18088

作者：Marcuzzi, Xuefei Ning, Roy Schwartz, Iryna Gurevych

【6】Understanding Subword Compositionality of Large Language Models
标题：理解大型语言模型的子词组合
链接：https://arxiv.org/abs/2508.17953

作者：g, Yekun Chai, Anders Søgaard
备注：EMNLP 2025 Main

【7】Debiasing Multilingual LLMs in Cross-lingual Latent Space
标题：跨语言潜在空间中的多语言LLM去偏置
链接：https://arxiv.org/abs/2508.17948

作者：g, Guimin Hu, Yekun Chai, Anders Søgaard
备注：EMNLP 2025 Main

【8】ILRe: Intermediate Layer Retrieval for Context Compression in Causal Language Models
标题：ILRe：因果语言模型中上下文压缩的中间层检索
链接：https://arxiv.org/abs/2508.17892

作者：ang, Mandi Liu, Jiangzhou Ji, Huaijun Li, Haobo Yang, Yaohan He, Jinlong Li

【9】Group Expectation Policy Optimization for Stable Heterogeneous Reinforcement Learning in LLMs
标题：LLM中稳定的异类强化学习的组期望策略优化
链接：https://arxiv.org/abs/2508.17850

作者：, Ruibin Zheng, Zexuan Yi, Hanyang Peng, Hui Wang, Yue Yu

【10】TiKMiX: Take Data Influence into Dynamic Mixture for Language Model Pre-training
标题：TiKMiX：将数据影响纳入语言模型预训练的动态混合中
链接：https://arxiv.org/abs/2508.17677

作者：g, Binbin Liu, Fengze Liu, Yuanfan Guo, Jiyao Deng, Xuecheng Wu, Weidong Zhou, Xiaohuan Zhou, Taifeng Wang

【11】Towards Synthesizing Normative Data for Cognitive Assessments Using Generative Multimodal Large Language Models
标题：使用生成式多模式大型语言模型合成认知评估的规范性数据
链接：https://arxiv.org/abs/2508.17675

作者：Yan, Honor Chotkowski, Fengran Wang, Alex Fedorov
备注：Preprint

【12】Attacking LLMs and AI Agents: Advertisement Embedding Attacks Against Large Language Models
标题：攻击LLM和AI代理：针对大型语言模型的广告嵌入攻击
链接：https://arxiv.org/abs/2508.17674

作者：o, Jinwen Tang, Xingran Huang
备注：7 pages, 2 figures

【13】UQ: Assessing Language Models on Unsolved Questions
标题：UQ：评估未解决问题的语言模型
链接：https://arxiv.org/abs/2508.17580

作者：Ken Ziyu Liu, Zihao Wang, Rui Sun, Wei Liu, Weijia Shi, Huaxiu Yao, Linjun Zhang, Andrew Y. Ng, James Zou, Sanmi Koyejo, Yejin Choi, Percy Liang, Niklas Muennighoff
备注：FN, KZL, and NM are project co-leads and contributed equally. Project website: this https URL

【14】Evaluating Retrieval-Augmented Generation Strategies for Large Language Models in Travel Mode Choice Prediction
标题：评估旅行模式选择预测中大型语言模型的检索增强生成策略
链接：https://arxiv.org/abs/2508.17527

作者：, Junfeng Jiao

【15】MoE-Inference-Bench: Performance Evaluation of Mixture of Expert Large Language and Vision Models
标题：教育部推理台：专家大型语言和视觉模型混合的性能评估
链接：https://arxiv.org/abs/2508.17467

作者：eja Chitty-Venkata, Sylvia Howland, Golara Azar, Daria Soboleva, Natalia Vassilieva, Siddhisanket Raskar, Murali Emani, Venkatram Vishwanath
备注：Preprint

【16】Retrieval Capabilities of Large Language Models Scale with Pretraining FLOPs
标题：具有预训练FLOP的大型语言模型规模的检索能力
链接：https://arxiv.org/abs/2508.17400

作者：tes, Connor Jennings, Erica Ji Yuen, Sasha Doubov, Michael Carbin
备注：15 pages, 4 figures

【17】Graph-R1: Incentivizing the Zero-Shot Graph Learning Capability in LLMs via Explicit Reasoning
标题：Graph-R1：通过显式推理激励LLM中的Zero-Shot图学习能力
链接：https://arxiv.org/abs/2508.17387

作者：, Guangyue Lu, Yuan Zuo, Huarong Zhang, Junjie Wu

【18】Trust Me, I Know This Function: Hijacking LLM Static Analysis using Bias
标题：相信我，我知道这个功能：使用偏差劫持LLM静态分析
链接：https://arxiv.org/abs/2508.17361

作者：stein, David Beste, Daniel Ayzenshteyn, Lea Schonherr, Yisroel Mirsky

【19】AdaptiveK Sparse Autoencoders: Dynamic Sparsity Allocation for Interpretable LLM Representations
标题：AdaptiveK稀疏自动编码器：可解释LLM表示的动态稀疏性分配
链接：https://arxiv.org/abs/2508.17320

作者：, Mengnan Du

【20】TokenLake: A Unified Segment-level Prefix Cache Pool for Fine-grained Elastic Long-Context LLM Serving
标题：TokenLake：用于细粒度弹性长上下文LLM服务的统一段级前置缓存池
链接：https://arxiv.org/abs/2508.17219

作者：Wu, Zili Zhang, Yinmin Zhong, Guanzhe Huang, Yibo Zhu, Xuanzhe Liu, Xin Jin

【21】BudgetThinker: Empowering Budget-aware LLM Reasoning with Control Tokens
标题：BudgetThinker：通过控制令牌增强预算感知LLM推理
链接：https://arxiv.org/abs/2508.17196

作者：Xinrui Wu, Yi Sun, Feifei Zhang, Liye Chen, Jie Wang, Yunxin Liu, Ya-Qin Zhang, Yuanchun Li

【22】LLM Assertiveness can be Mechanistically Decomposed into Emotional and Logical Components
标题：LLM断言可以机械地分解为情感和逻辑组成部分
链接：https://arxiv.org/abs/2508.17182

作者：ujimura, Arush Tagade
备注：This preprint is under review

【23】Towards Safeguarding LLM Fine-tuning APIs against Cipher Attacks
标题：保护LLM微调API免受密码攻击
链接：https://arxiv.org/abs/2508.17158

作者：tra, Mohammed Mahfoud, Yang Yan, Henry Sleight, Ethan Perez, Mrinank Sharma

【24】PlantVillageVQA: A Visual Question Answering Dataset for Benchmarking Vision-Language Models in Plant Science
标题：PlantVillageVQA：用于对植物科学中的视觉语言模型进行基准测试的视觉问题回答数据集
链接：https://arxiv.org/abs/2508.17117

作者：us Sakib, Nafiul Haque, Mohammad Zabed Hossain, Shifat E. Arman
备注：17 pages, 15 figures and Submittd to Nature Scientific Data

【25】Unveiling the Latent Directions of Reflection in Large Language Models
标题：揭示大型语言模型中反射的潜在方向
链接：https://arxiv.org/abs/2508.16989

作者：Chang, Yu-Ting Lee, Pei-Yuan Wu

【26】Breaking the Exploration Bottleneck: Rubric-Scaffolded Reinforcement Learning for General LLM Reasoning
标题：打破探索瓶颈：通用LLM推理的框架强化学习
链接：https://arxiv.org/abs/2508.16949

作者：, Sunzhu Li, Shunyu Liu, Wenkai Fang, Jiale Zhao, Jingwen Yang, Jianwei Lv, Kongcheng Zhang, Yihe Zhou, Hengtong Lu, Wei Chen, Yan Xie, Mingli Song

【27】Interpreting the Effects of Quantization on LLMs
标题：解释量化对LLM的影响
链接：https://arxiv.org/abs/2508.16785

作者：Singh, Hassan Sajjad

【28】Systematic Characterization of LLM Quantization: A Performance, Energy, and Quality Perspective
标题：LLM量化的系统性特征：性能、能量和质量的角度
链接：https://arxiv.org/abs/2508.16712

作者：hi, Yi Ding
备注：14 pages, 10 figures, 4 tables

【29】CALR: Corrective Adaptive Low-Rank Decomposition for Efficient Large Language Model Layer Compression
标题：CALR：用于高效大型语言模型层压缩的纠正性自适应低等级分解
链接：https://arxiv.org/abs/2508.16680

作者： Daniyal Kautsar, Afra Majida Hariono, Widyawan, Syukron Abu Ishaq Alfarozi, Kuntpong Wararatpanya
备注：Submitted to IEEE Transactions on Artificial Intelligence. This is the preprint version, not peer-reviewed. The final version may differ after peer review. (11 pages, 3 figures)

【30】Recall-Extend Dynamics: Enhancing Small Language Models through Controlled Exploration and Refined Offline Integration
标题：召回-扩展动力学：通过受控探索和精细离线集成增强小型语言模型
链接：https://arxiv.org/abs/2508.16677

作者：n, Likang Wu, Hongke Zhao, Jiahui Wang, Le Wu

【31】WISCA: A Lightweight Model Transition Method to Improve LLM Training via Weight Scaling
标题：WISCA：一种通过体重测量改善LLM训练的轻量级模型转换方法
链接：https://arxiv.org/abs/2508.16676

作者：Li, Jianchao Tan, Zhidong Yang, Pingwei Sun, Feiye Huo, Jiayu Qin, Yerui Sun, Yuchen Xie, Xunliang Cai, Xiangyu Zhang, Maoxin He, Guangming Tan, Weile Jia, Tong Zhao

【32】Learn to Memorize: Optimizing LLM-based Agents with Adaptive Memory Framework
标题：学习小型化：使用自适应内存框架优化基于LLM的代理
链接：https://arxiv.org/abs/2508.16629

作者：g, Quanyu Dai, Rui Li, Xiaohe Bo, Xu Chen, Zhenhua Dong
备注：17 pages, 4 figures, 5 tables

【33】Confidence-Modulated Speculative Decoding for Large Language Models
标题：大型语言模型的置信调制推测解码
链接：https://arxiv.org/abs/2508.15371

作者：n, Subhasis Dasgupta, Hetvi Waghela
备注：This is the preprint of the paper, which has been accepted for oral presentation and publication in the proceedings of IEEE INDISCON 2025. The conference will be organized at the National Institute of Technology, Rourkela, India, from August 21 to 23, 2025. The paper is 10 pages long, and it contains 2 figures and 5 tables

Graph相关(图学习|图神经网络|图优化等)(16篇)

【1】Unveiling the Actual Performance of Neural-based Models for Equation Discovery on Graph Dynamical Systems
标题：图动力系统方程发现神经模型的实际性能
链接：https://arxiv.org/abs/2508.18173

作者：Cappi, Paolo Frazzetto, Nicolò Navarin, Alessandro Sperduti
备注：Preprint. Under Review

【2】Weisfeiler-Lehman meets Events: An Expressivity Analysis for Continuous-Time Dynamic Graph Neural Networks
标题：Weisfeiler-Lehman遇上事件：连续时间动态图神经网络的表现性分析
链接：https://arxiv.org/abs/2508.18052

作者：ddar-Wiesing, Alice Moallemy-Oureh

【3】Ada-TransGNN: An Air Quality Prediction Model Based On Adaptive Graph Convolutional Networks
标题：Ada-TransGNN：基于自适应图卷积网络的空气质量预测模型
链接：https://arxiv.org/abs/2508.17867

作者： Feng Jiang, Zhanquan Wang
备注：15 pages, 4 figures, 3 tables. This paper is accepted by ICONIP2025 but not published

