NeurIPS

Enhancing the Outcome Reward-based RL Training of MLLMs with Self-Consistency Sampling

Jiahao Wang, Weiye Xu, Aijun Yang, Wengang Zhou, Lewei Lu, Houqiang Li, Xiaohua Wang, Jinguo Zhu

机构 * Xi’an Jiaotong University（西安交通大学） ； University of Science and Technology of China（中国科学技术大学） ； SenseTime Research（商汤科技研究院） 纠错

Comments Accepted to NeurIPS 2025 (The Thirty-Ninth Annual Conference on Neural Information Processing Systems)

Impact of Layer Norm on Memorization and Generalization in Transformers

Rishi Singhal, Jung-Eun Kim

Comments NeurIPS 2025

GraphFaaS: Serverless GNN Inference for Burst-Resilient, Real-Time Intrusion Detection

Lingzhi Wang, Vinod Yegneswaran, Xinyi Shi, Ziyu Li, Ashish Gehani, Yan Chen

Comments Accepted by ML For Systems workshop at Neural Information Processing Systems (NeurIPS 2025)

Improving Perturbation-based Explanations by Understanding the Role of Uncertainty Calibration

Thomas Decker, Volker Tresp, Florian Buettner

机构 * Siemens AG（西门子股份公司） ； LMU Munich（慕尼黑大学） ； Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心） ； Goethe University Frankfurt（法兰克福歌德大学） ； German Cancer Research Center (DKFZ)（德国癌症研究中心） ； German Cancer Consortium (DKTK)（德国癌症联合会） 纠错

Comments 39th Conference on Neural Information Processing Systems (NeurIPS 2025)

Product distribution learning with imperfect advice

Arnab Bhattacharyya, Davin Choo, Philips George John, Themis Gouleakis

机构 * Department of Computer Science University of Warwick（沃里克大学计算机科学系） ； Harvard University（哈佛大学） ； College of Computing & Data Science Nanyang Technological University（南洋理工大学计算机科学与数据科学学院） 纠错

Comments Full version (11 pages). To be published in NeurIPS 2025

Torch-Uncertainty: A Deep Learning Framework for Uncertainty Quantification

Adrien Lafage, Olivier Laurent, Firas Gabetni, Gianni Franchi

机构 * U2IS, ENSTA, Institut Polytechnique de Paris（U2IS、ENSTA、巴黎理工学院） 纠错

Comments NeurIPS 2025 Spotlight

Towards Leveraging Sequential Structure in Animal Vocalizations

Eklavya Sarkar, Mathew Magimai. -Doss

机构 * Idiap Research Institute（IDiap研究机构） ； Ecole Polytechnique Fédérale de Lausanne（洛桑联邦理工学院） 纠错

Comments Accepted at NeurIPS workshop (AI for Non-Human Animal Communication)

RobIA: Robust Instance-aware Continual Test-time Adaptation for Deep Stereo

Jueun Ko, Hyewon Park, Hyesong Choi, Dongbo Min

机构 * Ewha Womans University（成均馆大学） ； Soongsil University（顺天大学） 纠错

Comments Accepted by Neural Information Processing Systems (NeurIPS) 2025

Opinion: Towards Unified Expressive Policy Optimization for Robust Robot Learning

Haidong Huang, Haiyue Zhu. Jiayu Song, Xixin Zhao, Yaohua Zhou, Jiayi Zhang, Yuze Zhai, Xiaocong Li

机构 * Eastern Institute of Technology（东部技术研究所） ； University of Nottingham（诺丁汉大学） ； SIMTech, Agency for Science, Technology and Research (A*STAR)（SIMTech，科技研究局（A*STAR）） ； Southern University of Science and Technology（南方科技大学） 纠错

Comments Accepted by NeurIPS 2025 Workshop on Embodied World Models for Decision Making

