arXivDaily arXiv每日学术速递 周一至周五更新

期刊&会议

NeurIPS

Conference on Neural Information Processing Systems · 会议 · Machine Learning

共收录 17318
2608.12416 2026-08-14 cs.RO 新提交

RoboSynChallenge: Mastering Real-World Dexterity via Generalizing Synthesized Manipulation Skills

RoboSynChallenge:通过泛化合成操作技能掌握真实世界灵巧操作

Runyi Zhao, Ruixin Wu, Chengkun Li, Hongrui Zhang, Ang Li, Ruixing Jin, Yueci Deng, Yingying Guo, Lihe Ding, Shaocong Dong, Tianfan Xue, Yanjun Gao, Yudong Luo, Pascal Poupart, Simo Wu, Kui Jia, Wei-shi Zheng, Guiliang Liu

机构 * The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳)) DexForce(德克斯力公司) The Chinese University of Hong Kong(香港中文大学) The Hong Kong University of Science and Technology(香港科技大学) University of Colorado Anschutz(科罗拉多大学安舒茨医学中心) Mila - Quebec AI Institute(米拉-魁北克人工智能研究所) Vector Institute(向量研究所) University of Waterloo(滑铁卢大学) Fudan University(复旦大学) Sun Yat-sen University(中山大学) Shenzhen Loop Area Institute (SLAI)(深圳环区研究所)

AI总结 RoboSynChallenge竞赛推出统一基准,结合合成数据与真实评估,提供多类基准策略,旨在推动开发泛化性强、数据高效的机器人操作系统,助力通用机器人智能发展。

Comments NeurIPS 2026 Competition Track

详情

展开后加载摘要…

URL PDF HTML 收藏
2510.24232 2026-08-14 cs.CV 版本更新

Delving into Cascaded Instability: A Lipschitz Continuity View on Image Restoration and Object Detection Synergy

深入探讨级联不稳定:从Lipschitz连续性视角看图像恢复与目标检测的协同

Qing Zhao, Weijian Deng, Pengxu Wei, ZiYi Dong, Hannan Lu, Xiangyang Ji, Liang Lin

机构 * Sun Yat-sen University(中山大学) Australian National University(澳大利亚国立大学) Harbin Institute of Technology(哈尔滨工业大学) Tsinghua University(清华大学) Peng Cheng Laboratory(鹏城实验室)

AI总结 通过Lipschitz连续性视角,提出LR-YOLO框架,将图像恢复与目标检测整合,提升检测稳定性与准确性。

Comments NeurIPS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.20346 2026-08-13 cs.LG hep-ex physics.ins-det 交叉投稿

Inverse Autoregressive Flows for Zero Degree Calorimeter fast simulation

逆自回归流用于零度 calorimeter 快速模拟

Emilia Majerz, Witold Dzwinel, Jacek Kitowski

机构 * AGH University of Krakow(克拉科夫AGH大学)

AI总结 本文提出逆自回归流方法,通过引入新的损失函数和缩放机制,提升ZDC模拟的准确性和速度,比现有方法快421倍。

Comments Presented as a poster at the Machine Learning and the Physical Sciences Workshop, 39th Conference on Neural Information Processing Systems (NeurIPS), 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.10435 2026-08-12 cs.CV 新提交

DynaPPI: A Large-scale Dynamic Protein Dataset for AI-driven Advances in Protein Interactomics

DynaPPI:面向AI驱动蛋白质互作组学进展的大规模动态蛋白质数据集

Jiabao Wei, Zilong Geng, Yuze Wang, Jianjun Li, Ning Ding, Bowen Zhou, Bing Zhang, Zhiyuan Ma

机构 * BIT(北京理工大学) HUST(华中科技大学) SJTU(上海交通大学) Tsinghua University(清华大学) Shanghai AI Laboratory(上海人工智能实验室)

AI总结 本研究提出DynaPPI动态蛋白质数据集,填补现有数据集忽略多体结合动态过程的空白,助力扩散模型预测未知蛋白质复合物结构,推动AI驱动的蛋白质互作组学发展。

