Aria-MIDI: A Dataset of Piano MIDI Files for Symbolic Music Modeling
机构 * Queen Mary University of London（伦敦大学Queen Mary分校）
Journal ref International Conference on Learning Representations (ICLR), 2025
期刊&会议
ICLR
International Conference on Learning Representations · 会议 · Machine Learning
Metadata, Wavelet, and Time Aware Diffusion Models for Satellite Image Super Resolution
机构 * Dept. of Information Engineering, Electronics, and Telecomm., Sapienza University of Rome（信息工程、电子与电信系，罗马萨皮恩扎大学）
Comments ICLR 2025 Workshop on Machine Learning for Remote Sensing (ML4RS)
Beware of Calibration Data for Pruning Large Language Models
机构 * School of Computer Science and Technology, Soochow University（苏州大学计算机科学与技术学院） ； Key Laboratory of Data Intelligence and Advanced Computing, Soochow University（苏州大学数据智能与先进计算重点实验室） ； Huawei Cloud, China（华为云（中国））
Comments Published as a conference paper at ICLR 2025
High-Precision Dichotomous Image Segmentation via Probing Diffusion Capacity
机构 * Dalian University of Technology（大连理工大学） ； vivo Mobile Communication Co., Ltd（vivo移动通信有限公司）
Comments Published as a conference paper at ICLR 2025
Sample then Identify: A General Framework for Risk Control and Assessment in Multimodal Large Language Models
机构 * University of Electronic Science and Technology of China（电子科技大学） ； Southern University of Science and Technology（南方科技大学） ； University of Birmingham（伯明翰大学） ； The University of Hong Kong（香港大学）
Comments Accepted by ICLR 2025 Spotlights
Hybrid Generative Modeling for Incomplete Physics: Deep Grey-Box Meets Optimal Transport
Comments Workshop paper at ICLR 2025 (XAI4Science Workshop)
CAPM: Fast and Robust Verification on Maxpool-based CNN via Dual Network
机构 * Graduate Institute of Communication Engineering, National Taiwan University（国立台湾大学通信工程研究所） ； Department of Electrical Engineering, National Taiwan University（国立台湾大学电子工程系） ； MediaTek Research（联发科技研究）
Journal ref ICLR 2025 Workshop: VerifAI: AI Verification in the Wild
L2MAC: Large Language Model Automatic Computer for Extensive Code Generation
机构 * University of Cambridge（剑桥大学）
Comments Published in The Twelfth International Conference on Learning Representations (ICLR), 2024. Copyright 2023 by the author(s)
Bilinear MLPs enable weight-based mechanistic interpretability
机构 * University of Antwerp（安特卫普大学） ； sqIRL/IDLab ； Apollo Research
Comments Accepted to ICLR'25
Why Uncertainty Calibration Matters for Reliable Perturbation-based Explanations
机构 * Siemens AG（西门子公司） ； 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）
Comments ICLR 2025 Workshop: XAI4Science: From Understanding Model Behavior to Discovering New Scientific Knowledge
Local Look-Ahead Guidance via Verifier-in-the-Loop for Automated Theorem Proving
机构 * Language Technology Lab, University of Amsterdam（阿姆斯特丹大学语言技术实验室） ； Qualcomm AI Research（高通人工智能研究）
Comments Accepted at the Findings of ACL 2025, Accepted at ICLR 2025 Workshop on Reasoning and Planning for Large Language Models
Brain Mapping with Dense Features: Grounding Cortical Semantic Selectivity in Natural Images With Vision Transformers
Comments Accepted at ICLR 2025, code: https://github.com/aluo-x/BrainSAIL
SD-KDE: Score-Debiased Kernel Density Estimation
机构 * Stanford University（斯坦福大学）
Comments ICLR 2025 Workshop on Frontiers of Probabilistic Inference
UniDrive: Towards Universal Driving Perception Across Camera Configurations
机构 * University of Michigan, Ann Arbor（密歇根大学） ； University of California, Berkeley（加州大学伯克利分校）
Comments ICLR 2025; 15 pages, 7 figures, 2 tables; Code at https://github.com/ywyeli/UniDrive
Directional Gradient Projection for Robust Fine-Tuning of Foundation Models
机构 * Georgia Institute of Technology（佐治亚理工学院）
Comments Accepted to ICLR 2025
LocoVR: Multiuser Indoor Locomotion Dataset in Virtual Reality
机构 * University of California Santa Barbara（加州大学圣芭芭拉分校） ； Toyota Motor North America（丰田北美公司）
Comments This paper has been accepted to Proceedings of the Thirteenth International Conference on Learning Representations (ICLR 2025), 2025
Journal ref "LocoVR: Multiuser Indoor Locomotion Dataset in Virtual Reality." Proceedings of the Thirteenth International Conference on Learning Representations (ICLR 2025), 2025
