NeurIPS

Perturb a Model, Not an Image: Towards Robust Privacy Protection via Anti-Personalized Diffusion Models

Tae-Young Lee, Juwon Seo, Jong Hwan Ko, Gyeong-Moon Park

机构 * Korea University（韩国大学） ； Kyung Hee University（庆熙大学） ； Sungkyunkwan University（成均馆大学） 纠错

Comments 26 pages, 9 figures, 16 tables, NeurIPS 2025

Detecting Generated Images by Fitting Natural Image Distributions

Yonggang Zhang, Jun Nie, Xinmei Tian, Mingming Gong, Kun Zhang, Bo Han

机构 * The Hong Kong University of Science and Technology（香港科学与技术大学） ； TMLR Group, Hong Kong Baptist University（香港 Baptist 大学 TMLR 组） ； University of Science and Technology of China（中国科学技术大学） ； The University of Melbourne, Australia（澳大利亚墨尔本大学） ； Carnegie Mellon University（卡内基梅隆大学） ； Mohamed bin Zayed University of Artificial Intelligence（Mohamed bin Zayed 人工智能大学） 纠错

Comments 25 pages, 9 figures, NeurIPS 2025 spotlight

FirstAidQA: A Synthetic Dataset for First Aid and Emergency Response in Low-Connectivity Settings

Saiyma Sittul Muna, Rezwan Islam Salvi, Mushfiqur Rahman Mushfique, Ajwad Abrar

机构 * Islamic University of Technology（伊斯兰技术大学） 纠错

Comments Accepted at the 5th Muslims in Machine Learning (MusIML) Workshop, co-located with NeurIPS 2025

Evolve to Inspire: Novelty Search for Diverse Image Generation

Alex Inch, Passawis Chaiyapattanaporn, Yuchen Zhu, Yuan Lu, Ting-Wen Ko, Davide Paglieri

机构 * University College London（伦敦大学学院） ； University of Oxford（牛津大学） ； Tesco Technology（百事科技） ； Microsoft Research（微软研究院） 纠错

Comments 14 pages, 10 figures, Accepted to Neurips 2025 GenProCC Workshop

Position Paper: If Innovation in AI Systematically Violates Fundamental Rights, Is It Innovation at All?

Josu Eguiluz Castañeira, Axel Brando, Migle Laukyte, Marc Serra-Vidal

机构 * Adevinta ServicesCo S.L.U. ； Pompeu Fabra University (UPF)（庞培法布拉大学） ； Barcelona Supercomputing Center (BSC-CNS)（巴塞罗那超级计算中心） ； Kleinanzeigen.de GmbH (Adevinta) 纠错

Comments NeurIPS 2025 Position Paper track; accepted for oral and poster presentation at the Thirty-Ninth Annual Conference on Neural Information Processing Systems

Towards Transparent Reasoning: What Drives Faithfulness in Large Language Models?

Teague McMillan, Gabriele Dominici, Martin Gjoreski, Marc Langheinrich

机构 * ETH Zurich（苏黎世联邦理工学院） ； Università della Svizzera italiana（瑞士意大利大学） 纠错

Comments 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Workshop: NeurIPS 2025 Workshop on Evaluating the Evolving LLM Lifecycle: Benchmarks, Emergent Abilities, and Scaling

Contextual Tokenization for Graph Inverted Indices

Pritish Chakraborty, Indradyumna Roy, Soumen Chakrabarti, Abir De

机构 * Department of Computer Science and Engineering, IIT Bombay（印度理工学院班加罗尔计算机科学与工程系） 纠错

Journal ref NeurIPS 2025 paper

Bridging Symmetry and Robustness: On the Role of Equivariance in Enhancing Adversarial Robustness

Longwei Wang, Ifrat Ikhtear Uddin, KC Santosh, Chaowei Zhang, Xiao Qin, Yang Zhou

机构 * AI Research Lab, Department of Computer Science, University of South Dakota（人工智能研究实验室、计算机科学系、南达科他大学） ； School of Information and Artificial Intelligence, Yangzhou University（信息与人工智能学院、扬州大学） ； Department of Computer Science and Software Engineering, Auburn University（计算机科学与软件工程系、阿伯丁大学） 纠错

Comments Accepted for the proceedings of 39th Conference on Neural Information Processing Systems (NeurIPS 2025)

Sound Clouds: Exploring ambient intelligence in public spaces to elicit deep human experience of awe, wonder, and beauty

Chengzhi Zhang, Dashiel Carrera, Daksh Kapoor, Jasmine Kaur, Jisu Kim, Brian Magerko

机构 * Georgia Institute of Technology（佐治亚理工学院） ； University of Toronto（多伦多大学） 纠错

Comments 4 pages, Artwork accepted by NeurIPS Creative AI Track 2025

Riemannian Consistency Model

Chaoran Cheng, Yusong Wang, Yuxin Chen, Xiangxin Zhou, Nanning Zheng, Ge Liu

机构 * University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校） ； Xi’an Jiaotong University（西安交通大学） ； University of Chinese Academy of Sciences（中国科学院大学） 纠错