Comments 9 pages, 1 figure, 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Workshop: AI4Science

详情

展开后加载摘要…

URL PDF HTML 收藏
2602.15159 2026-08-12 cs.LG 版本更新

Learning Representations from Incomplete EHR Data with Dual-Masked Autoencoding

从不完整电子健康记录数据中学习表示的双掩码自编码

Xiao Xiang, David Restrepo, Hyewon Jeong, Yugang Jia, Leo Anthony Celi

机构 * Massachusetts Institute of Technology(麻省理工学院) EPFL(苏黎世联邦理工学院) Harvard University(哈佛大学) Beth Israel Deaconess Medical Center(贝塞斯达医院)

AI总结 AID-MAE通过双掩码技术直接从不完整EHR时间序列中学习表示,有效提升临床任务性能并实现患者群体的自然分层。

Comments MLHC 2026 camera-ready. Spotlight at NeurIPS TS4H 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.23399 2026-08-12 cs.AI 版本更新

GAM-Agent: Game-Theoretic and Uncertainty-Aware Collaboration for Complex Visual Reasoning

GAM-Agent:面向复杂视觉推理的博弈论与感知不确定性感知协作框架

Jusheng Zhang, Yijia Fan, Wenjun Lin, Ruiqi Chen, Haoyi Jiang, Wenhao Chai, Jian Wang, Keze Wang

AI总结 该研究提出GAM-Agent博弈论多智能体框架,通过基础感知智能体与关键验证智能体的非零和博弈及不确定性感知协作,在四个视觉推理基准上显著提升了中小及强规模VLM的性能,为可靠可解释多模态推理提供了新路径。

Comments Accepted at NeurIPS 2025. Code available at https://github.com/jushengzhang/Gam-Agent

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.08081 2026-08-11 cs.NE cs.LG 新提交

RotaryQuant: Fitting 120B MoE Models on Consumer Hardware via Fused Compressed-Space Attention

RotaryQuant:通过融合压缩空间注意力在消费级硬件上运行1200亿参数的混合专家模型

Anthony. Lui, Mohamed. Elsaied, N. P. Savani

AI总结 RotaryQuant通过三轴压缩系统及IsoQuant等技术,在消费级硬件上实现1200亿参数MoE模型的高效运行,内存占用低且性能损失极小。

Comments 9 Pages, 9 Tables, Initial submission into NeurIPS 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.18566 2026-08-11 eess.SY cs.SY 版本更新

HJ-Gauss: A Monte-Carlo HJ Reachability Scheme

HJ-Gauss: 一种蒙特卡洛HJ可达性方案

Lekan Molu, Venkatraman Renganathan, Namhoon Cho

AI总结 本文提出了一种基于局部PDE线性化的方法,通过冻结系数采样方案解决高维系统中经典网格基HJ求解器内存消耗大的问题,实现了存储和网格无关的算法,适用于高维可达性分析。

Comments NeurIPS 2026 Submission

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.06522 2026-08-11 cs.CL cs.AI

Fixing It in Post: A Comparative Study of LLM Post-Training Data Quality and Model Performance

事后修复:对LLM事后训练数据质量和模型性能的比较研究

Aladin Djuhera, Swanand Ravindra Kadhe, Syed Zawad, Farhan Ahmed, Heiko Ludwig, Holger Boche

机构 * Technical University Munich(慕尼黑技术大学) IBM Research(IBM研究院)

AI总结 本文通过对比两个开源事后训练数据集,提出了一种系统化编纂方法,生成性能更优的TuluTalk数据集,提升模型表现。

Journal ref The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS), 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2509.16625 2026-08-11 cs.LG cs.CR

Self-Supervised Learning of Graph Representations for Network Intrusion Detection

图表示学习用于网络入侵检测的自监督学习

Lorenzo Guerra, Thomas Chapuis, Guillaume Duc, Pavlo Mozharovskyi, Van-Tam Nguyen

