IBGS: Image-Based Gaussian Splatting
机构 * Australian National University（澳大利亚国立大学） ； NVIDIA（英伟达）
Comments Accepted to NeurIPS 2025
期刊&会议
NeurIPS
Conference on Neural Information Processing Systems · 会议 · Machine Learning
INC: An Indirect Neural Corrector for Auto-Regressive Hybrid PDE Solvers
机构 * Technical University of Munich（慕尼黑技术大学）
Comments Accepted at NeurIPS 2025. 35 pages, 10 figures
Optimality and NP-Hardness of Transformers in Learning Markovian Dynamical Functions
机构 * Department of Computer Science, Rensselaer Polytechnic Institute（拉特格斯理工学院计算机科学系） ； Department of Computer Science and Engineering, The Chinese University of Hong Kong（香港中文大学计算机科学与工程系） ； Mathematics of Computation, IBM Research（IBM研究院数学部）
Comments NeurIPS 2025
Not All Deepfakes Are Created Equal: Triaging Audio Forgeries for Robust Deepfake Singer Identification
机构 * Universal Music Group（环球音乐集团）
Comments Accepted for presentation at the NeurIPS 2025 Workshop on Generative and Protective AI for Content Creation (non-archival)
A More Realistic Evaluation of Cross-Frequency Transfer Learning and Foundation Forecasting Models
机构 * Amazon（亚马逊）
Comments NeurIPS 2025 Workshop on Recent Advances in Time Series Foundation Models (BERT2S)
Explaining Similarity in Vision-Language Encoders with Weighted Banzhaf Interactions
机构 * University of Warsaw（华沙大学） ； Warsaw University of Technology（华沙理工大学） ； LMU Munich（慕尼黑大学） ； MCML ； DFKI（德累斯顿大学） ； Bielefeld University（比勒菲尔德大学）
Comments NeurIPS 2025. Code: https://github.com/hbaniecki/fixlip
Bridging the Gap: Examining Vision Foundation Models for Optical and Radio Astronomy Applications
Comments 12 pages, 5 figures, submitted to Astronomy and Astrophysics. A previous version of this work was accepted to the Foundation Models for Science Workshop at NeurIPS 2024
Journal ref A&A 703, A217 (2025)
NeuroPath: Neurobiology-Inspired Path Tracking and Reflection for Semantically Coherent Retrieval
机构 * Institute of Intelligent Computing, University of Electronic Science and Technology of China（电子科技大学智能计算研究院）
Comments Accepted by NeurIPS 2025
Learning Skill-Attributes for Transferable Assessment in Video
机构 * University of Texas at Austin（德克萨斯大学奥斯汀分校）
Comments NeurIPS 2025, Project webpage: https://vision.cs.utexas.edu/projects/CrossTrainer/
CTRL-ALT-DECEIT: Sabotage Evaluations for Automated AI R&D
机构 * LawZero ； Apollo Research ； Imperial College London（帝国理工学院伦敦校区）
Comments 53 pages, 21 figures, 8 tables. Accepted as a spotlight at NeurIPS 2025
Discovering EV Charging Site Archetypes Through Few Shot Forecasting: The First U.S.-Wide Study
机构 * Alpha Grid ； University of Nebraska-Lincoln（内布拉斯加大学林肯分校）
Comments Tackling Climate Change with Machine Learning: Workshop at NeurIPS 2025
Posterior Sampling by Combining Diffusion Models with Annealed Langevin Dynamics
机构 * UT Austin（得克萨斯大学） ； Microsoft Research（微软研究院）
Comments NeurIPS 2025
Retrosynthesis Planning via Worst-path Policy Optimisation in Tree-structured MDPs
机构 * University of Warwick（沃里克大学） ； The Alan Turing Institute（艾伦·图灵研究所）
Comments Published as a conference paper at NeurIPS 2025
EvoLM: In Search of Lost Language Model Training Dynamics
机构 * Harvard（哈佛大学） ； Stanford（斯坦福大学） ； EPFL（苏黎世联邦理工学院） ； CMU（卡内基梅隆大学）
Comments NeurIPS 2025 (Oral)
Vision Transformers with Self-Distilled Registers
Comments NeurIPS 2025 Spotlight. Website: https://github.com/0raiser0/PH-Reg
Preference Learning with Lie Detectors can Induce Honesty or Evasion
Comments NeurIPS 2025
Availability-aware Sensor Fusion via Unified Canonical Space
机构 * CCS Graduate School of Mobility（移动系统联合研究生院） ； KAIST（韩国科学技术院）
Comments Accepted at NeurIPS 2025