【4】Quantum Graph Attention Network: A Novel Quantum Multi-Head Attention Mechanism for Graph Learning
标题：量子图注意力网络：一种用于图学习的新型量子多头注意力机制
链接：https://arxiv.org/abs/2508.17630

作者：Tai Yue Li, Nan Yow Chen

【5】Bridging Graph and State-Space Modeling for Intensive Care Unit Length of Stay Prediction
标题：重症监护室住院时间预测的桥图和状态空间模型
链接：https://arxiv.org/abs/2508.17554

作者： Haitz Sáez de Ocáriz Borde, Emma Rocheteau, Pietro Lio'

【6】Gumbel-MPNN: Graph Rewiring with Gumbel-Softmax
标题：Gumbel-MPNN：用Gumbel-Softmax重新布线图形
链接：https://arxiv.org/abs/2508.17531

作者：ffmann, Lukas Galke, Ansgar Scherp

【7】Effective Clustering for Large Multi-Relational Graphs
标题：大型多关系图的有效聚集
链接：https://arxiv.org/abs/2508.17388

作者：Lin, Runhao Jiang, Renchi Yang
备注：23 pages. The technical report for the paper titled "Effective Clustering for Large Multi-Relational Graphs" in SIGMOD 2026

【8】Scaling Graph Transformers: A Comparative Study of Sparse and Dense Attention
标题：缩放图变形者：稀疏和密集注意力的比较研究
链接：https://arxiv.org/abs/2508.17175

作者：trov

【9】Two Birds with One Stone: Enhancing Uncertainty Quantification and Interpretability with Graph Functional Neural Process
标题：一箭双雕：用图函数神经过程增强不确定性量化和可解释性
链接：https://arxiv.org/abs/2508.17097

作者：ong, Haotian Sun, Yuchen Zhuang, Haorui Wang, Wenhao Mu, Chao Zhang
备注：AISTATS'25

【10】Physics-Inspired Spatial Temporal Graph Neural Networks for Predicting Industrial Chain Resilience
标题：物理启发的时空图神经网络预测产业链弹性
链接：https://arxiv.org/abs/2508.16836

作者：ang, Junping Wang, Yibo Xue

【11】Understanding and Tackling Over-Dilution in Graph Neural Networks
标题：理解和解决图神经网络中的过度稀释
链接：https://arxiv.org/abs/2508.16829

作者：ee, Veronika Thost, Bumsoo Kim, Jaewoo Kang, Tengfei Ma
备注：Extended version of KDD '25 paper. 22 pages including appendix. Conference version: KDD '25 (Toronto, Aug 3-7, 2025), pp. 1253-1261. Code: https://github.com/LeeJunHyun/NATR

【12】Latent Graph Learning in Generative Models of Neural Signals
标题：神经信号生成模型中的潜在图学习
链接：https://arxiv.org/abs/2508.16776

作者： Kodama, Kenneth A. Loparo

【13】DR-CircuitGNN: Training Acceleration of Heterogeneous Circuit Graph Neural Network on GPUs
标题：DR-CircuitGNN：在图形处理器上进行异类电路图神经网络的训练加速
链接：https://arxiv.org/abs/2508.16769

作者：, Shiyang Li, Junran Tao, Kiran Thorat, Xi Xie, Hongwu Peng, Nuo Xu, Caiwen Ding, Shaoyi Huang
备注：None

【14】STGAtt: A Spatial-Temporal Unified Graph Attention Network for Traffic Flow Forecasting
标题：STGAtt：用于交通流预测的时空统一图注意力网络
链接：https://arxiv.org/abs/2508.16685

作者：iang, Jianxun Cui, Qingshuang Zeng, Feng Liu, Nenad Filipovic, Tijana Geroski

【15】A Novel Unified Extended Matrix for Graph Signal Processing: Theory and Application
标题：一种新型的图形信号处理统一扩展矩阵：理论与应用
链接：https://arxiv.org/abs/2508.16633

作者：eng, Zhichao Zhang, Wei Yao

【16】GraphPPD: Posterior Predictive Modelling for Graph-Level Inference
标题：GraphCPD：图级推理的后验预测建模
链接：https://arxiv.org/abs/2508.16995

作者：dar Pal, Liheng Ma, Amine Natik, Yingxue Zhang, Mark Coates

Transformer(12篇)

【1】Frozen in Time: Parameter-Efficient Time Series Transformers via Reservoir-Induced Feature Expansion and Fixed Random Dynamics
标题：冻结在时间中：通过水库诱导的特征扩展和固定随机动态学的参数高效时间序列变换器
链接：https://arxiv.org/abs/2508.18130

作者：ingh, Mehak Sharma, Anupriya Dey, Balasubramanian Raman
备注：8 pages, 5 tables, 3 figures, accepted at ECAI 2025

【2】Arnold: a generalist muscle transformer policy
标题：阿诺德：多面手肌肉Transformer政策
链接：https://arxiv.org/abs/2508.18066

作者：ilvio Chiappa, Boshi An, Merkourios Simos, Chengkun Li, Alexander Mathis
备注：A.S.C. and B.A. contributed equally. Code is available at this https URL

【3】Training Transformers for Mesh-Based Simulations
标题：训练变形器以进行基于网格的模拟
链接：https://arxiv.org/abs/2508.18051

作者：ier, Vincent Lannelongue, Jonathan Viquerat, Elie Hachem

【4】In-Context Algorithm Emulation in Fixed-Weight Transformers
标题：固定权重变形器中的上下文算法仿真
链接：https://arxiv.org/abs/2508.17550

作者：-Chieh Hu, Hude Liu, Jennifer Yuntong Zhang, Han Liu
备注：Code is available at this https URL

【5】Module-Aware Parameter-Efficient Machine Unlearning on Transformers
标题：Transformer上的模块感知参数高效机器学习
链接：https://arxiv.org/abs/2508.17233

作者：o, Jian Lou, Yuke Hu, Xiaochen Li, Zhihao Liu, Jiaqi Liu, Zhan Qin, Kui Ren

【6】GPG-HT: Generalized Policy Gradient with History-Aware Decision Transformer for Probabilistic Path Planning
标题：GPG-HT：具有历史感知决策Transformer的广义政策梯度，用于概率路径规划
链接：https://arxiv.org/abs/2508.17218

作者： Yuqi Ouyang

【7】TriagerX: Dual Transformers for Bug Triaging Tasks with Content and Interaction Based Rankings
标题：TriagerX：双Transformer，通过基于内容和交互的排名进行漏洞分类任务
链接：https://arxiv.org/abs/2508.16860

作者：l Mamun, Gias Uddin, Lan Xia, Longyu Zhang
备注：This work is currently under review at IEEE Transactions on Software Engineering. The replication package will be made publicly available upon acceptance

【8】The Loupe: A Plug-and-Play Attention Module for Amplifying Discriminative Features in Vision Transformers
标题：The Loupe：一个即插即用的注意力模块，用于放大视觉Transformer中的区分特征
链接：https://arxiv.org/abs/2508.16663

作者：godan

【9】A Laplace diffusion-based transformer model for heart rate forecasting within daily activity context
标题：用于日常活动背景下心率预测的基于拉普拉斯扩散的Transformer模型
链接：https://arxiv.org/abs/2508.16655

作者：teescu, Ioana Hadarau, Ionut Anghel, Tudor Cioara, Ovidiu Anchidin, Ancuta Nemes

【10】Enhancing Transformer-Based Foundation Models for Time Series Forecasting via Bagging, Boosting and Statistical Ensembles
标题：通过Bagging、Boosting和统计集成增强基于转换器的时间序列预测基础模型
链接：https://arxiv.org/abs/2508.16641

作者：Modi, Rong Pan

【11】Recurrent Transformer U-Net Surrogate for Flow Modeling and Data Assimilation in Subsurface Formations with Faults
标题：递归Transformer U-Net代理用于地下断层地层流动模拟和数据同化
链接：https://arxiv.org/abs/2508.16631

作者： Louis J. Durlofsky

【12】CrystalDiT: A Diffusion Transformer for Crystal Generation
标题：CrystalDiT：一种用于晶体产生的扩散Transformer
链接：https://arxiv.org/abs/2508.16614

作者：i, Guikun Xu, Xi Xiao, Zhong Zhang, Liu Liu, Yatao Bian, Peilin Zhao
备注：18 pages, 18 figures. Code available at this https URL

GAN|对抗|攻击|生成相关(13篇)

【1】Does simple trump complex? Comparing strategies for adversarial robustness in DNNs
标题：简单胜过复杂吗？比较DNN中对抗鲁棒性的策略
链接：https://arxiv.org/abs/2508.18019

作者：rooks, Marelie H. Davel, Coenraad Mouton

【2】FasterVoiceGrad: Faster One-step Diffusion-Based Voice Conversion with Adversarial Diffusion Conversion Distillation
标题：FasterEqualGrad：更快的基于扩散的一步语音转换，采用对抗扩散转换蒸馏
链接：https://arxiv.org/abs/2508.17868

作者：Kaneko, Hirokazu Kameoka, Kou Tanaka, Yuto Kondo
备注：Accepted to Interspeech 2025. Project page: this https URL

【3】SuperGen: An Efficient Ultra-high-resolution Video Generation System with Sketching and Tiling
标题：SuperGen：一个具有草图和拼贴功能的高效超高分辨率视频生成系统
链接：https://arxiv.org/abs/2508.17756

作者：Ye, Zepeng Zhao, Yi Mu, Jucheng Shen, Renjie Li, Kaijian Wang, Desen Sun, Saurabh Agarwal, Myungjin Lee, Triston Cao, Aditya Akella, Arvind Krishnamurthy, T.S. Eugene Ng, Zhengzhong Tu, Yuke Wang

【4】Multi-layer Abstraction for Nested Generation of Options (MANGO) in Hierarchical Reinforcement Learning
标题：分层强化学习中的嵌套期权生成（MANGO）的多层抽象
链接：https://arxiv.org/abs/2508.17751

作者：rcudi, Davide Sartor, Alberto Sinigaglia, Vincent François-Lavet, Gian Antonio Susto

【5】Robustness Feature Adapter for Efficient Adversarial Training
标题：用于高效对抗训练的鲁棒性特征适配器
链接：https://arxiv.org/abs/2508.17680

作者：u, Jun Guo, Wei Wang, Yi Wang
备注：The paper has been accepted for presentation at ECAI 2025

【6】ChartMaster: Advancing Chart-to-Code Generation with Real-World Charts and Chart Similarity Reinforcement Learning
标题：ChartMaster：使用真实世界图表和图表相似性强化学习推进图表到代码生成
链接：https://arxiv.org/abs/2508.17608

作者：n, Qiong Cao, Chao Xue, Yibing Zhan, Changxing Ding, Xiaodong He

【7】MetaGen: A DSL, Database, and Benchmark for VLM-Assisted Metamaterial Generation
标题：MetaGen：VLM辅助超材料生成的一个数字线路、数据库和基准
链接：https://arxiv.org/abs/2508.17568

作者：atura, Benjamin Jones, Siyuan Bian, Wojciech Matusik

【8】Adversarial Examples Are Not Bugs, They Are Superposition
标题：对抗性例子不是错误，而是叠加
链接：https://arxiv.org/abs/2508.17456

作者：n, Owen Lewis

【9】FRAME : Comprehensive Risk Assessment Framework for Adversarial Machine Learning Threats
标题：FRAME：对抗性机器学习威胁的全面风险评估框架
链接：https://arxiv.org/abs/2508.17405

作者：hapira, Simon Shigol, Asaf Shabtai

【10】ShortListing Model: A Streamlined SimplexDiffusion for Discrete Variable Generation
标题：入围列表模型：离散变量生成的简化SimplexDistribution
链接：https://arxiv.org/abs/2508.17345