AI总结 GraphIDS通过自监督学习统一图表示学习与异常检测,利用掩码自动编码器和Transformer架构,在网络入侵检测中实现高精度性能。

Comments Accepted at NeurIPS 2025

Journal ref Advances in Neural Information Processing Systems 38 (NeurIPS 2025)

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.10465 2026-08-11 cs.LG cs.AI cs.CL

Superposition Yields Robust Neural Scaling

叠加产生稳健的神经扩展

Yizhou Liu, Ziming Liu, Jeff Gore

机构 * Massachusetts Institute of Technology(麻省理工学院)

AI总结 研究发现表示叠加是神经扩展定律的核心驱动因素,揭示了损失与模型规模之间的反比关系。

Comments Best Paper Runner-up at NeurIPS 2025

Journal ref Advances in Neural Information Processing Systems 38 (2025) 159269--159305

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.09816 2026-08-11 q-bio.NC cs.NE physics.bio-ph

Slow Transition to Low-Dimensional Chaos in Heavy-Tailed Recurrent Neural Networks

Yi Xie, Stefan Mihalas, Łukasz Kuśmierz

Journal ref Advances in Neural Information Processing Systems 38, 136401-136437 (2025)

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.19141 2026-08-11 eess.SP cs.LG 版本更新

EEG Foundation Challenge: From Cross-Task to Cross-Subject EEG Decoding

脑电(EEG)基础挑战赛:从跨任务到跨主体的脑电解码

Bruno Aristimunha, Dung Truong, Pierre Guetschel, Seyed Yahya Shirazi, Isabelle Guyon, Alexandre R. Franco, Michael P. Milham, Aviv Dotan, Scott Makeig, Alexandre Gramfort, Jean-Remi King, Marie-Constance Corsi, Pedro A. Valdés-Sosa, Amit Majumdar, Alan Evans, Terrence J Sejnowski, Oren Shriki, Sylvain Chevallier, Arnaud Delorme

AI总结 本文介绍一项包含跨任务跨主体脑电解码、精神病理因素预测两项挑战的大规模脑电竞赛,提供对应基线模型,旨在推动泛化性脑电解码模型及相关临床应用的发展。

Comments Approved at Neurips Competition track. webpage: https://eeg2025.github.io/

Journal ref The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS), 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2509.16664 2026-08-10 cs.LG cs.CV 交叉投稿

$\boldsymbolλ$-Orthogonality Regularization for Compatible Representation Learning

λ-正交性正则化用于兼容表示学习

Simone Ricci, Niccolò Biondi, Federico Pernici, Ioannis Patras, Alberto Del Bimbo

机构 * DINFO (Department of Information Engineering), University of Florence, Italy(意大利佛罗伦萨大学信息工程系) MICC (Media Integration and Communication Center)(媒体整合与通信中心) Queen Mary University of London, UK(英国伦敦女王学院)

AI总结 本文提出λ-正交性正则化方法,通过学习仿射变换在保持原有表示的同时实现分布特定的适应,验证了其在不同架构和数据集上的有效性,保持了零样本性能并确保模型更新的兼容性。

Comments Accepted at NeurIPS2025

Journal ref Advances in Neural Information Processing Systems 38 (NeurIPS 2025), pp. 29036-29063

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.11583 2026-08-10 stat.ML cs.LG

Distributional Autoencoders Know the Score

分布式自编码器知晓得分

Andrej Leban

机构 * Department of Statistics, University of Michigan(密歇根大学统计学系)

AI总结 本文提出分布式主元自编码器,通过理论保证实现分布正确重建与编码可解释性,证明模型能同时学习数据分布和内在维度。

Comments NeurIPS 2025 - camera-ready version

Journal ref Advances in Neural Information Processing Systems 38 (NeurIPS 2025), 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.01582 2026-08-10 cs.LG cond-mat.dis-nn cs.IT math.IT stat.ML