VLF-MSC: Vision-Language Feature-Based Multimodal Semantic Communication System

Gwangyeon Ahn, Jiwan Seo, Joonhyuk Kang

Comments To appear in the AI4NextG Workshop at NeurIPS 2025

From Static Structures to Ensembles: Studying and Harnessing Protein Structure Tokenization

Zijing Liu, Bin Feng, He Cao, Yu Li

机构 * International Digital Economy Academy Shenzhen, China（国际数字经济学院深圳） 纠错

Comments NeurIPS 2025 AI for Science Workshop

AffordBot: 3D Fine-grained Embodied Reasoning via Multimodal Large Language Models

Xinyi Wang, Xun Yang, Yanlong Xu, Yuchen Wu, Zhen Li, Na Zhao

机构 * University of Science and Technology of China（科学技术大学） ； Singapore University of Technology and Design（新加坡科技设计大学） ； Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）） 纠错

Comments NeurIPS 2025

MVU-Eval: Towards Multi-Video Understanding Evaluation for Multimodal LLMs

Tianhao Peng, Haochen Wang, Yuanxing Zhang, Zekun Wang, Zili Wang, Gavin Chang, Jian Yang, Shihao Li, Yanghai Wang, Xintao Wang, Houyi Li, Wei Ji, Pengfei Wan, Steven Huang, Zhaoxiang Zhang, Jiaheng Liu

机构 * Nanjing University（南京大学） ； CASIA（中国科学院自动化研究所） ； Kuaishou Technology（快手科技） ； M-A-P 纠错

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

Quantifying Climate Policy Action and Its Links to Development Outcomes: A Cross-National Data-Driven Analysis

Aditi Dutta

机构 * School of Government University of Birmingham（政府学院伯明翰大学） 纠错

Comments This paper/proposal has been accepted as a poster in the NeurIPS 2025

Retrieval-Augmented Generation for Reliable Interpretation of Radio Regulations

Zakaria El Kassimi, Fares Fourati, Mohamed-Slim Alouini

机构 * KAUST（卡斯泰尔大学） 纠错

Comments 12 pages, 7 figures, AI4NextG @ NeurIPS 2025

Cameras as Relative Positional Encoding

Ruilong Li, Brent Yi, Junchen Liu, Hang Gao, Yi Ma, Angjoo Kanazawa

机构 * UC Berkeley（伯克利大学） ； NVIDIA（NVIDIA公司） ； HKU（香港大学） 纠错

Comments Project Page: https://www.liruilong.cn/prope/

Journal ref NeurIPS 2025

Generalized Linear Mode Connectivity for Transformers

Alexander Theus, Alessandro Cabodi, Sotiris Anagnostidis, Antonio Orvieto, Sidak Pal Singh, Valentina Boeva

Comments Accepted to NeurIPS 2025 (Oral)

Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning

Shenzhi Wang, Le Yu, Chang Gao, Chujie Zheng, Shixuan Liu, Rui Lu, Kai Dang, Xionghui Chen, Jianxin Yang, Zhenru Zhang, Yuqiong Liu, An Yang, Andrew Zhao, Yang Yue, Shiji Song, Bowen Yu, Gao Huang, Junyang Lin

机构 * Qwen Team, Alibaba Inc.（通义团队，阿里巴巴公司） ； LeapLab, Tsinghua University（清华大学跃实验室） 纠错

Comments Accepted to NeurIPS 2025. 25 pages, 17 figures, 2 tables

R1-Compress: Long Chain-of-Thought Compression via Chunk Compression and Search

Yibo Wang, Haotian Luo, Huanjin Yao, Tiansheng Huang, Haiying He, Rui Liu, Naiqiang Tan, Jiaxing Huang, Xiaochun Cao, Dacheng Tao, Li Shen