作者：ng, Zhe Zhang, Yu Pei, Jingjing Gong, Qiying Yu, Zheng Zhang, Mingxuan Wang, Hao Zhou, Jingjing Liu, Wei-Ying Ma

【11】MC3G: Model Agnostic Causally Constrained Counterfactual Generation
标题：MC 3G：模型不可知的因果约束反事实生成
链接：https://arxiv.org/abs/2508.17221

作者：gupta, Sadaf MD Halim, Joaquín Arias, Elmer Salazar, Gopal Gupta

【12】GRAID: Synthetic Data Generation with Geometric Constraints and Multi-Agentic Reflection for Harmful Content Detection
标题：GRAID：用于有害内容检测的几何约束和多尺度反射合成数据生成
链接：https://arxiv.org/abs/2508.17057

作者：azemi Rad, Alberto Purpura, Himanshu Kumar, Emily Chen, Mohammad Shahed Sorower
备注：19 pages, 12 figures

【13】Online Learning for Approximately-Convex Functions with Long-term Adversarial Constraints
标题：具有长期对抗约束的近凸函数的在线学习
链接：https://arxiv.org/abs/2508.16992

作者：kar, Samrat Mukhopadhyay, Abhishek Sinha

半/弱/无/有监督|不确定性|主动学习(12篇)

【1】A Novel Framework for Uncertainty Quantification via Proper Scores for Classification and Beyond
标题：通过分类及超越的适当分数进行不确定性量化的新框架
链接：https://arxiv.org/abs/2508.18001

作者： G. Gruber
备注：PhD Thesis (cumulative, spanning 6 peer-reviewed publications)

【2】Proximal Supervised Fine-Tuning
标题：近端监督微调
链接：https://arxiv.org/abs/2508.17784

作者：hu, Ruobing Xie, Rui Wang, Xingwu Sun, Di Wang, Pengfei Liu

【3】Evaluating the Quality of the Quantified Uncertainty for (Re)Calibration of Data-Driven Regression Models
标题：评估数据驱动回归模型（重新）校准的量化不确定性的质量
链接：https://arxiv.org/abs/2508.17761

作者：beke, Nico Schönfisch, Sebastian Rohjans, Andreas Rauh

【4】Deep Learning with Self-Attention and Enhanced Preprocessing for Precise Diagnosis of Acute Lymphoblastic Leukemia from Bone Marrow Smears in Hemato-Oncology
标题：深度学习、自我注意力和增强预处理，用于从血液肿瘤学中的骨髓涂片精确诊断急性淋巴细胞白血病
链接：https://arxiv.org/abs/2508.17216

作者：, Md.Mahbubul Haque, Bishowjit Paul
备注：26 pages, 15 figures, 8 tables. VGG19+MHSA with Focal Loss; test accuracy 99.25%

【5】Quantifying Out-of-Training Uncertainty of Neural-Network based Turbulence Closures
标题：量化基于神经网络的湍流关闭的训练外不确定性
链接：https://arxiv.org/abs/2508.16891

作者：an, Som Dhulipala, Mauricio Tano, Izabela Gutowska, Som Dutta

【6】UM3: Unsupervised Map to Map Matching
标题：UM3：无监督地图到地图匹配
链接：https://arxiv.org/abs/2508.16874

作者：Ying, Yinan Zhang, Lei Zhang, Jiazhuang Wang, Shujun Jia, Tianshu Yu
备注：Accepted by ACM SIGSPATIAL 2025

【7】Uncertainty Propagation Networks for Neural Ordinary Differential Equations
标题：神经常微方程的不确定性传播网络
链接：https://arxiv.org/abs/2508.16815

作者：nshahi, Zheng H. Zhu

【8】FAIRWELL: Fair Multimodal Self-Supervised Learning for Wellbeing Prediction
标题：FAIRWELL：健康预测的公平多模式自我监督学习
链接：https://arxiv.org/abs/2508.16748

作者：ong, Abtin Mogharabin, Paul Liang, Hatice Gunes, Sinan Kalkan

【9】OASIS: Open-world Adaptive Self-supervised and Imbalanced-aware System
标题：OASIS：开放世界自适应自我监督和不平衡感知系统
链接：https://arxiv.org/abs/2508.16656

作者： Mugon Joe, Minhae Kwon
备注：Accepted at the 34th ACM International Conference on Information and Knowledge Management (CIKM 2025)

【10】On the sample complexity of semi-supervised multi-objective learning
标题：半监督多目标学习的样本复杂性
链接：https://arxiv.org/abs/2508.17152

作者：gel, Geelon So, Junhyung Park, Fanny Yang

【11】A Decoupled LOB Representation Framework for Multilevel Manipulation Detection with Supervised Contrastive Learning
标题：基于监督对比学习的多层次操作检测解耦LOB表示框架
链接：https://arxiv.org/abs/2508.17086

作者：, Peng Yang

【12】Generative Latent Diffusion Model for Inverse Modeling and Uncertainty Analysis in Geological Carbon Sequestration
标题：地质碳封存逆建模和不确定性分析的生成潜在扩散模型
链接：https://arxiv.org/abs/2508.16640

作者：, Xin-Yang Liu, Meet Hemant Parikh, Junyi Guo, Pan Du, Bicheng Yan, Jian-Xun Wang

迁移|Zero/Few/One-Shot|自适应(15篇)

【1】Type-Compliant Adaptation Cascades: Adapting Programmatic LM Workflows to Data
标题：符合类型的自适应级联：将编程LM工作流适应数据
链接：https://arxiv.org/abs/2508.18244

作者： Lin, Daiyi Peng, Yifeng Lu, Ming Zhang, Eugene Ie

【2】AQ-PCDSys: An Adaptive Quantized Planetary Crater Detection System for Autonomous Space Exploration
标题：AQ-PCDS：用于自主太空探索的自适应量化行星陨石坑检测系统
链接：https://arxiv.org/abs/2508.18025

作者：ul, Archan Paul
备注：17 pages, 6 figures. A research paper on a novel deep learning framework for planetary crater detection

【3】Spectrum Prediction in the Fractional Fourier Domain with Adaptive Filtering
标题：自适应过滤的分数傅里叶域频谱预测
链接：https://arxiv.org/abs/2508.17872

作者：in, Bo Zhou, Guangliang Pan, Qihui Wu, Meixia Tao
备注：Accepted by IEEE Wireless Communications Letters

【4】A Contrastive Learning-Guided Confident Meta-learning for Zero Shot Anomaly Detection
标题：用于Zero-Shot异常检测的对比学习引导自信元学习
链接：https://arxiv.org/abs/2508.17827

作者：Aqeel, Danijel Skocaj, Marco Cristani, Francesco Setti
备注：Accepted to VISION Workshop at ICCV 2025

【5】Adaptive Ensemble Learning with Gaussian Copula for Load Forecasting
标题：基于高斯Copula的自适应集成学习用于负荷预测
链接：https://arxiv.org/abs/2508.17700

作者：ang, Gang Lu, Xiaoqing Yan, Peng Xia, Di Wu

【6】Towards Optimal Convolutional Transfer Learning Architectures for Breast Lesion Classification and ACL Tear Detection
标题：迈向乳腺病变分类和ACL撕裂检测的最佳卷积转移学习架构
链接：https://arxiv.org/abs/2508.17567

作者：ees, Moritz Bolling, Aditri Bhagirath

【7】Efficient Zero-Shot Long Document Classification by Reducing Context Through Sentence Ranking
标题：通过句子排序减少上下文的高效Zero-Shot长文档分类
链接：https://arxiv.org/abs/2508.17490

作者：h Kokate, Mitali Sarnaik, Manavi Khopade, Mukta Takalikar, Raviraj Joshi

【8】FedKLPR: Personalized Federated Learning for Person Re-Identification with Adaptive Pruning
标题：FedKLPO：通过自适应修剪进行人员重新识别的个性化联邦学习
链接：https://arxiv.org/abs/2508.17431

作者：Yu, Yu-Syuan Tseng, Shao-Yi Chien

【9】DropLoRA: Sparse Low-Rank Adaptation for Parameter-Efficient Fine-Tuning
标题：DropLoRA：稀疏低等级自适应，实现参数高效的微调
链接：https://arxiv.org/abs/2508.17337

作者：ang
备注：8 pages

【10】Tri-Accel: Curvature-Aware Precision-Adaptive and Memory-Elastic Optimization for Efficient GPU Usage
标题：Tri-Accel：曲线感知精确自适应和内存弹性优化，以实现高效的图形处理
链接：https://arxiv.org/abs/2508.16905

作者：eibanian, Pouya Shaeri, Alimohammad Beigi, Ryan T. Woo, Aryan Keluskar

【11】WST: Weak-to-Strong Knowledge Transfer via Reinforcement Learning
标题：WST：通过强化学习从弱到强的知识转移
链接：https://arxiv.org/abs/2508.16741

作者：, Shuo Li, Lianghuan Huang

【12】AdapSNE: Adaptive Fireworks-Optimized and Entropy-Guided Dataset Sampling for Edge DNN Training
标题：AdapSNE：用于边缘DNN训练的自适应烟花优化和信息引导数据集采样
链接：https://arxiv.org/abs/2508.16647

作者：o, Hetian Liu, Zihang Yuan, Li Zhu, Fan Yang, Lina Xie Tian Xia, Wenzhe Zhao, Pengju Ren

【13】Few-shot Class-incremental Fault Diagnosis by Preserving Class-Agnostic Knowledge with Dual-Granularity Representations
标题：通过用双粒度表示保留类不可知知识进行少次类增量故障诊断
链接：https://arxiv.org/abs/2508.16634

作者：Yang, Jie Wang, Liansong Zong, Xiaorong Liu, Quan Qian, Shiqian Chen

【14】Adaptive Variance-Penalized Continual Learning with Fisher Regularization
标题：具有Fisher正规化的自适应方差惩罚连续学习
链接：https://arxiv.org/abs/2508.16632

作者：arkar

【15】Unseen Speaker and Language Adaptation for Lightweight Text-To-Speech with Adapters
标题：使用适配器实现轻量级文本到语音的隐形说话者和语言适应
链接：https://arxiv.org/abs/2508.18006

作者：alai, Ziyao Zhang, Akos Gangoly
备注：Accepted at IEEE MLSP 2025

强化学习(3篇)

【1】ReviBranch: Deep Reinforcement Learning for Branch-and-Bound with Revived Trajectories
标题：ReviBranch：具有复兴轨迹的分支绑定深度强化学习
链接：https://arxiv.org/abs/2508.17452

作者：o, Nie Jiayi, Yihang Cheng, Jinwei Liu, Yingrui Ji, Canran Xiao, Feixiang Du, Jiaping Xiao
备注：conference

【2】Rectified Robust Policy Optimization for Model-Uncertain Constrained Reinforcement Learning without Strong Duality
标题：无强二元性的模型不确定约束强化学习的纠正鲁棒策略优化
链接：https://arxiv.org/abs/2508.17448

作者：Ma, Ziyi Chen, Yi Zhou, Heng Huang

【3】Autonomous UAV Flight Navigation in Confined Spaces: A Reinforcement Learning Approach
标题：受限空间中的自主无人机飞行导航：强化学习方法
链接：https://arxiv.org/abs/2508.16807

作者：Tayar, Lucas K. de Oliveira, Juliano D. Negri, Thiago H. Segreto, Ricardo V. Godoy, Marcelo Becker

元学习(1篇)

【1】Robust Anomaly Detection in Industrial Environments via Meta-Learning
标题：通过元学习在工业环境中进行鲁棒的异常检测
链接：https://arxiv.org/abs/2508.17789

作者：Aqeel, Shakiba Sharifi, Marco Cristani, Francesco Setti
备注：Accepted to VISION Workshop at ICCV 2025