Bayes optimal learning of attention-indexed models

贝叶斯最优学习注意力索引模型

Fabrizio Boncoraglio, Emanuele Troiani, Vittorio Erba, Lenka Zdeborová

机构 * Statistical Physics of Computation Laboratory, École polytechnique fédérale de Lausanne (EPFL)(计算物理学实验室,瑞士联邦理工学院(EPFL))

AI总结 本文提出注意力索引模型,通过理论分析和算法设计,探讨深度注意力层的贝叶斯最优学习问题。

Journal ref NeurIPS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.17958 2026-08-10 stat.ML cond-mat.dis-nn cs.IT cs.LG math.IT

The Nuclear Route: Sharp Asymptotics of ERM in Overparameterized Quadratic Networks

核路:在过参数化二次网络中ERM的尖锐渐近性

Vittorio Erba, Emanuele Troiani, Lenka Zdeborová, Florent Krzakala

机构 * Statistical Physics of Computation Laboratory(计算物理学统计力学实验室) Information, Learning and Physics Laboratory(信息、学习与物理实验室)

AI总结 该研究通过将过参数化二次网络的ERM问题转化为凸矩阵感知任务,揭示了低秩结构对容量控制的影响,并确定了目标函数宽度对可学习性的作用。

Journal ref NeurIPS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.06769 2026-08-10 cs.CV cs.AI

Stitch and Tell: A Structured Multimodal Data Augmentation Method for Spatial Understanding

拼接与讲述:一种结构化多模态数据增强方法用于空间理解

Hang Yin, Xiaomin He, PeiWen Yuan, Yiwei Li, Jiayi Shi, Wenxiao Fan, Shaoxiong Feng, Kan Li

机构 * School of Computer Science, Beijing Institute of Technology(北京理工大学计算机科学学院) School of Software and Microelectronics, Peking University(北京大学软件与微电子学院) Xiaohongshu Inc(小红书公司)

AI总结 Stitch and Tell通过结构化空间监督提升视觉-语言模型的空间理解能力,有效缓解空间幻觉并提高相关任务性能。

Journal ref Advances in Neural Information Processing Systems 38 (NeurIPS 2025)

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.13961 2026-08-10 stat.ML cs.LG

The Computational Advantage of Depth: Learning High-Dimensional Hierarchical Functions with Gradient Descent

Yatin Dandi, Luca Pesce, Lenka Zdeborová, Florent Krzakala

机构 * Information, Learning and Physics Laboratory. Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland.(信息、学习与物理实验室。瑞士洛桑联邦理工学院(EPFL)) Statistical Physics of Computation Laboratory. Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland.(计算统计物理实验室。瑞士洛桑联邦理工学院(EPFL))

Journal ref NeurIPS 2025 (Spotlight)

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.09432 2026-08-10 cs.LG 版本更新

Equivariant Sparse Autoencoders: Mechanistic Interpretability of Neural Networks on Symmetric Data

等变稀疏自编码器:对称数据上神经网络的机制可解释性

Ege Erdogan, Ana Lucic

机构 * University of Amsterdam(阿姆斯特丹大学)

AI总结 该研究针对稀疏自编码器在对称数据上的不可识别问题,提出等变稀疏自编码器,可避免缺陷并发现更有用的下游任务特征,表明重构质量与特征实用性在对称下可能负相关。

Comments NeurIPS 2025 Mechanistic Interpretability and UniReps workshops

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.02651 2026-08-10 stat.ML cond-mat.dis-nn cs.LG

Asymptotics of SGD in Sequence-Single Index Models and Single-Layer Attention Networks

Luca Arnaboldi, Bruno Loureiro, Ludovic Stephan, Florent Krzakala, Lenka Zdeborova

Journal ref NeurIPS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.18046 2026-08-10 cs.LG cond-mat.dis-nn stat.ML

Learning with Restricted Boltzmann Machines: Asymptotics of AMP and GD in High Dimensions