Probe before You Talk: Towards Black-box Defense against Backdoor Unalignment for Large Language Models
机构 * College of Cyber Science, Nankai University（南开大学网络科学学院） ； Alibaba Group（阿里巴巴集团） ； Zhejiang University（浙江大学） ； Nanyang Technological University（新加坡国立大学）
Comments Accepted at ICLR 2025
Journal ref Proceedings of The Thirteenth International Conference on Learning Representations (ICLR 2025)
A CLIP-Powered Framework for Robust and Generalizable Data Selection
机构 * National Key Laboratory for Novel Software Technology, Nanjing University（南京大学新型软件技术国家重点实验室） ； Shanghai Artificial Intelligence Laboratory（上海人工智能实验室） ； The Chinese University of Hong Kong（香港中文大学）
Comments ICLR 2025 Spotlight
Journal ref ICLR 2025
Open-Set Graph Anomaly Detection via Normal Structure Regularisation
机构 * Monash University（蒙纳士大学） ； Singapore Management University（新加坡管理大学） ； The University of Tokyo（东京大学） ； The University of Melbourne（墨尔本大学）
Comments Accepted by ICLR 2025
Event-Driven Online Vertical Federated Learning
机构 * Western University（西部大学） ； Vector Institute（向量研究所） ； Jilin University（吉林大学）
Comments Published as a conference paper at ICLR 2025
Journal ref The Thirteenth International Conference on Learning Representations (2025)
Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think
Comments ICLR 2025 (Oral). Project page: https://sihyun.me/REPA
Oscillatory State-Space Models
机构 * MIT（麻省理工学院）
Comments ICLR 2025 (Oral)
A General Framework for Off-Policy Learning with Partially-Observed Reward
机构 * Waseda University（早稻田大学） ； HAKUHODO Technologies Inc.（HAKUHODO技术公司） ； Cornell University（康奈尔大学）
Comments 10 pages, 5 figures. Published as a conference paper at ICLR 2025
Journal ref International Conference on Learning Representations (ICLR) 2025
ClimateChat: Designing Data and Methods for Instruction Tuning LLMs to Answer Climate Change Queries
机构 * Tsinghua University（清华大学）
Comments ICLR 2025 camera ready, 13 pages, 4 figures, 4 tables
LARP: Tokenizing Videos with a Learned Autoregressive Generative Prior
机构 * University of Maryland, College Park（马里兰大学 College Park 分校）
Comments ICLR 2025. Project page: https://hywang66.github.io/larp/
Uncovering Overfitting in Large Language Model Editing
机构 * Shandong University（山东大学） ； School of Computer Science, Beijing University of Posts and Telecommunications（北京邮电大学计算机学院） ； New Laboratory of Pattern Recognition (NLPR) State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS) Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）
Comments ICLR 2025
Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents
机构 * The Ohio State University（俄亥俄州立大学） ； Orby AI（Orby人工智能）
Comments Accepted to ICLR 2025 (Oral). Project Homepage: https://osu-nlp-group.github.io/UGround/
Interpretable Causal Representation Learning for Biological Data in the Pathway Space
机构 * CIMA University of Navarra, CCUN, IdiSNA（纳瓦拉大学CIMA、CCUN、IdiSNA） ； TECNUN, University of Navarra, San Sebastián（纳瓦拉大学TECNUN、桑塞巴斯蒂安） ； Biological and Environmental Science and Engineering Division, King Abdullah University of Science and Technology (KAUST)（科学技术大学（KAUST）生物与环境科学与工程部门） ； Dept. of Molecular Pharmacology and Experimental Therapeutics, Mayo Clinic（梅奥诊所分子药理学与实验治疗学部门） ； Institute of Chemical Biology, Ilia State University（伊利亚州立大学化学生物学研究所） ； Center for Data Science (DATAI), University of Navarra（纳瓦拉大学数据科学中心（DATAI））
Comments ICLR 2025, 28 pages, 14 figures, 10 tables
Let SSMs be ConvNets: State-space Modeling with Optimal Tensor Contractions
机构 * Brainchip Inc.（Brainchip公司）
Comments 25 pages, 7 figures
Journal ref International Conference on Learning Representations 2025
An Auditing Test To Detect Behavioral Shift in Language Models
机构 * UCL Centre for Artificial Intelligence, University College London（伦敦大学学院人工智能中心，大学学院伦敦） ； School of Informatics, University of Edinburgh（信息学院，爱丁堡大学） ； Miniml.AI, United Kingdom（Miniml.AI，英国） ； Polytechnique Montréal, Canada（蒙特利尔理工学院，加拿大） ； Mila – Quebec AI Institute, Canada（魁北克人工智能研究所，加拿大）
Comments 25 pages, 12 figures
Journal ref International Conference on Learning Representations (ICLR) 2025