医学相关(6篇)

【1】Diffusion-Based Data Augmentation for Medical Image Segmentation
标题：基于扩散的数据增强用于医学图像分割
链接：https://arxiv.org/abs/2508.17844

作者：ir, Muhammad Aqeel, Francesco Setti
备注：Accepted to CVAMD Workshop at ICCV 2025

【2】Multi-domain Distribution Learning for De Novo Drug Design
标题：用于从头开始药物设计的多域分布学习
链接：https://arxiv.org/abs/2508.17815

作者：euing, Ilia Igashov, Adrian W. Dobbelstein, Thomas Castiglione, Michael Bronstein, Bruno Correia
备注：None

【3】Segmentation and Classification of Pap Smear Images for Cervical Cancer Detection Using Deep Learning
标题：使用深度学习进行子宫颈抹片图像的分割和分类以检测宫颈癌
链接：https://arxiv.org/abs/2508.17728

作者：lbzour, Sarah S. Lam

【4】Longitudinal Progression Prediction of Alzheimer's Disease with Tabular Foundation Model
标题：表格基金会模型预测阿尔茨海默病的纵向进展
链接：https://arxiv.org/abs/2508.17649

作者：ng, Jiawen Ren, Jiaying Lu, Gloria Hyunjung Kwak, Armin Iraji, Alex Fedorov

【5】How to make Medical AI Systems safer? Simulating Vulnerabilities, and Threats in Multimodal Medical RAG System
标题：如何让医疗人工智能系统更安全？模拟多模式医疗RAG系统中的漏洞和威胁
链接：https://arxiv.org/abs/2508.17215

作者：o, Zelin Liu, Raman Dutt, Ziyang Wang, Zhongtian Sun, Yeming Wang, Fan Mo, Pietro Liò
备注：Sumbitted to 2025 AAAI main track

【6】Clinical characteristics, complications and outcomes of critically ill patients with Dengue in Brazil, 2012-2024: a nationwide, multicentre cohort study
标题：2012-2024年巴西登革热重症患者的临床特征、并发症和结局：一项全国性、多中心队列研究
链接：https://arxiv.org/abs/2508.18207

作者： Peres, Otavio T. Ranzani, Leonardo S.L. Bastos, Silvio Hamacher, Tom Edinburgh, Esteban Garcia-Gallo, Fernando Augusto Bozza
备注：None

蒸馏|知识提取(1篇)

【1】Sig-DEG for Distillation: Making Diffusion Models Faster and Lighter
标题：Sig-DEG用于蒸馏：使扩散模型更快、更轻
链接：https://arxiv.org/abs/2508.16939

作者：, Wen Ge, Niels Cariou-Kotlarek, Mingxuan Yi, Po-Yu Chen, Lingyi Yang, Francois Buet-Golfouse, Gaurav Mittal, Hao Ni

推荐(3篇)

【1】PCR-CA: Parallel Codebook Representations with Contrastive Alignment for Multiple-Category App Recommendation
标题：PCR-CA：用于多类别应用程序推荐的具有对比对齐的并行码本表示
链接：https://arxiv.org/abs/2508.18166

作者：Wangyao Ge, Yidi Wang, Xin Liu, Jeff Burtoft, Hao Fan, Hui Wang
备注：9 pages, 4 figures, conference

【2】Test-Time Scaling Strategies for Generative Retrieval in Multimodal Conversational Recommendations
标题：多模态会话推荐中生成检索的测试时尺度策略
链接：https://arxiv.org/abs/2508.18132

作者： Hsu, Yuan-Ching Kuo, Chao-Han Huck Yang, Szu-Wei Fu, Hanrong Ye, Hongxu Yin, Yu-Chiang Frank Wang, Ming-Feng Tsai, Chuan-Ju Wang

【3】Bootstrapping Conditional Retrieval for User-to-Item Recommendations
标题：用户对项目推荐的引导条件检索
链接：https://arxiv.org/abs/2508.16793

作者：in, Haoyu Chen, Jaewon Jang, Jiajing Xu

聚类(1篇)

【1】SACA: Selective Attention-Based Clustering Algorithm
标题：SACA：基于选择性注意力的分簇算法
链接：https://arxiv.org/abs/2508.17150

作者：irdel Bilehsavar, Razieh Ghaedi, Samira Seyed Taheri, Xinqi Fan, Christian O'Reilly
备注：22 pages, 10 figures

超分辨率|去噪|去模糊|去雾(1篇)

【1】Multidimensional Distributional Neural Network Output Demonstrated in Super-Resolution of Surface Wind Speed
标题：多维分布神经网络输出在地表风速超分辨率中证明
链接：https://arxiv.org/abs/2508.16686

作者：J. Goldwyn, Mitchell Krock, Johann Rudi, Daniel Getter, Julie Bessac

自动驾驶|车辆|车道检测等(1篇)

【1】A Retrieval Augmented Spatio-Temporal Framework for Traffic Prediction
标题：用于交通预测的检索增强时空框架
链接：https://arxiv.org/abs/2508.16623

作者：an, Xilin Dang, Ziyu Zhou, Sisuo Lyu, Yuxuan Liang

点云|SLAM|雷达|激光|深度RGBD相关(1篇)

【1】Beyond Memorization: Extending Reasoning Depth with Recurrence, Memory and Test-Time Compute Scaling
标题：超越推理：用递归、内存和测试时计算缩放扩展推理深度
链接：https://arxiv.org/abs/2508.16745

作者：in, Daniil Orel, Konstantin Smirnov, Arman Bolatov, Bilal Elbouardi, Besher Hassan, Yuri Kuratov, Aydar Bulatov, Preslav Nakov, Timothy Baldwin, Artem Shelmanov, Mikhail Burtsev

联邦学习|隐私保护|加密(5篇)

【1】FedGreed: A Byzantine-Robust Loss-Based Aggregation Method for Federated Learning
标题：FedGreed：一种用于联邦学习的拜占庭鲁棒的基于损失的聚合方法
链接：https://arxiv.org/abs/2508.18060

作者： Kritharakis, Antonios Makris, Dusan Jakovetic, Konstantinos Tserpes
备注：8 pages, 4 figures

【2】Choice Outweighs Effort: Facilitating Complementary Knowledge Fusion in Federated Learning via Re-calibration and Merit-discrimination
标题：选择大于努力：通过重新校准和优点区分促进联邦学习中的互补知识融合
链接：https://arxiv.org/abs/2508.17954

作者：, Dongrun Li, Xin Wang, Xiaoyang Yu, Xiaoming Wu, Shibo He

【3】Rethinking Federated Learning Over the Air: The Blessing of Scaling Up
标题：重新思考空中联邦学习：扩大规模的祝福
链接：https://arxiv.org/abs/2508.17697

作者：, Bikramjit Das, Yong Xie, Nikolaos Pappas, Howard H. Yang

【4】FedERL: Federated Efficient and Robust Learning for Common Corruptions
标题：FedERL：针对常见腐败的联合高效且稳健的学习
链接：https://arxiv.org/abs/2508.17381

作者：ache, Naresh Shanbhag

【5】Degree of Staleness-Aware Data Updating in Federated Learning
标题：联邦学习中感知停滞的数据更新程度
链接：https://arxiv.org/abs/2508.16931

作者：Xuehe Wang
备注：accepted by European Conference on Artificial Intelligence

推理|分析|理解|解释(12篇)

【1】A Proportional-Integral Controller-Incorporated SGD Algorithm for High Efficient Latent Factor Analysis
标题：一种结合比例积分控制器的高效潜在因子分析SGD算法
链接：https://arxiv.org/abs/2508.17609

作者： Shiyu Long, Minglian Han

【2】TreePO: Bridging the Gap of Policy Optimization and Efficacy and Inference Efficiency with Heuristic Tree-based Modeling
标题：TreePO：利用启发式基于树的建模弥合政策优化、有效性和推理效率的差距
链接：https://arxiv.org/abs/2508.17445

作者： Qingshui Gu, Zhoufutu Wen, Ziniu Li, Tianshun Xing, Shuyue Guo, Tianyu Zheng, Xin Zhou, Xingwei Qu, Wangchunshu Zhou, Zheng Zhang, Wei Shen, Qian Liu, Chenghua Lin, Jian Yang, Ge Zhang, Wenhao Huang

【3】Explainable AI (XAI) for Arrhythmia detection from electrocardiograms
标题：可解释人工智能（XAI），用于从心电图中检测心律失常
链接：https://arxiv.org/abs/2508.17294

作者：eck, Arlene John

【4】MaRVL-QA: A Benchmark for Mathematical Reasoning over Visual Landscapes
标题：MaRVL-QA：视觉景观数学推理的基准
链接：https://arxiv.org/abs/2508.17180

作者：de, Sahiti Yerramilli, Jayant Sravan Tamarapalli, Rynaa Grover

【5】Frequency Response Identification of Low-Order Systems: Finite-Sample Analysis
标题：低频系统的频率响应识别：伪样本分析
链接：https://arxiv.org/abs/2508.17142

作者：rpisheh, Mario Sznaier
备注：15 pages, Submitted to IEEE Transactions on Automatic Control

【6】EduRABSA: An Education Review Dataset for Aspect-based Sentiment Analysis Tasks
标题：EduRABSA：用于基于Aspects的情感分析任务的教育评论数据集
链接：https://arxiv.org/abs/2508.17008

作者： Hua, Paul Denny, Jörg Wicker, Katerina Taskova

【7】Reinforcement-Guided Hyper-Heuristic Hyperparameter Optimization for Fair and Explainable Spiking Neural Network-Based Financial Fraud Detection
标题：增强引导的超启发式超参数优化，用于公平且可解释的尖峰神经网络的金融欺诈检测
链接：https://arxiv.org/abs/2508.16915

作者：hammad Nasif, Md Abrar Jahin, M. F. Mridha

【8】Neural Contrast Expansion for Explainable Structure-Property Prediction and Random Microstructure Design
标题：神经对比扩展用于可解释结构性能预测和随机微结构设计
链接：https://arxiv.org/abs/2508.16857

作者：ie, Yang Jiao, Yi Ren

【9】PuzzleJAX: A Benchmark for Reasoning and Learning
标题：PuzzleJAX：推理和学习的基准
链接：https://arxiv.org/abs/2508.16821

作者：, Graham Todd, Yuchen Li, Ahmed Khalifa, Muhammad Umair Nasir, Zehua Jiang, Andrzej Banburski-Fahey, Julian Togelius
备注：25 pages, 11 figures, 2 tables

【10】Explainable AI for Predicting and Understanding Mathematics Achievement: A Cross-National Analysis of PISA 2018
标题：预测和理解数学成绩的可解释人工智能：2018年PISA的跨国分析
链接：https://arxiv.org/abs/2508.16747

作者：Rui Dai

【11】CP4SBI: Local Conformal Calibration of Credible Sets in Simulation-Based Inference
标题：CP 4SBI：基于模拟的推理中可信集的局部共形校准
链接：https://arxiv.org/abs/2508.17077

作者：C. Cabezas, Vagner S. Santos, Thiago R. Ramos, Pedro L. C. Rodrigues, Rafael Izbicki

【12】Analysis of Transferability Estimation Metrics for Surgical Phase Recognition
标题：手术阶段识别的可移植性估计子菜单分析
链接：https://arxiv.org/abs/2508.16730

作者：Singh, Yiping Li, Yasmina Al Khalil
备注：Accepted at DEMI workshop MICCAI 2025

检测相关(13篇)

【1】BirdRecorder's AI on Sky: Safeguarding birds of prey by detection and classification of tiny objects around wind turbines
标题：BirdRecorder在Sky上的人工智能：通过检测和分类风力涡轮机周围的微小物体来保护猛禽
链接：https://arxiv.org/abs/2508.18136

