Local Steps Speed Up Local GD for Heterogeneous Distributed Logistic Regression
机构 * George Mason University（乔治·玛森大学） ； George Washington University（乔治·华盛顿大学）
Comments ICLR 2025
期刊&会议
ICLR
International Conference on Learning Representations · 会议 · Machine Learning
Multi-Robot Motion Planning with Diffusion Models
机构 * Carnegie Mellon University（卡内基梅隆大学） ； Brown University（布朗大学）
Comments The first three authors contributed equally to this work. Published at ICLR 2025
Spectral-Refiner: Accurate Fine-Tuning of Spatiotemporal Fourier Neural Operator for Turbulent Flows
机构 * School of Science and Engineering University of Missouri-Kansas City（科学与工程学院，密苏里大学-堪萨斯城分校） ； Department of Mathematics Emory University（数学系，埃默里大学） ； Center for Applied Scientific Computing Lawrence Livermore National Laboratory（应用科学计算中心，劳伦斯利弗莫尔国家实验室）
Comments Published as a conference paper at ICLR 2025
The Steganographic Potentials of Language Models
Comments Published at Building Trust Workshop at ICLR 2025
HadamRNN: Binary and Sparse Ternary Orthogonal RNNs
Journal ref International Conference on Learning Representations (ICLR), Apr 2025, Singapour, Singapore
Patterns and Mechanisms of Contrastive Activation Engineering
机构 * Georgia Institute of Technology（佐治亚理工学院）
Comments Published at the ICLR 2025 Bi-Align, HAIC, and Building Trust workshops
VLM Q-Learning: Aligning Vision-Language Models for Interactive Decision-Making
机构 * The University of Texas at Austin（德克萨斯大学奥斯汀分校） ； Salesforce AI Research（Salesforce人工智能研究）
Comments SSI-FM Workshop ICLR 2025
Null Counterfactual Factor Interactions for Goal-Conditioned Reinforcement Learning
机构 * The University of Texas at Austin（德克萨斯大学奥斯汀分校） ； University of California San Diego（加州大学圣地亚哥分校） ； MBZUAI（马克斯·普朗克人工智能研究所） ； University of Massachusetts Amherst（马萨诸塞大学阿姆赫斯特分校）
Comments Published at ICLR 2025
Journal ref The Thirteenth International Conference on Learning Representations. 2025
Beyond Single Concept Vector: Modeling Concept Subspace in LLMs with Gaussian Distribution
机构 * New Jersey Institute of Technology（新泽西理工学院） ； Wake Forest University（威克森林大学） ； Cisco Research（思科研究）
Comments Accepted by ICLR 2025
Language Models Trained to do Arithmetic Predict Human Risky and Intertemporal Choice
机构 * Department of Computer Science（计算机科学系） ； Princeton University（普林斯顿大学） ； Department of Psychology（心理学系） ； University of Warwick（沃里克大学）
Journal ref ICLR 2025
FreDF: Learning to Forecast in the Frequency Domain
机构 * Department of Control Science and Engineering, Zhejiang University（浙江大学控制科学与工程学院） ； School of Automation, Central South University（中南大学自动化学院） ； Department of Computer Science and Engineering, Shanghai Jiao Tong University（上海交通大学计算机科学与工程学院） ； Trust and Safety Team, TikTok Sydney, ByteDance Inc.（字节跳动悉尼信任与安全团队） ； Center for Data Science, Peking University（北京大学数据科学中心） ； Generative AI Lab, College of Computing and Data Science, Nanyang Technological University（南洋理工大学计算与数据科学学院生成式人工智能实验室）
Comments Accepted by ICLR 2025
Exploring Design Choices for Autoregressive Deep Learning Climate Models
机构 * Data Science Chair, Center for Artificial Intelligence and Data Science (CAIDAS), University of Würzburg（数据科学主席、人工智能与数据科学中心（CAIDAS）、乌尔姆大学）
Comments Tackling Climate Change with Machine Learning Workshop @ ICLR 2025
CircuitFusion: Multimodal Circuit Representation Learning for Agile Chip Design
Comments Accepted by ICLR 2025 (https://openreview.net/forum?id=rbnf7oe6JQ)
A Survey on Personalized Alignment -- The Missing Piece for Large Language Models in Real-World Applications
机构 * Ant Group（蚂蚁集团） ； Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所） ； Gaoling School of Artificial Intelligence, Renmin University of China（中国人民大学人工智能学院）