Journal ref Proceedings of the Neural Information Processing Systems (NeurIPS 2025)
PRIMUS: Pretraining IMU Encoders with Multimodal Self-Supervision
机构 * Nokia Bell Labs Cambridge, UK（诺基亚贝尔实验室（剑桥，英国）） ； University of Washington, USA（华盛顿大学（美国）） ； University of Glasgow, UK（格拉斯哥大学（英国））
Comments Presented at ICASSP 2025. Also presented under the title "PRIMUS: Pretraining IMU Encoders with Multimodal and Self-Supervised Learning" at NeurIPS 2024 TSALM Workshop (Time Series in the Age of Large Models)
Autoregressive Image Diffusion: Generation of Image Sequence and Application in MRI
机构 * University Medical Center Göttingen（哥廷根大学医学中心） ； Shenzhen Technology University（深圳科技大学） ； Graz University of Technology（格拉茨技术大学）
Journal ref Advances in Neural Information Processing Systems 2024;37:129094-129119
The Last Vote: A Multi-Stakeholder Framework for Language Model Governance
Comments This paper has been accepted to the NeurIPS 2025 Workshop on Algorithmic Collective Action (ACA@NeurIPS 2025). The submission is 26 pages including the appendix and includes the NeurIPS checklist. A big thanks to Avijit Ghosh
Natural gradient descent for improving variational inference based classification of radio galaxies
Comments Accepted to Machine Learning and the Physical Sciences Workshop, NeurIPS 2025
TokenSqueeze: Performance-Preserving Compression for Reasoning LLMs
机构 * State Key Lab of CAD&CG, Zhejiang University（浙江大学CAD&CG国家重点实验室） ； School of Software Technology, Zhejiang University（浙江大学软件学院） ； Alibaba Cloud（阿里云） ； Zhiyuan Research Institute（智源研究院）
Comments Accepted to NeurIPS 2025
Toward Explainable Offline RL: Analyzing Representations in Intrinsically Motivated Decision Transformers
机构 * DISPOC University of Siena（DISPOC 硅恩纳大学） ； University of Siena（硅恩纳大学） ； University of Milan（米兰大学）
Comments Accepted for poster presentation at the NeurIPS 2025 workshop "CogInterp: Interpreting Cognition in Deep Learning Models", San Diego, CA, USA
Compress, Gather, and Recompute: REFORMing Long-Context Processing in Transformers
机构 * KAIST（韩国科学技术院） ； Amazon AGI（亚马逊人工智能实验室）
Comments NeurIPS 2025
ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS
机构 * ZIP Lab, Zhejiang University（浙江大学ZIP实验室） ； Monash University（墨尔本大学）
Comments NeurIPS 2025, Project Page: https://lhmd.top/zpressor, Code: https://github.com/ziplab/ZPressor
Meta-Learning an In-Context Transformer Model of Human Higher Visual Cortex
Comments Accepted to NeurIPS 2025. Website: https://github.com/leomqyu/BraInCoRL
Hogwild! Inference: Parallel LLM Generation via Concurrent Attention
机构 * Yandex ； HSE University（俄罗斯高等经济大学） ； IST Austria（国际研究所）
Comments 39th Conference on Neural Information Processing Systems (NeurIPS 2025)
A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders
机构 * LASR Labs（LASR实验室） ； University College London（伦敦大学学院） ； Tübingen AI Center, University of Tübingen（图宾根大学人工智能中心） ； MATS ； Decode Research（Decode研究）
Comments Accepted at NeurIPS 2025 (Oral)
Learning Time-Scale Invariant Population-Level Neural Representations
机构 * Computing & Mathematical Sciences（计算与数学科学） ； Computation & Neural Systems（计算与神经系统） ； California Institute of Technology（加州理工学院）
Comments 10 pages, 5 figures, NeurIPS 2025 Foundation Models for the Brain and Body
The Good, The Bad, and The Hybrid: A Reward Structure Showdown in Reasoning Models Training
机构 * Berkeley AI Safety Initiative (BASIS) UC Berkeley（伯克利人工智能安全计划（BASIS）加州大学伯克利分校）
Comments Paper accepted to the 2nd Workshop on Aligning Reinforcement Learning Experimentalists and Theorists (ARLET 2025) at NeurIPS; the paper consists of 14 pages (including the appendix) and contains 3 figures