作者：, Nizam Gifary, Felix P. G. Ziegler, Frank Sehnke, Anton Kaifel, Eric Price, Aamir Ahmad
备注：18 pages, 1 figures, to appear in Proceedings of the 19th International Conference on Intelligent Autonomous Systems (IAS-19), Genoa, Italy, 2025

【2】Quantum-Classical Hybrid Framework for Zero-Day Time-Push GNSS Spoofing Detection
标题：用于零日时间推送式全球导航卫星欺骗检测的量子经典混合框架
链接：https://arxiv.org/abs/2508.18085

作者：n, Mashrur Chowdhury, Sagar Dasgupta, Mizanur Rahman
备注：This work has been submitted to the IEEE Internet of Things Journal for possible publication

【3】Riemannian Change Point Detection on Manifolds with Robust Centroid Estimation
标题：具有鲁棒性中心估计的流体上的Riemann变点检测
链接：https://arxiv.org/abs/2508.18045

作者：ang, Ricardo Borsoi, Arnaud Breloy, Cédric Richard

【4】Development of a Neural Network Model for Currency Detection to aid visually impaired people in Nigeria
标题：开发货币检测神经网络模型以帮助尼日利亚的视障人士
链接：https://arxiv.org/abs/2508.18012

作者：a Nwokoye, Desmond Moru

【5】Learning to Detect Label Errors by Making Them: A Method for Segmentation and Object Detection Datasets
标题：学习通过制造标签错误来检测标签错误：分割和对象检测数据集的方法
链接：https://arxiv.org/abs/2508.17930

作者：nquitt, Tobias Riedlinger, Timo Heller, Markus Reischl, Matthias Rottmann

【6】Interpretable Early Failure Detection via Machine Learning and Trace Checking-based Monitoring
标题：通过机器学习和基于轨迹检查的监控进行可解释的早期故障检测
链接：https://arxiv.org/abs/2508.17786

作者：unello, Luca Geatti, Angelo Montanari, Nicola Saccomanno
备注：Full version of the paper accepted for publication at the 28th European Conference on Artificial Intelligence (ECAI 2025)

【7】Text Meets Topology: Rethinking Out-of-distribution Detection in Text-Rich Networks
标题：文本符合布局：重新思考文本丰富网络中的分发外检测
链接：https://arxiv.org/abs/2508.17690

作者：g, Ruihong Qiu, Guangdong Bai, Zi Huang
备注：EMNLP2025 Main

【8】MahaParaphrase: A Marathi Paraphrase Detection Corpus and BERT-based Models
标题：Maha Paraphrase：Marathi Paraphrase检测数据库和基于BERT的模型
链接：https://arxiv.org/abs/2508.17444

作者：adhav, Abhay Shanbhag, Amogh Thakurdesai, Ridhima Sinare, Ananya Joshi, Raviraj Joshi

【9】Detecting Struggling Student Programmers using Proficiency Taxonomies
标题：使用熟练程度分类法检测陷入困境的学生程序员
链接：https://arxiv.org/abs/2508.17353

作者：artz, Roy Fairstein, Avi Segal, Kobi Gal
备注：appears at ECAI 2025

【10】Sharpness-Aware Geometric Defense for Robust Out-Of-Distribution Detection
标题：用于鲁棒性分布外检测的敏锐度感知几何防御
链接：https://arxiv.org/abs/2508.17174

作者：Li, Ming-Ching Chang, Wei-Chao Chen
备注：under review

【11】Out of Distribution Detection for Efficient Continual Learning in Quality Prediction for Arc Welding
标题：非分布检测实现电焊质量预测中的高效连续学习
链接：https://arxiv.org/abs/2508.16832

作者：hn, Jan Voets, Antonin Koenigsfeld, Hasan Tercan, Tobias Meisen
备注：Accepted at CIKM 2025 (Applied Research Papers)

【12】Leveraging the Christoffel Function for Outlier Detection in Data Streams
标题：利用Christoffel函数进行数据流中的异常值检测
链接：https://arxiv.org/abs/2508.16617

作者：harlet, Louise Travé-Massuyès, Jean-Bernard Lasserre, Marie-Véronique Le Lann, Youssef Miloudi
备注：None

【13】Entanglement Detection with Quantum-inspired Kernels and SVMs
标题：利用量子启发的核和支持器进行纠缠检测
链接：https://arxiv.org/abs/2508.17909

作者：nez-Sabiote, Michalis Skotiniotis, Jara J. Bermejo-Vega, Daniel Manzano, Carlos Cano

分类|识别(10篇)

【1】Limits of message passing for node classification: How class-bottlenecks restrict signal-to-noise ratio
标题：节点分类的消息传递限制：类瓶颈如何限制信噪比
链接：https://arxiv.org/abs/2508.17822

作者：Rubin, Sahil Loomba, Nick S. Jones

【2】Randomly Removing 50% of Dimensions in Text Embeddings has Minimal Impact on Retrieval and Classification Tasks
标题：随机删除文本嵌入中50%的维度对检索和分类任务的影响最小
链接：https://arxiv.org/abs/2508.17744

作者：keshita, Yurina Takeshita, Daniel Ruffinelli, Simone Paolo Ponzetto
备注：Accepted to EMNLP 2025 Main Conference, submitted version

【3】CausalSent: Interpretable Sentiment Classification with RieszNet
标题：发送的CASEARCH：使用RieszNet进行可解释的情绪分类
链接：https://arxiv.org/abs/2508.17576

作者：ees, Martin Pollack

【4】TANDEM: Temporal Attention-guided Neural Differential Equations for Missingness in Time Series Classification
标题：TANEM：时间序列分类中缺失的时间注意力引导神经微方程
链接：https://arxiv.org/abs/2508.17519

作者： Oh, Dong-Young Lim, Sungil Kim, Alex Bui

【5】Learning Interpretable Differentiable Logic Networks for Time-Series Classification
标题：学习用于时间序列分类的可解释可微逻辑网络
链接：https://arxiv.org/abs/2508.17512

作者：, Niraj K. Jha

【6】A Synthetic Dataset for Manometry Recognition in Robotic Applications
标题：机器人应用中的压力识别合成数据集
链接：https://arxiv.org/abs/2508.17468

作者：onio Rabelo Saraiva, Enzo Ferreira de Souza, Joao Manoel Herrera Pinheiro, Thiago H. Segreto, Ricardo V. Godoy, Marcelo Becker

【7】A Systematic Literature Review on Multi-label Data Stream Classification
标题：多标签数据流分类的系统文献综述
链接：https://arxiv.org/abs/2508.17455

作者：-Oliveira, E. R. F. Paiva, J. Gama, L. Khan, R. Cerri
备注：48 pages, 12 figures

【8】CLIFF: Continual Learning for Incremental Flake Features in 2D Material Identification
标题：CLFF：2D材料识别中增量片状特征的持续学习
链接：https://arxiv.org/abs/2508.17261

作者：andey, Xuan Bac Nguyen, Nicholas Borys, Hugh Churchill, Khoa Luu

【9】WHAR Datasets: An Open Source Library for Wearable Human Activity Recognition
标题：WHAR数据集：可穿戴人类活动识别的开源库
链接：https://arxiv.org/abs/2508.16604

作者：n Burzer, Tobias King, Till Riedel, Michael Beigl, Tobias Röddiger
备注：6 pages, 7 figures, to appear in Companion of the 2025 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp), OpenWearables Workshop (accepted paper)

【10】Hybrid Quantum-Classical Learning for Multiclass Image Classification
标题：用于多类图像分类的混合量子经典学习
链接：https://arxiv.org/abs/2508.18161

作者：ta Anwar, Sowmitra Das, Muhammad Iqbal Hossain, Jishnu Mahmud
备注：13 pages, 8 figures

表征(5篇)

【1】Multimodal Representation Learning Conditioned on Semantic Relations
标题：基于语义关系的多模式表示学习
链接：https://arxiv.org/abs/2508.17497

作者：, Yuntong Hu, Liang Zhao

【2】Learned Structure in CARTRIDGES: Keys as Shareable Routers in Self-Studied Representations
标题：CARTRIDGES中的学习结构：自学习表示中作为可共享路由器的密钥
链接：https://arxiv.org/abs/2508.17032

作者：Diaz

【3】Hyperbolic Multimodal Representation Learning for Biological Taxonomies
标题：生物分类学的双曲多模态表示学习
链接：https://arxiv.org/abs/2508.16744

作者：ng, Chuanqi Tang, Xiaoliang Huo, Nicholas Pellegrino, Austin T. Wang, Graham W. Taylor, Angel X. Chang, Scott C. Lowe, Joakim Bruslund Haurum

【4】Native Logical and Hierarchical Representations with Subspace Embeddings
标题：具有子空间嵌入的本地逻辑和分层表示
链接：https://arxiv.org/abs/2508.16687

作者：oreira, Zita Marinho, Manuel Marques, João Paulo Costeira, Chenyan Xiong

【5】Bridging Foundation Models and Efficient Architectures: A Modular Brain Imaging Framework with Local Masking and Pretrained Representation Learning
标题：桥梁基础模型和高效架构：具有局部掩蔽和预训练的表示学习的模块化大脑成像框架
链接：https://arxiv.org/abs/2508.16597

作者：ng, Xinglin Zhao, Yijin Song, Xiaobo Liu, Yanrong Hao, Rui Cao, Xin Wen

3D|3D重建等相关(1篇)

【1】Copyright Protection for 3D Molecular Structures with Watermarking
标题：带有水印的3D分子结构的版权保护
链接：https://arxiv.org/abs/2508.17702

作者：, Peilin Chen, Keyan Ding, Shiqi Wang

优化|敛散性(5篇)

【1】Riemannian Optimization for LoRA on the Stiefel Manifold
标题：Stiefel Manifold上LoRA的Riemann优化
链接：https://arxiv.org/abs/2508.17901

作者： Park, Minjae Kang, Seongbae Lee, Haegang Lee, Seongwan Kim, Jaeho Lee
备注：EMNLP 2025 Findings

【2】Convergence and Generalization of Anti-Regularization for Parametric Models
标题：参数模型反正规化的收敛性和推广
链接：https://arxiv.org/abs/2508.17412

作者：Kim, Wonjun Jeong, Gisung Oh
备注：39 pages, 1 figure

【3】Aligning Distributionally Robust Optimization with Practical Deep Learning Needs
标题：将分布稳健优化与实际深度学习需求保持一致
链接：https://arxiv.org/abs/2508.16734

作者：eoktistov, Igor Ignashin, Andrey Veprikov, Nikita Borovko, Alexander Bogdanov, Savelii Chezhegov, Aleksandr Beznosikov
备注：13 pages, 1 table, 4 figures

【4】Sparse and Dense Retrievers Learn Better Together: Joint Sparse-Dense Optimization for Text-Image Retrieval
标题：稀疏和密集检索器一起学习更好：文本图像检索的稀疏-密集联合优化
链接：https://arxiv.org/abs/2508.16707

作者：Song, Youngjune Lee, Gyu-Hwung Cho, Ilhyeon Song, Saehun Kim, Yohan Jo
备注：accepted to CIKM 2025 short research paper track

【5】Quantum-Inspired DRL Approach with LSTM and OU Noise for Cut Order Planning Optimization
标题：具有LSTM和RST噪音的量子启发DRL方法用于切割订单规划优化
链接：https://arxiv.org/abs/2508.16611

作者：erry Chrisnanto, Julian Evan Chrisnanto
备注：14 pages,3 figures, 4 tables

预测|估计(13篇)

【1】WOMAC: A Mechanism For Prediction Competitions
标题：WOMAC：预测竞争机制
链接：https://arxiv.org/abs/2508.17907