Comments Survey paper; 11 pages; Literature reviewed up to ICLR 2025
SiMHand: Mining Similar Hands for Large-Scale 3D Hand Pose Pre-training
机构 * The University of Tokyo（东京大学） ； Hunan University（湖南大学）
Comments ICLR 2025. arXiv admin note: text overlap with arXiv:2409.09714
3D Vision-Language Gaussian Splatting
机构 * Center for Research in Computer Vision, University of Central Florida（计算机视觉研究中心，佛罗里达大学）
Comments Accepted at ICLR 2025. Main paper + supplementary material
SoftCVI: Contrastive variational inference with self-generated soft labels
机构 * School of Mathematics, Bristol University, UK（数学系，布里斯托尔大学）
Comments Updated to match camera ready version accepted as ICLR spotlight
The Alignment Problem from a Deep Learning Perspective
机构 * OpenAI ； UC Berkeley（加州大学伯克利分校） ； University of Oxford（牛津大学）
Comments Published in ICLR 2024
Co$^{3}$Gesture: Towards Coherent Concurrent Co-speech 3D Gesture Generation with Interactive Diffusion
机构 * The Hong Kong University of Science and Technology（香港科学与技术大学） ； Peking University（北京大学）
Comments Accepted as ICLR 2025 (Spotlight)
Deconstructing Bias: A Multifaceted Framework for Diagnosing Cultural and Compositional Inequities in Text-to-Image Generative Models
机构 * Algoverse AI Research（Algoverse AI研究）
Comments Published at ICLR 2025 Workshop SynthData
Conformal Transformations for Symmetric Power Transformers
机构 * Stanford University（斯坦福大学） ； Manifest AI
Comments SCOPE Workshop at ICLR 2025
T-JEPA: Augmentation-Free Self-Supervised Learning for Tabular Data
机构 * Université Paris-Saclay, CNRS, CentraleSupélec, Laboratoire Interdisciplinaire des Sciences du Numérique（巴黎萨克雷大学、法国国家科学研究中心、中央超导电子学院、跨学科数字科学实验室）
Comments Accepted at ICLR 2025: https://openreview.net/forum?id=gx3LMRB15C
Improving Large Language Model Planning with Action Sequence Similarity
机构 * Google DeepMind（谷歌DeepMind） ； Carnegie Mellon University（卡内基梅隆大学）
Comments 25 pages, 11 figures
Journal ref The Thirteenth International Conference on Learning Representations (ICLR 2025)
Robust Root Cause Diagnosis using In-Distribution Interventions
Comments Accepted at ICLR-25
ICLR: In-Context Learning of Representations
机构 * CBS-NTT Program in Physics of Intelligence, Harvard University（哈佛大学物理智能联合项目） ； Department of Physics, Harvard University（哈佛大学物理系） ； Physics & Informatics Lab, NTT Research Inc.（NTT研究公司物理与信息学实验室） ； SEAS, Harvard University（哈佛大学科学与工程学院） ； CSE, University of Michigan, Ann Arbor（密歇根大学安娜堡分校计算机科学系）
Comments ICLR 2025
Journal ref International Conference on Learning Representations, 2025
Competition Dynamics Shape Algorithmic Phases of In-Context Learning
机构 * Harvard University（哈佛大学） ； NTT Research, Inc.（NTT研究公司） ； University of Michigan, Ann Arbor（密歇根大学安阿伯分校）
Comments ICLR 2025 Spotlight
Journal ref International Conference on Learning Representations, 2025
Offline Model-Based Optimization by Learning to Rank
机构 * National Key Laboratory for Novel Software Technology（新型软件技术国家重点实验室） ； School of Artificial Intelligence（人工智能学院） ； Key Laboratory of Water Big Data Technology of Ministry of Water Resources（水利部水大数据技术重点实验室） ； College of Computer Science and Software Engineering（计算机科学与软件工程学院） ； Advanced Computing and Storage Lab（先进计算与存储实验室） ； Huawei Technologies Co., Ltd.（华为技术有限公司）
Comments ICLR 2025
Deep Kernel Posterior Learning under Infinite Variance Prior Weights
Comments Published as a conference paper at ICLR 2025
MoDeGPT: Modular Decomposition for Large Language Model Compression
机构 * Samsung Research America（三星美国研究实验室）
Comments ICLR 2025 Oral
An evaluation of unconditional 3D molecular generation methods
Comments Published at the GEM workshop, ICLR 2025