Yizhou Xu, Florent Krzakala, Lenka Zdeborová

机构 * Statistical Physics of Computation Laboratory (SPOC), EPFL, Switzerland(计算统计物理实验室(SPOC),瑞士联邦理工学院) Information, Learning, and Physics Laboratory (IDEPHICS), EPFL, Switzerland(信息、学习与物理实验室(IDEPHICS),瑞士联邦理工学院)

Journal ref NeurIPS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2205.13503 2026-08-10 cs.IT math.IT

Multi-layer State Evolution Under Random Convolutional Design

Mara Daniels, Cédric Gerbelot, Florent Krzakala, Lenka Zdeborová

Comments Accepted to NeurIPS 2022

Journal ref Advances in Neural Information Processing Systems (2022), vol 52, pages 7089--7102

详情

展开后加载摘要…

URL PDF HTML 收藏
2408.03733 2026-08-10 stat.ML cond-mat.dis-nn cs.IT cs.LG math.IT math.PR

Bayes-optimal learning of an extensive-width neural network from quadratically many samples

Antoine Maillard, Emanuele Troiani, Simon Martin, Florent Krzakala, Lenka Zdeborová

Comments 47 pages

Journal ref Advances in Neural Information Processing Systems 37 (NeurIPS 2024)

详情

展开后加载摘要…

URL PDF HTML 收藏
2402.03902 2026-08-10 cs.LG

A phase transition between positional and semantic learning in a solvable model of dot-product attention

Hugo Cui, Freya Behrens, Florent Krzakala, Lenka Zdeborová

Journal ref Advances in Neural Information Processing Systems 37 (NeurIPS 2024)

详情

展开后加载摘要…

URL PDF HTML 收藏
2305.11041 2026-08-10 cs.LG cond-mat.dis-nn stat.ML

High-dimensional Asymptotics of Denoising Autoencoders

Hugo Cui, Lenka Zdeborová

Journal ref Advances in Neural Information Processing Systems 36 (2023)

详情

展开后加载摘要…

URL PDF HTML 收藏
2302.08933 2026-08-10 math.ST stat.ML stat.TH

Universality laws for Gaussian mixtures in generalized linear models

Yatin Dandi, Ludovic Stephan, Florent Krzakala, Bruno Loureiro, Lenka Zdeborová

Journal ref Advances in Neural Information Processing Systems 36 (2023)

详情

展开后加载摘要…

URL PDF HTML 收藏
2202.00293 2026-08-10 stat.ML cond-mat.dis-nn cs.LG

Phase diagram of Stochastic Gradient Descent in high-dimensional two-layer neural networks

Rodrigo Veiga, Ludovic Stephan, Bruno Loureiro, Florent Krzakala, Lenka Zdeborová

Comments 20 pages

Journal ref Advances in Neural Information Processing Systems (2022), vol 35, pages {23244--23255)

详情

展开后加载摘要…

URL PDF HTML 收藏
2205.13527 2026-08-10 stat.ML cond-mat.dis-nn cs.LG math.PR math.ST stat.TH

Subspace clustering in high-dimensions: Phase transitions & Statistical-to-Computational gap

Luca Pesce, Bruno Loureiro, Florent Krzakala, Lenka Zdeborová

Comments NeurIPS camera-ready version

Journal ref Advances in Neural Information Processing Systems (2022), vol 35, pages 27087--27099

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.05422 2026-08-07 cs.LG cs.MA 新提交

IFlowNets: Extending Generative Samplers to Learn Strategies in Incomplete Information Games

IFlowNets:将生成采样器扩展至不完美信息博弈中学习策略

Conor M. Artman, Nicholas Di, Scott Perkins

AI总结 该研究将AFlowNets扩展为适用于不完美信息博弈的IFlowNets,解决了原有约束无法得到有效密度与训练目标的问题,在标准博弈环境中性能与速度优于或相当于OSMCCFR等方法。

Comments Accepted at the NeurIPS 2025 Workshop on Dynamics at the Frontiers of Optimization, Sampling, and Games

详情

展开后加载摘要…

URL PDF HTML 收藏