作者：Srinivasan, Tao Lin, Connacher Murphy, Anish Thilagar, Yiling Chen, Ezra Karger

【2】ControlEchoSynth: Boosting Ejection Fraction Estimation Models via Controlled Video Diffusion
标题：Control EchoSynth：通过受控视频扩散增强射弹分数估计模型
链接：https://arxiv.org/abs/2508.17631

作者：ori, Hanwen Liang, Hooman Vaseli, Bingyu Xie, Christina Luong, Purang Abolmaesumi, Teresa Tsang, Renjie Liao
备注：Data Curation and Augmentation in Medical Imaging CVPR 2024

【3】GateTS: Versatile and Efficient Forecasting via Attention-Inspired routed Mixture-of-Experts
标题：Gateway TS：通过注重注意力的路由混合专家进行多功能、高效的预测
链接：https://arxiv.org/abs/2508.17515

作者：mets, Mykola Lukashchuk, Ivan Izonin

【4】A Human-In-The-Loop Approach for Improving Fairness in Predictive Business Process Monitoring
标题：提高预测性业务流程监控公平性的人在环方法
链接：https://arxiv.org/abs/2508.17477

作者：ppel, Julian Neuberger, Felix Möhrlein, Sven Weinzierl, Martin Matzner, Stefan Jablonski

【5】Physics-informed neural network for fatigue life prediction of irradiated austenitic and ferritic/martensitic steels
标题：物理信息神经网络用于预测受辐射的奥氏体和铁氏体/马氏体钢疲劳寿命
链接：https://arxiv.org/abs/2508.17303

作者：Kori, Abhinav Chandraker, Syed Abdur Rahman, Punit Rathore, Ankur Chauhan

【6】MoE-Beyond: Learning-Based Expert Activation Prediction on Edge Devices
标题：MoE-Beyond：边缘设备上基于学习的专家激活预测
链接：https://arxiv.org/abs/2508.17137

作者：avhane, Arush Mehrotra, Rohit Chawla, Peter Proenca

【7】Walk-on-Interfaces: A Monte Carlo Estimator for an Elliptic Interface Problem with Nonhomogeneous Flux Jump Conditions and a Neumann Boundary Condition
标题：步行界面：具有非齐次通量跳变条件和诺伊曼边界条件的椭圆界面问题的Monte Carlo估计
链接：https://arxiv.org/abs/2508.16767

作者：ng, Adam R Stinchcombe
备注：49 pages, 14 figures

【8】CellEcoNet: Decoding the Cellular Language of Pathology with Deep Learning for Invasive Lung Adenocarcinoma Recurrence Prediction
标题：CellEcoNet：利用深度学习解码病理学的细胞语言，用于侵袭性肺腺癌复发预测
链接：https://arxiv.org/abs/2508.16742

作者：man Akbar, Usama Sajjad, Ziyu Su, Wencheng Li, Fei Xing, Jimmy Ruiz, Wei Chen, Muhammad Khalid Khan Niazi

【9】COVID19 Prediction Based On CT Scans Of Lungs Using DenseNet Architecture
标题：使用DenseNet架构基于肺部CT扫描的COVID 19预测
链接：https://arxiv.org/abs/2508.16670

作者：anyal

【10】STRelay: A Universal Spatio-Temporal Relaying Framework for Location Prediction with Future Spatiotemporal Contexts
标题：STRelay：一个通用时空中继框架，用于未来时空上下文的位置预测
链接：https://arxiv.org/abs/2508.16620

作者：Deng, Lianhua Ji, Chunhua Chen, Xin Jing, Ling Ding, Bingqing QU, Pengyang Wang, Dingqi Yang

【11】Predicting User Grasp Intentions in Virtual Reality
标题：预测虚拟现实中的用户掌握意图
链接：https://arxiv.org/abs/2508.16582

作者：eng
备注：45 pages, 24 figures. This is a Master's thesis submitted as part of the M2 IASD (Artificial Intelligence, Systems, Data) program at Université PSL

【12】Boltzina: Efficient and Accurate Virtual Screening via Docking-Guided Binding Prediction with Boltz-2
标题：Boltzina：通过Boltz-2对接引导结合预测高效准确的虚拟筛选
链接：https://arxiv.org/abs/2508.17555

作者：ui, Masahito Ohue

【13】Factor Informed Double Deep Learning For Average Treatment Effect Estimation
标题：因子知情双重深度学习用于平均治疗效果估计
链接：https://arxiv.org/abs/2508.17136

作者：Fan, Soham Jana, Sanjeev Kulkarni, Qishuo Yin
备注：41 pages, 3 figures, 4 tables

其他神经网络|深度学习|模型|建模(39篇)

【1】Deep Learning and Matrix Completion-aided IoT Network Localization in the Outlier Scenarios
标题：离群场景中的深度学习和矩阵完成辅助物联网网络本地化
链接：https://arxiv.org/abs/2508.18225

作者：m
备注：4 pages, 2 figures

【2】HypER: Hyperbolic Echo State Networks for Capturing Stretch-and-Fold Dynamics in Chaotic Flows
标题：HypER：用于捕捉混乱流中的拉伸折叠动力学的双曲回声状态网络
链接：https://arxiv.org/abs/2508.18196

作者：ingh, Sutirtha Ghosh, Ashutosh Kumar, Hrishit B P, Balasubramanian Raman
备注：8 pages, accepted in ECAI 2025

【3】Emerging Semantic Segmentation from Positive and Negative Coarse Label Learning
标题：来自积极和消极粗标签学习的新兴语义分割
链接：https://arxiv.org/abs/2508.18186

作者： Fuping Wu, Arun Thirunavukarasu, Kevin Bronik, Thomas Nichols, Bartlomiej W. Papiez

【4】Introduction to Regularization and Learning Methods for Inverse Problems
标题：反问题的正规化和学习方法简介
链接：https://arxiv.org/abs/2508.18178

作者：Bednarski, Tim Roith
备注：These lecture notes are based on a lecture taught by the authors in the winter semester 2024/2025 at the University of Hamburg

【5】The Computational Complexity of Satisfiability in State Space Models
标题：状态空间模型中可满足性的计算复杂性
链接：https://arxiv.org/abs/2508.18162

作者：ann, Martin Lange
备注：Accepted at ECAI 25

【6】Assessing the Noise Robustness of Class Activation Maps: A Framework for Reliable Model Interpretability
标题：评估类激活地图的噪音稳健性：可靠模型解释性的框架
链接：https://arxiv.org/abs/2508.18154

作者： Sarkar, Revoti P. Bora, Bhupender Kaushal, Sudhish N George, Kiran Raja
备注：Image and Vision Computing (2025)

【7】Provable Mixed-Noise Learning with Flow-Matching
标题：具有流匹配的可证明混合噪音学习
链接：https://arxiv.org/abs/2508.18122

作者：mann, Robert Gruhlke, Bernhard Stankewitz, Claudia Schillings, Gabriele Steidl

【8】Incorporating Pre-trained Diffusion Models in Solving the Schrödinger Bridge Problem
标题：利用预先训练的扩散模型解决薛定汉桥问题
链接：https://arxiv.org/abs/2508.18095

作者：ang, Tiankai Hang, Shuyang Gu, Dong Chen, Baining Guo

【9】DesCartes Builder: A Tool to Develop Machine-Learning Based Digital Twins
标题：DesCartes Builder：开发基于机器学习的数字双胞胎的工具
链接：https://arxiv.org/abs/2508.17988

作者：e Conto, Blaise Genest, Arvind Easwaran, Nicholas Ng, Shweta Menon
备注：5 pages, 4 figures. Accepted at EDTconf 2025

【10】Puzzle: Scheduling Multiple Deep Learning Models on Mobile Device with Heterogeneous Processors
标题：谜题：在具有异类处理器的移动终端上调度多个深度学习模型
链接：https://arxiv.org/abs/2508.17764

作者：ng, Yunseong Lee, Junghoon Kim

【11】On the Edge of Memorization in Diffusion Models
标题：扩散模型中的子化边缘
链接：https://arxiv.org/abs/2508.17689

作者：nan, Druv Pai, Yi Ma, Valentin De Bortoli
备注：10 main body pages, 43 total pages

【12】Characterizing the Behavior of Training Mamba-based State Space Models on GPUs
标题：在图形处理器上描述训练基于Mamba的状态空间模型的行为
链接：https://arxiv.org/abs/2508.17679

作者：Baruah, Kaustubh Shivdikar, Sara Prescott, David Kaeli

【13】GWM: Towards Scalable Gaussian World Models for Robotic Manipulation
标题：GWM：迈向机器人操纵的可扩展高斯世界模型
链接：https://arxiv.org/abs/2508.17600

作者：Lu, Baoxiong Jia, Puhao Li, Yixin Chen, Ziwei Wang, Yansong Tang, Siyuan Huang
备注：Published at ICCV 2025. Project page: this https URL

【14】Exploring Efficient Learning of Small BERT Networks with LoRA and DoRA
标题：利用LoRA和DoRA探索小型BERT网络的有效学习
链接：https://arxiv.org/abs/2508.17586

作者：ees, Aditri Bhagirath, Moritz Bolling

【15】Modeling Irregular Astronomical Time Series with Neural Stochastic Delay Differential Equations
标题：用神经随机延迟方程建模不规则天文时间序列
链接：https://arxiv.org/abs/2508.17521

作者： Oh, Seungsu Kam, Dong-Young Lim, Sungil Kim

【16】Optimizing Grasping in Legged Robots: A Deep Learning Approach to Loco-Manipulation
标题：优化腿机器人的抓取：一种用于局部操纵的深度学习方法
链接：https://arxiv.org/abs/2508.17466

作者：o Almeida, Guilherme Lazzarini, Juliano Negri, Thiago H. Segreto, Ricardo V. Godoy, Marcelo Becker

【17】Modular MeanFlow: Towards Stable and Scalable One-Step Generative Modeling
标题：模块化MeanFlow：迈向稳定和可扩展的一步生成建模
链接：https://arxiv.org/abs/2508.17426

作者：ou, Baojing Liu, Hongyang He
备注：Accepted as a conference paper at PRCV 2025

【18】Who Wins the Race? (R Vs Python) - An Exploratory Study on Energy Consumption of Machine Learning Algorithms
标题：谁赢得了比赛？（R Vs Python）-机器学习算法能耗的探索性研究
链接：https://arxiv.org/abs/2508.17344

作者：hattaraj, Sridhar Chimalakonda, Vibhu Saujanya Sharma, Vikrant Kaulgud
备注：18 pages including references, 5 figures

【19】Quickly Tuning Foundation Models for Image Segmentation
标题：快速调整图像分割的基础模型
链接：https://arxiv.org/abs/2508.17283

作者：as, Lennart Purucker, Timur Carstensen, Frank Hutter
备注：Accepted as a short paper at the non-archival content track of AutoML 2025

【20】DeepCFD: Efficient near-ground airfoil lift coefficient approximation with deep convolutional neural networks
标题：DeepCFA：利用深度卷积神经网络实现高效的近地面机翼升力系数逼近
链接：https://arxiv.org/abs/2508.17278

作者：Amin Esabat, Saeed Jaamei, Fatemeh Asadi

【21】Learning Short-Term and Long-Term Patterns of High-Order Dynamics in Real-World Networks
标题：学习现实世界网络中高级动态的短期和长期模式
链接：https://arxiv.org/abs/2508.17236

作者：o, Da Eun Lee, Song Kyung Yu, Sang-Wook Kim
备注：5 pages, 4 figures, 2 tables, ACM International Conference on Information and Knowledge Management (CIKM) 2025

【22】Curvature Learning for Generalization of Hyperbolic Neural Networks
标题：基于曲率学习的双曲神经网络推广
链接：https://arxiv.org/abs/2508.17232