Comments Accepted to NeurIPS 2025

SynBrain: Enhancing Visual-to-fMRI Synthesis via Probabilistic Representation Learning

Weijian Mai, Jiamin Wu, Yu Zhu, Zhouheng Yao, Dongzhan Zhou, Andrew F. Luo, Qihao Zheng, Wanli Ouyang, Chunfeng Song

机构 * Shanghai Artificial Intelligence Laboratory（上海人工智能实验室） ； The University of Hong Kong（香港大学） ； The Chinese University of Hong Kong（香港中文大学） 纠错

Comments Accepted by NeurIPS 2025

CosmoBench: A Multiscale, Multiview, Multitask Cosmology Benchmark for Geometric Deep Learning

Ningyuan Huang, Richard Stiskalek, Jun-Young Lee, Adrian E. Bayer, Charles C. Margossian, Christian Kragh Jespersen, Lucia A. Perez, Lawrence K. Saul, Francisco Villaescusa-Navarro

机构 * Centers for Computational Mathematics（计算数学中心） ； Computational Astrophysics（计算天文学） ； University of Oxford（牛津大学） ； Seoul National University（首尔国立大学） ； Princeton University（普林斯顿大学） 纠错

Comments Accepted at NeurIPS 2025

Representation Consistency for Accurate and Coherent LLM Answer Aggregation

Junqi Jiang, Tom Bewley, Salim I. Amoukou, Francesco Leofante, Antonio Rago, Saumitra Mishra, Francesca Toni

机构 * Imperial College London（帝国理工学院伦敦分校） ； J.P. Morgan AI Research（摩根大通人工智能研究） ； King’s College London（国王学院伦敦分校） 纠错

Comments Accepted at NeurIPS 2025. Camera-ready version

TabArena: A Living Benchmark for Machine Learning on Tabular Data

Nick Erickson, Lennart Purucker, Andrej Tschalzev, David Holzmüller, Prateek Mutalik Desai, David Salinas, Frank Hutter

机构 * Amazon Web Services（亚马逊网络服务） ； University of Freiburg（弗赖堡大学） ； University of Mannheim（曼海姆大学） ； INRIA Paris（巴黎国家信息与自动化研究所） ； Ecole Normale Supérieure（高等师范学校） ； PSL Research University（巴黎科学实验室大学） ； Prior Labs（先驱实验室） ； ELLIS Institute Tübingen（图宾根ELLIS研究所） 纠错

Comments Accepted (spotlight) at NeurIPS 2025 Datasets and Benchmarks Track. v4: fixed links in comments. v3: NeurIPS camera-ready version. v2: fixed author list. 51 pages. Code available at https://tabarena.ai/code and examples at https://tabarena.ai/code-examples and dataset curation at https://tabarena.ai/data-tabular-ml-iid-study and https://tabarena.ai/dataset-curation

Instance-Specific Test-Time Training for Speech Editing in the Wild

Taewoo Kim, Uijong Lee, Hayoung Park, Choongsang Cho, Nam In Park, Young Han Lee

Comments Accepted to NeurIPS 2025 Workshop on GenProCC

WildCAT3D: Appearance-Aware Multi-View Diffusion in the Wild

Morris Alper, David Novotny, Filippos Kokkinos, Hadar Averbuch-Elor, Tom Monnier

机构 * Tel Aviv University（特拉维夫大学） ； Meta AI ； Cornell University（康奈尔大学） 纠错

Comments Accepted to NeurIPS 2025. Project page: https://wildcat3d.github.io

ConTextTab: A Semantics-Aware Tabular In-Context Learner

Marco Spinaci, Marek Polewczyk, Maximilian Schambach, Sam Thelin

机构 * SAP France（SAP法国分公司） ； SAP SE（SAP德国分公司） 纠错

Comments Accepted as spotlight at NeurIPS 2025

Deep Video Discovery: Agentic Search with Tool Use for Long-form Video Understanding

Xiaoyi Zhang, Zhaoyang Jia, Zongyu Guo, Jiahao Li, Bin Li, Houqiang Li, Yan Lu

机构 * Microsoft Research Asia（微软亚洲研究院） ； University of Science and Technology of China（中国科学技术大学） 纠错

Comments Accepted to NeurIPS 2025

EgoBlind: Towards Egocentric Visual Assistance for the Blind

Junbin Xiao, Nanxin Huang, Hao Qiu, Zhulin Tao, Xun Yang, Richang Hong, Meng Wang, Angela Yao

机构 * National University of Singapore（新加坡国立大学） ； Communication University of China（中国传媒大学） ； University of Science and Technology of China（中国科学技术大学） ； Hefei University of Technoloy（合肥工业大学） 纠错

Comments NeurIPS'25 (D&B Track)

XIFBench: Evaluating Large Language Models on Multilingual Instruction Following

Zhenyu Li, Kehai Chen, Yunfei Long, Xuefeng Bai, Yaoyin Zhang, Xuchen Wei, Juntao Li, Min Zhang