Comments Accepted by NeurIPS FoRLM Workshop

Constructing an Optimal Behavior Basis for the Option Keyboard

Lucas N. Alegre, Ana L. C. Bazzan, André Barreto, Bruno C. da Silva

机构 * Institute of Informatics Federal University of Rio Grande do Sul（里约格朗德杜斯尔市联邦大学信息学院） ； Federal University of Rio Grande do Sul（里约格朗德杜斯尔市联邦大学） ； Google DeepMind（谷歌DeepMind） ； University of Massachusetts（马萨诸塞大学） 纠错

Comments To appear in the Proceedings of the Thirty-ninth Conference on Neural Information Processing Systems (NeurIPS), 2025

Matryoshka Pilot: Learning to Drive Black-Box LLMs with LLMs

Changhao Li, Yuchen Zhuang, Rushi Qiang, Haotian Sun, Hanjun Dai, Chao Zhang, Bo Dai

机构 * Georgia Institute of Technology（佐治亚理工学院） ； Precur AI ； Google DeepMind（谷歌DeepMind） 纠错

Comments Accepted by NeurIPS 2025

Federated Behavioural Planes: Explaining the Evolution of Client Behaviour in Federated Learning

Dario Fenoglio, Gabriele Dominici, Pietro Barbiero, Alberto Tonda, Martin Gjoreski, Marc Langheinrich

机构 * Università della Svizzera italiana（瑞士大学） ； INRAE Paris（巴黎国家农业研究院） 纠错

Comments [v3] Pre-print of the paper accepted to NeurIPS 2024 (30 pages)

Distribution Learning Meets Graph Structure Sampling

Arnab Bhattacharyya, Sutanu Gayen, Philips George John, Sayantan Sen, N. V. Vinodchandran

机构 * University of Warwick（沃里克大学） ； IIT Kanpur（印度理工学院坎普尔分校） ； National University of Singapore（新加坡国立大学） ； Centre for Quantum Technologies（量子技术中心） ； University of Nebraska-Lincoln（内布拉斯加大学林肯分校） 纠错

Comments Full version (50 pages, 2 figures). Shortened abstract as per arXiv criteria. To be published in NeurIPS 2025

Neural Correlates of Language Models Are Specific to Human Language

Iñigo Parra

机构 * Department of Linguistics University of California, Berkeley（语言学系加州大学伯克利分校） 纠错

Comments To be presented at NeurIPS 2025 Workshops

Automated algorithm design via Nevanlinna-Pick interpolation

Ibrahim K. Ozaslan, Tryphon T. Georgiou, Mihailo R. Jovanovic

Comments Accepted to DynaFront Workshop at NeurIPS 2025

Succeed or Learn Slowly: Sample Efficient Off-Policy Reinforcement Learning for Mobile App Control

Georgios Papoudakis, Thomas Coste, Jianye Hao, Jun Wang, Kun Shao

机构 * Huawei Noah’s Ark Lab（华为诺亚实验室） ； University College London（伦敦大学学院） 纠错

Comments NeurIPS 2025

DuoGPT: Training-free Dual Sparsity through Activation-aware Pruning in LLMs

Ruokai Yin, Yuhang Li, Donghyun Lee, Priyadarshini Panda

机构 * Yale University（耶鲁大学） 纠错

Comments NeurIPS 2025. The code is available on Github (see hyperlink in the paper)

Caption This, Reason That: VLMs Caught in the Middle

Zihan Weng, Lucas Gomez, Taylor Whittington Webb, Pouya Bashivan

机构 * Integrated Program in Neuroscience (IPN) McGill University（神经科学联合计划 麦吉尔大学） ； Mila, University of Montreal（蒙特利尔大学Mila） ； Microsoft Research USA（微软研究院美国总部） ； Department of Physiology McGill University（生理学系 麦吉尔大学） 纠错

Comments Paper accepted by nips 2025

Regression Trees Know Calculus

Nathan Wycoff

机构 * Department of Mathematics and Statistics（数学与统计学系） ； University of Massachusetts Amherst（马萨诸塞大学阿默斯特分校） 纠错

Comments Camera ready version for Neurips 2025