作者：Fan, Yuwei Wu, Zhi Gao, Mehrtash Harandi, Yunde Jia

【23】Reconciling Communication Compression and Byzantine-Robustness in Distributed Learning
标题：分布式学习中的通信压缩和拜占庭鲁棒性
链接：https://arxiv.org/abs/2508.17129

作者：pta, Nirupam Gupta, Chuan Xu, Giovanni Neglia
备注：78 Pages, 1 figure

【24】SugarcaneShuffleNet: A Very Fast, Lightweight Convolutional Neural Network for Diagnosis of 15 Sugarcane Leaf Diseases
标题：SugarcaneShuffleNet：一个非常快速、轻量级的卷积神经网络，用于诊断15种甘蔗叶片病
链接：https://arxiv.org/abs/2508.17107

作者： Arman, Hasan Muhammad Abdullah, Syed Nazmus Sakib, RM Saiem, Shamima Nasrin Asha, Md Mehedi Hasan, Shahrear Bin Amin, S M Mahin Abrar
备注：18 pages, 19 figures, Submitted in Computers and Electronics in Agriculture

【25】Convolutional Neural Networks for Accurate Measurement of Train Speed
标题：卷积神经网络用于精确测量列车速度
链接：https://arxiv.org/abs/2508.17096

作者：an, Argyrios Zolotas, Miguel Arana-Catania
备注：15 pages, 12 figures, 2 tables. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit

【26】Learning ON Large Datasets Using Bit-String Trees
标题：使用位串树学习大数据集
链接：https://arxiv.org/abs/2508.17083

作者：Gupta
备注：PhD thesis

【27】TabResFlow: A Normalizing Spline Flow Model for Probabilistic Univariate Tabular Regression
标题：TabResFlow：一个概率单变量表回归的规范化样条流模型
链接：https://arxiv.org/abs/2508.17056

作者：husudhanan, Vijaya Krishna Yalavarthi, Jonas Sonntag, Maximilian Stubbemann, Lars Schmidt-Thieme
备注：To be published in The European Conference on Artificial Intelligence, 2025

【28】KL-Regularised Q-Learning: A Token-level Action-Value perspective on Online RLHF
标题：KL规范化的Q-Learning：在线RL HF的代币级目标-价值视角
链接：https://arxiv.org/abs/2508.17000

作者：rown, Lennie Wells, Edward James Young, Sergio Bacallado

【29】NinA: Normalizing Flows in Action. Training VLA Models with Normalizing Flows
标题：NinA：正常化流程正在行动。通过规范化流程训练VLA模型
链接：https://arxiv.org/abs/2508.16845

作者：asov, Alexander Nikulin, Ilya Zisman, Albina Klepach, Nikita Lyubaykin, Andrei Polubarov, Alexander Derevyagin, Vladislav Kurenkov

【30】TaDiCodec: Text-aware Diffusion Speech Tokenizer for Speech Language Modeling
标题：TaDiCodec：用于语音语言建模的文本感知扩散语音令牌器
链接：https://arxiv.org/abs/2508.16790

作者： Wang, Dekun Chen, Xueyao Zhang, Junan Zhang, Jiaqi Li, Zhizheng Wu

【31】Deep Learning for Markov Chains: Lyapunov Functions, Poisson's Equation, and Stationary Distributions
标题：马尔可夫链的深度学习：李雅普诺夫函数、泊松方程和平稳分布
链接：https://arxiv.org/abs/2508.16737

作者：, Jose Blanchet, Peter Glynn

【32】A novel auxiliary equation neural networks method for exactly explicit solutions of nonlinear partial differential equations
标题：一种新型辅助方程神经网络方法精确显式解非线性偏微方程
链接：https://arxiv.org/abs/2508.16702

作者：uan, Yanqin Liu, Runfa Zhang, Limei Yan, Shunjun Wu, Libo Feng

【33】HiCL: Hippocampal-Inspired Continual Learning
标题：HiCL：受河马启发的持续学习
链接：https://arxiv.org/abs/2508.16651

作者：poor, Wyatt Mackey, Yiannis Aloimonos, Xiaomin Lin
备注：Submitted to AAAI

【34】From Classical Probabilistic Latent Variable Models to Modern Generative AI: A Unified Perspective
标题：从经典概率潜在变量模型到现代生成人工智能：统一视角
链接：https://arxiv.org/abs/2508.16643

作者：hen
备注：This is a substantially improved and expanded version of an earlier manuscript hosted on SSRN: this https URL

【35】GreenTEA: Gradient Descent with Topic-modeling and Evolutionary Auto-prompting
标题：GreenTEA：具有主题建模和进化自动提示的渐变下降
链接：https://arxiv.org/abs/2508.16603

作者：g, Luming Shang, Gabriela Olinto

【36】Increasing Interaction Fidelity: Training Routines for Biomechanical Models in HCI
标题：提高交互保真度：人机交互中生物力学模型训练课程
链接：https://arxiv.org/abs/2508.16581

作者：tryk Miazga, Patrick Ebel
备注：None

【37】Algebraic Approach to Ridge-Regularized Mean Squared Error Minimization in Minimal ReLU Neural Network
标题：最小ReLU神经网络中岭正则均方误差最小化的代数方法
链接：https://arxiv.org/abs/2508.17783

作者：asaku, Yutaro Kabata, Akifumi Okuno
备注：44 pages, 5 figres

【38】Programmable k-local Ising Machines and all-optical Kolmogorov-Arnold Networks on Photonic Platforms
标题：光子平台上的可编程k-局域Ising机和全光Kolmogorov-Arnold网络
链接：https://arxiv.org/abs/2508.17440

作者：roev, Natalia G. Berloff
备注：16 pages, 6 figures

【39】Predictability Enables Parallelization of Nonlinear State Space Models
标题：非线性状态空间模型的可预测性实现了可预测化
链接：https://arxiv.org/abs/2508.16817

作者：nzalez, Leo Kozachkov, David M. Zoltowski, Kenneth L. Clarkson, Scott W. Linderman

其他(55篇)

【1】ANO : Faster is Better in Noisy Landscape
标题：ANO：在喧闹的环境中，速度越快越好
链接：https://arxiv.org/abs/2508.18258

作者：greisz
备注：Work in progress, 26 pages total with appendix, 7 figures, 12 tables

【2】Aligning the Evaluation of Probabilistic Predictions with Downstream Value
标题：将概率预测的评估与下游价值保持一致
链接：https://arxiv.org/abs/2508.18251

作者：hroudi, Viacheslav Komisarenko, Meelis Kull

【3】Flash Sparse Attention: An Alternative Efficient Implementation of Native Sparse Attention Kernel
标题：Flash稀疏注意力：原生稀疏注意力核心的替代有效实现
链接：https://arxiv.org/abs/2508.18224

作者：Youhe Jiang, Binhang Yuan

【4】Practical GPU Choices for Earth Observation: ResNet-50 Training Throughput on Integrated, Laptop, and Cloud Accelerators
标题：用于地球观测的实用图形处理器选择：集成、笔记本电脑和云加速器上的ResNet-50训练投入
链接：https://arxiv.org/abs/2508.18206

作者：aturvedi
备注：10 pages, 5 figures

【5】Scene-Aware Vectorized Memory Multi-Agent Framework with Cross-Modal Differentiated Quantization VLMs for Visually Impaired Assistance
标题：具有跨模式区分量化VLM的场景感知载体存储多代理框架用于视觉障碍的协助
链接：https://arxiv.org/abs/2508.18177

作者：g Wang, Xuanyu Wang, YiJia Luo, Yongbin Yu, Manping Fan, Jingtao Zhang, Liyong Ren
备注：28 pages,9 figures

【6】Amortized Sampling with Transferable Normalizing Flows
标题：具有可转移正化流的摊销抽样
链接：https://arxiv.org/abs/2508.18175

作者：. Tan, Majdi Hassan, Leon Klein, Saifuddin Syed, Dominique Beaini, Michael M. Bronstein, Alexander Tong, Kirill Neklyudov

【7】SpotEdit: Evaluating Visually-Guided Image Editing Methods
标题：SpotEdit：评估视觉引导图像编辑方法
链接：https://arxiv.org/abs/2508.18159

作者：anfari, Wei-An Lin, Haitong Tian, Ersin Yumer

【8】The AI Data Scientist
标题：人工智能数据科学家
链接：https://arxiv.org/abs/2508.18113

作者：kimov, Munachiso Samuel Nwadike, Zangir Iklassov, Martin Takáč

【9】Enhancing Differentially Private Linear Regression via Public Second-Moment
标题：通过公共二次时刻增强差异私人线性回归
链接：https://arxiv.org/abs/2508.18037

作者：o (1), Hai Zhang (1) ((1) The School of Mathematics, Northwest University)

【10】Topology Aware Neural Interpolation of Scalar Fields
标题：纯量场的结构感知神经内插
链接：https://arxiv.org/abs/2508.17995

作者：issi, Keanu Sisouk, Joshua A. Levine, Julien Tierny

【11】Generative Feature Imputing - A Technique for Error-resilient Semantic Communication
标题：生成性特征输入--一种抗错误语义通信技术
链接：https://arxiv.org/abs/2508.17957

作者：uang, Qunsong Zeng, Hongyang Du, Kaibin Huang

【12】Vocoder-Projected Feature Discriminator
标题：声码器投影特征鉴别器
链接：https://arxiv.org/abs/2508.17874

作者：Kaneko, Hirokazu Kameoka, Kou Tanaka, Yuto Kondo
备注：Accepted to Interspeech 2024. Project page: this https URL

【13】Alternating Training-based Label Smoothing Enhances Prompt Generalization
标题：交替基于训练的标签平滑增强提示概括
链接：https://arxiv.org/abs/2508.17846

作者：, Yanbin Wei, Ke Jin, Yi Kong, James Kwok, Yu Zhang

【14】Limitations of Normalization in Attention Mechanism
标题：注意机制正常化的局限性
链接：https://arxiv.org/abs/2508.17821

作者：arisov, Mikhail Burtsev, Tatiana Petrova, Radu State
备注：10 pages, 4 figures

【15】MeshSplat: Generalizable Sparse-View Surface Reconstruction via Gaussian Splatting
标题：MeshSplat：通过高斯飞溅的可推广稀疏视图表面重建
链接：https://arxiv.org/abs/2508.17811

作者：ang, Ruijie Zhu, Wenjie Chang, Mulin Yu, Yanzhe Liang, Jiahao Lu, Zhuoyuan Li, Tianzhu Zhang
备注：17 pages, 15 figures, 5 tables

【16】ISACL: Internal State Analyzer for Copyrighted Training Data Leakage
标题：ISAL：受版权保护的训练数据泄露的内部状态分析器
链接：https://arxiv.org/abs/2508.17767

作者：Zhang, Qisheng Su, Jiateng Liu, Cheng Qian, Yanzhou Pan, Yanjie Fu, Denghui Zhang

【17】Speculative Safety-Aware Decoding
标题：推测性安全意识解码
链接：https://arxiv.org/abs/2508.17739

作者：ang, Shengyu Zhu, Xueqi Cheng
备注：EMNLP'2025 main conference; more experiments will be added to the coming camera-ready version

【18】Unlearning as Ablation: Toward a Falsifiable Benchmark for Generative Scientific Discovery
标题：放弃学习作为消融：迈向生成性科学发现的可证伪基准
链接：https://arxiv.org/abs/2508.17681

作者：ng
备注：6 pages. NeurIPS 2025 AI4Science Workshop submission

【19】Heterogeneous co-occurrence embedding for visual information exploration
标题：用于视觉信息探索的异类共生嵌入
链接：https://arxiv.org/abs/2508.17663