机构 * Harbin Institute of Technology, Shenzhen（哈尔滨工业大学（深圳）） ； Queen Mary University of London（伦敦大学女王学院） ； Soochow University（苏州大学） 纠错

Comments Accepted by the NeurIPS 2025 Datasets and Benchmarks Track

Deep Modularity Networks with Diversity-Preserving Regularization

Yasmin Salehi, Dennis Giannacopoulos

机构 * Department of Electrical and Computering Engineering（电气与计算机工程系） ； McGill University（麦吉尔大学） 纠错

Comments Accepted at the 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Workshop: New Perspectives in Graph Machine Learning (NPGML)

FlexEvent: Towards Flexible Event-Frame Object Detection at Varying Operational Frequencies

Dongyue Lu, Lingdong Kong, Gim Hee Lee, Camille Simon Chane, Wei Tsang Ooi

机构 * National University of Singapore（新加坡国立大学） ； IPAL, CNRS IRL（IPAL与CNRS IRL） ； ETIS UMR 8051, CY Cergy Paris University, ENSEA, CNRS, France（ETIS UMR 8051，CY巴黎中央理工大学，ENSEA，CNRS，法国） 纠错

Comments NeurIPS 2025; 28 pages, 14 figures, 10 tables; Code at https://flexevent.github.io/

Epistemic Uncertainty for Generated Image Detection

Jun Nie, Yonggang Zhang, Tongliang Liu, Yiu-ming Cheung, Bo Han, Xinmei Tian

机构 * MoE Key Laboratory of Brain-inspired Intelligent Perception and Cognition, University of Science and Technology of China（脑启发智能感知与认知联合实验室，中国科学技术大学） ； Hong Kong Baptist University（香港 Baptist 大学） ； Sydney AI Centre, The University of Sydney（悉尼人工智能中心，悉尼大学） ； The Hong Kong University of Science and Technology（香港科学与技术大学） 纠错

Comments 28 pages, 10 figures, NeurIPS 2025

Double Descent Meets Out-of-Distribution Detection: Theoretical Insights and Empirical Analysis on the role of model complexity

Mouïn Ben Ammar, David Brellmann, Arturo Mendoza, Antoine Manzanera, Gianni Franchi

机构 * U2IS Lab ENSTA Paris（ENSTA巴黎大学U2IS实验室） ； Safran Tech（萨弗兰技术） 纠错

Comments Accepted at NeurIPS 2025 (Conference on Neural Information Processing Systems)

Can Large Language Models Analyze Graphs like Professionals? A Benchmark, Datasets and Models

Xin Li, Weize Chen, Qizhi Chu, Haopeng Li, Zhaojun Sun, Ran Li, Chen Qian, Yiwei Wei, Zhiyuan Liu, Chuan Shi, Maosong Sun, Cheng Yang

机构 * School of Computer Science, Beijing University of Posts and Telecommunications（北京邮电大学计算机科学学院） ； Department of Computer Science and Technology, Tsinghua University（清华大学计算机科学与技术系） ； School of Artificial Intelligence, Beijing University of Posts and Telecommunications（北京邮电大学人工智能学院） ； College of Petroleum Engineering, China University of Petroleum (Beijing) at Karamay（中国石油大学（北京）克拉玛依校区石油工程学院） ； Institute for Artificial Intelligence, Tsinghua University（清华大学人工智能研究院） 纠错

Comments NeurIPS 2024

Neural Entropy

Akhil Premkumar

Comments 29 pages + references, 18 figures. Camera-ready version from NeurIPS 2025

Memory-Enhanced Neural Solvers for Routing Problems

Felix Chalumeau, Refiloe Shabe, Noah De Nicola, Arnu Pretorius, Thomas D. Barrett, Nathan Grinsztajn

机构 * InstaDeep ； University of Cape Town（南非开普敦大学） 纠错

Comments Neurips '25 version

Web-Scale Collection of Video Data for 4D Animal Reconstruction

Brian Nlong Zhao, Jiajun Wu, Shangzhe Wu

机构 * Stanford University（斯坦福大学） ； University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校） ； University of Cambridge（剑桥大学） 纠错

Comments NeurIPS 2025 Datasets and Benchmarks

AI Progress Should Be Measured by Capability-Per-Resource, Not Scale Alone: A Framework for Gradient-Guided Resource Allocation in LLMs

David McCoy, Yulun Wu, Zachary Butzin-Dozier

机构 * Division of Biostatistics University of California, Berkeley（生物统计学系加州大学伯克利分校） ； Capital One（Capital One公司） 纠错

Comments 9 pages (main) + appendix, 3 figures. Accepted at NeurIPS 2025 (Position Paper Track), submission #491. OpenReview: https://openreview.net/forum?id=6plSmhBI33&noteId=KP5ZqY7JLg

Dynamic Diameter in High-Dimensions against Adaptive Adversary and Beyond

Kiarash Banihashem, Jeff Giliberti, Samira Goudarzi, MohammadTaghi Hajiaghayi, Peyman Jabbarzade, Morteza Monemizadeh

Comments NeurIPS'25