作者：hida, Tetsuo Furukawa
备注：36pages, 9 figures, Accepted to International Journal of Innovative Computing, Information and Control (IJICIC), 2025

【20】Spacer: Towards Engineered Scientific Inspiration
标题：间隔物：迈向工程科学灵感
链接：https://arxiv.org/abs/2508.17661

作者： Lee, Suyoung Hwang, Seunghyun Moon, Geonho Nah, Donghyun Koh, Youngjun Cho, Johyun Park, Hojin Yoo, Jiho Park, Haneul Choi, Sungbin Moon, Taehoon Hwang, Seungwon Kim, Jaeyeong Kim, Seongjun Kim, Juneau Jung

【21】Citizen Centered Climate Intelligence: Operationalizing Open Tree Data for Urban Cooling and Eco-Routing in Indian Cities
标题：以公民为中心的气候情报：在印度城市运营开放树数据以实现城市降温和生态路线
链接：https://arxiv.org/abs/2508.17648

作者：avi, Andreas Brück
备注：Forthcoming book chapter, currently under review for the "HackYourDistrict" initiative at TU Berlin. 20 pages, 9 figures, 1 table

【22】Consciousness as a Functor
标题：作为功能者的意识
链接：https://arxiv.org/abs/2508.17561

作者：ahadevan
备注：31 pages

【23】LodeStar: Long-horizon Dexterity via Synthetic Data Augmentation from Human Demonstrations
标题：LodeStar：通过人类演示的合成数据增强实现长期灵活性
链接：https://arxiv.org/abs/2508.17547

作者：an, Jiawei Fu, Xiaodi Yuan, Yifeng Zhu, Hao Su
备注：CoRL 2025

【24】Activation Transport Operators
标题：激活运输运营商
链接：https://arxiv.org/abs/2508.17540

作者：zablewski, Marek Masiak
备注：4 pages, 4 figures, references and appendices

【25】Mutual Information Surprise: Rethinking Unexpectedness in Autonomous Systems
标题：相互信息惊喜：重新思考自治系统中的意外
链接：https://arxiv.org/abs/2508.17403

作者：ang, Xiao Liu, Quan Zeng, Yu Ding
备注：Pre-Submission Version

【26】ShaLa: Multimodal Shared Latent Space Modelling
标题：ShaLa：多模式共享潜空间建模
链接：https://arxiv.org/abs/2508.17376

作者：, Yan-Ying Chen, Yanxia Zhang, Matthew Klenk

【27】MetaFed: Advancing Privacy, Performance, and Sustainability in Federated Metaverse Systems
标题：MetaFed：在联邦元宇宙系统中推进隐私、性能和可持续性
链接：https://arxiv.org/abs/2508.17341

作者：Anil Yagiz, Zeynep Sude Cengiz, Polat Goktas
备注：2025 IEEE International Symposium on Emerging Metaverse (ISEMV)

【28】Mind the (Language) Gap: Towards Probing Numerical and Cross-Lingual Limits of LVLMs
标题：注意（语言）差距：探索LVLM的数值和跨语言限制
链接：https://arxiv.org/abs/2508.17334

作者：utam, Abhirama Subramanyam Penamakuri, Abhishek Bhandari, Gaurav Harit

【29】Is the Frequency Principle always valid?
标题：频率原则总是有效的吗？
链接：https://arxiv.org/abs/2508.17323

作者：i

【30】MEENA (PersianMMMU): Multimodal-Multilingual Educational Exams for N-level Assessment
标题：MEENA（PersianMMMU）：N级评估的多模式多语言教育考试
链接：https://arxiv.org/abs/2508.17290

作者：roodi, Arshia Hemmat, Marzia Nouri, Seyed Mohammad Hadi Hosseini, Doratossadat Dastgheib, Mohammad Vali Sanian, Alireza Sahebi, Reihaneh Zohrabi, Mohammad Hossein Rohban, Ehsaneddin Asgari, Mahdieh Soleymani Baghshah

【31】Provable Generalization in Overparameterized Neural Nets
标题：过度参数化神经网络中的可证明推广
链接：https://arxiv.org/abs/2508.17256

作者：ingra
备注：8 Pages

【32】Multi-Metric Preference Alignment for Generative Speech Restoration
标题：生成式语音恢复中的多度量偏好对齐
链接：https://arxiv.org/abs/2508.17229

作者：ng, Xueyao Zhang, Jing Yang, Yuancheng Wang, Fan Fan, Zhizheng Wu
备注：16 pages, 10 figures. demopage: this https URL

【33】VROOM - Visual Reconstruction over Onboard Multiview
标题：VROOM -车载多视图视觉重建
链接：https://arxiv.org/abs/2508.17172

作者：av, Varun Bharadwaj, Jathin Korrapati, Tanish Baranwal
备注：Project page with videos and interactive 4D visualizations: this https URL, Code: this https URL

【34】ONG: Orthogonal Natural Gradient Descent
标题：ONG：垂直自然梯度下降
链接：https://arxiv.org/abs/2508.17169

作者：av, Jathin Korrapati, Patrick Mendoza
备注：Code at this https URL

【35】Stochastic Gradient Descent with Strategic Querying
标题：具有战略查询的随机梯度下降
链接：https://arxiv.org/abs/2508.17144

作者：ang, Hoi-To Wai, Mahnoosh Alizadeh
备注：18 pages, 2 figures. Accepted to IEEE Conference on Decision and Control (CDC) 2025. Includes appendix and supplementary discussion

【36】Token Homogenization under Positional Bias
标题：位置偏差下的令牌均匀化
链接：https://arxiv.org/abs/2508.17126

作者：v Yusupov, Danil Maksimov, Ameliia Alaeva, Tatiana Zaitceva, Antipina Anna, Anna Vasileva, Chenlin Liu, Rayuth Chheng, Danil Sazanakov, Andrey Chetvergov, Alina Ermilova, Egor Shvetsov

【37】Enhancing Knowledge Tracing through Leakage-Free and Recency-Aware Embeddings
标题：通过无泄漏和恢复意识嵌入增强知识追踪
链接：https://arxiv.org/abs/2508.17092

作者：ran, Christine Preisach

【38】Preserving Domain Generalization in Fine-Tuning via Joint Parameter Selection
标题：通过联合参数选择在微调中保持域推广
链接：https://arxiv.org/abs/2508.16976

作者：Shiyu Shen, Zongbin Wang, Zhenwei Shi, Xia Xu

【39】Disentangling Polysemantic Neurons with a Null-Calibrated Polysemanticity Index and Causal Patch Interventions
标题：用零校准的多义性指数和因果补片干预解开多义性神经元
链接：https://arxiv.org/abs/2508.16950

作者：ta, Dhruv Kumar
备注：Under review. 13 pages

【40】Attention Layers Add Into Low-Dimensional Residual Subspaces
标题：注意层添加到低维剩余子空间
链接：https://arxiv.org/abs/2508.16929

作者：ang, Xuyang Ge, Wentao Shu, Zhengfu He, Xipeng Qiu

【41】Anchor-MoE: A Mean-Anchored Mixture of Experts For Probabilistic Regression
标题：锚定-教育部：概率回归的平均锚定专家混合体
链接：https://arxiv.org/abs/2508.16802

作者：u, Zhengxian Qu

【42】Dynamic Sparse Attention on Mobile SoCs
标题：移动SOC上的动态稀疏注意力
链接：https://arxiv.org/abs/2508.16703

作者：Yin, Daliang Xu, Mengwei Xu, Gang Huang, Xuanzhe Liu
备注：Technical Report

【43】QueryBandits for Hallucination Mitigation: Exploiting Semantic Features for No-Regret Rewriting
标题：用于缓解幻觉的CredyBandits：利用语义特征进行无悔重写
链接：https://arxiv.org/abs/2508.16697

作者：o, William Watson, Alec Koppel, Sumitra Ganesh, Manuela Veloso

【44】LatentFlow: Cross-Frequency Experimental Flow Reconstruction from Sparse Pressure via Latent Mapping
标题：LatentFlow：通过潜伏映射从稀疏压力重建跨频实验流量
链接：https://arxiv.org/abs/2508.16648

作者：, Chang Liu, Yanyu Ke, Qiuxiang Huang, Jiachen Zhao, Wenliang Chen, K.T. Tse, Gang Hu
备注：The paper is submitted to IAAI26. Total 9 pages with 8 figures

【45】Multimodal Appearance based Gaze-Controlled Virtual Keyboard with Synchronous Asynchronous Interaction for Low-Resource Settings
标题：基于多模态外观的低资源环境下同步异步交互的注视控制虚拟键盘
链接：https://arxiv.org/abs/2508.16606

作者：mar Meena, Manish Salvi

【46】Flexibility-Conditioned Protein Structure Design with Flow Matching
标题：流动匹配的相容性条件蛋白质结构设计
链接：https://arxiv.org/abs/2508.18211

作者：Viliuga, Leif Seute, Nicolas Wolf, Simon Wagner, Arne Elofsson, Jan Stühmer, Frauke Gräter
备注：ICML 2025

【47】The Statistical Fairness-Accuracy Frontier
标题：统计公平性-准确性前沿
链接：https://arxiv.org/abs/2508.17622

作者：allah, Michael I. Jordan, Annie Ulichney

【48】High-Order Langevin Monte Carlo Algorithms
标题：高级Langevin Monte Carlo算法
链接：https://arxiv.org/abs/2508.17545

作者：g, Mert Gurbuzbalaban, Mohammad Rafiqul Islam, Nian Yao, Lingjiong Zhu
备注：73 pages, 3 figures, 1 table

【49】Integrative Experiments Identify How Punishment Impacts Welfare in Public Goods Games
标题：综合实验确定公共产品游戏中惩罚如何影响福利
链接：https://arxiv.org/abs/2508.17151

作者：Alsobay, David G. Rand, Duncan J. Watts, Abdullah Almaatouq

【50】Rao Differential Privacy
标题：差异Rao隐私
链接：https://arxiv.org/abs/2508.17135

作者：to
备注：13 pages

【51】HV Metric For Time-Domain Full Waveform Inversion
标题：时间域全波倒置的高压指标
链接：https://arxiv.org/abs/2508.17122

作者：mann, Yunan Yang
备注：30 Pages

【52】Neural Stochastic Differential Equations on Compact State-Spaces
标题：紧状态空间上的神经随机微分方程
链接：https://arxiv.org/abs/2508.17090

作者：Liu, Malinda Lu, Matthew K. Nock, Yaniv Yacoby
备注：Accepted at Methods and Opportunities at Small Scale (MOSS), ICML 2025, Vancouver, Canada

【53】Limitations of refinement methods for weak to strong generalization
标题：从弱到强推广的精化方法的局限性
链接：https://arxiv.org/abs/2508.17018

作者：merstep, Ya'acov Ritov, Mikhail Yurochkin, Subha Maity, Yuekai Sun
备注：COLM 2025

【54】The compressible Neural Particle Method for Simulating Compressible Viscous Fluid Flows
标题：模拟可压缩粘性流体流的可压缩粒子神经方法
链接：https://arxiv.org/abs/2508.16916

作者：ibukawa, Naoya Ozaki, Maximilien Berthet
备注：13 pages, 5 figures, submitted to PASJ

【55】HemePLM-Diffuse: A Scalable Generative Framework for Protein-Ligand Dynamics in Large Biomolecular System
标题：HemePLM-diffuse：大生物分子系统中蛋白质配体动力学的可扩展生成框架
链接：https://arxiv.org/abs/2508.16587

作者：akur, Riya Gupta
备注：7 pages, 9 figures and 1 table

机器翻译由腾讯交互翻译提供，仅供参考

点击“阅读原文”获取带摘要的学术速递