ICML

Re-ranking Reasoning Context with Tree Search Makes Large Vision-Language Models Stronger

Qi Yang, Chenghao Zhang, Lubin Fan, Kun Ding, Jieping Ye, Shiming Xiang

机构 * School of Artificial Intelligence, University of Chinese Academy of Sciences, China（中国科学院大学人工智能学院） ； MAIS, Institute of Automation, Chinese Academy of Sciences, China（中国科学院自动化研究所MAIS部） ； Alibaba Cloud Computing, China（阿里巴巴云计算） 纠错

Comments ICML 2025 Spotlight. 22 pages, 16 figures

E-LDA: Toward Interpretable LDA Topic Models with Strong Guarantees in Logarithmic Parallel Time

Adam Breuer

机构 * Department of Computer Science, Dartmouth College, Hanover, NH, USA（计算机科学系，达特茅斯学院） 纠错

Comments ICML 2025; Code available at: https://github.com/BreuerLabs/E- LDA

Journal ref In Proceedings of the 42nd International Conference on Machine Learning (ICML 2025), Vancouver, Canada. Proceedings of Machine Learning Research, Vol. 267, 2025

Improving Memory Efficiency for Training KANs via Meta Learning

Zhangchi Zhao, Jun Shu, Deyu Meng, Zongben Xu

机构 * School of Mathematics（数学系） ； Statistics, Ministry of Education Key Lab of Intelligent Networks（统计学，教育部智能网络重点实验室） ； Network Security, Xi'an Jiaotong University（网络安全，西安交通大学） ； Pengcheng Laboratory（鹏城实验室） 纠错

Comments ICML 2025

CoCoA-Mix: Confusion-and-Confidence-Aware Mixture Model for Context Optimization

Dasol Hong, Wooju Lee, Hyun Myung

机构 * Urban Robotics Lab, School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Republic of Korea（乌尔班机器人实验室，电气工程学院，韩国科学技术院，大韩民国） 纠错

Comments 8 pages, 5 figures; accepted at ICML 2025

Federated In-Context Learning: Iterative Refinement for Improved Answer Quality

Ruhan Wang, Zhiyong Wang, Chengkai Huang, Rui Wang, Tong Yu, Lina Yao, John C. S. Lui, Dongruo Zhou

机构 * Indiana University（印第安纳大学） ； The Chinese University of Hong Kong（香港中文大学） ； The University of New South Wales（新南威尔士大学） ； Adobe Research（Adobe研究） 纠错

Comments 27 pages, 16 figures. Accepted to ICML 2025

Adapter Naturally Serves as Decoupler for Cross-Domain Few-Shot Semantic Segmentation

Jintao Tong, Ran Ma, Yixiong Zou, Guangyao Chen, Yuhua Li, Ruixuan Li

机构 * School of Computer Science（计算机科学学院） ； Technology, Huazhong University of Science（科学技术大学技术学院） ； Peking University（北京大学） 纠错

Comments ICML 2025 Spotlight

Tokenized Bandit for LLM Decoding and Alignment

Suho Shin, Chenghao Yang, Haifeng Xu, Mohammad T. Hajiaghayi

机构 * University of Maryland（马里兰大学） ； University of Chicago（芝加哥大学） 纠错

Comments To appear at ICML 2025

Promoting Ensemble Diversity with Interactive Bayesian Distributional Robustness for Fine-tuning Foundation Models

Ngoc-Quan Pham, Tuan Truong, Quyen Tran, Tan Nguyen, Dinh Phung, Trung Le

机构 * Qualcomm AI Research, Qualcomm Vietnam Company Limited（高通人工智能研究, 高通越南公司） ； National University of Singapore, Singapore（新加坡国立大学） ； Monash University, Australia（墨尔本大学） 纠错

Comments ICML 2025 (Poster)

Towards Universal Offline Black-Box Optimization via Learning Language Model Embeddings

Rong-Xi Tan, Ming Chen, Ke Xue, Yao Wang, Yaoyuan Wang, Sheng Fu, Chao Qian

Comments ICML 2025

Text-to-LoRA: Instant Transformer Adaption

Rujikorn Charakorn, Edoardo Cetin, Yujin Tang, Robert Tjarko Lange

Comments Accepted at ICML 2025

FLAM: Frame-Wise Language-Audio Modeling

Yusong Wu, Christos Tsirigotis, Ke Chen, Cheng-Zhi Anna Huang, Aaron Courville, Oriol Nieto, Prem Seetharaman, Justin Salamon

Comments Accepted at ICML 2025 V2: fixed small typo on eq. 15 and eq. 17

Tree-Sliced Wasserstein Distance with Nonlinear Projection

Thanh Tran, Viet-Hoang Tran, Thanh Chu, Trang Pham, Laurent El Ghaoui, Tam Le, Tan M. Nguyen

Comments Accepted at ICML 2025

STAMP Your Content: Proving Dataset Membership via Watermarked Rephrasings

Saksham Rastogi, Pratyush Maini, Danish Pruthi

机构 * Saksham Rastogi（未知） ； Pratyush Maini（未知） ； Danish Pruthi（未知） 纠错

Comments Published at ICML 25, Code is available at https://github.com/codeboy5/stamp

MIB: A Mechanistic Interpretability Benchmark

Aaron Mueller, Atticus Geiger, Sarah Wiegreffe, Dana Arad, Iván Arcuschin, Adam Belfki, Yik Siu Chan, Jaden Fiotto-Kaufman, Tal Haklay, Michael Hanna, Jing Huang, Rohan Gupta, Yaniv Nikankin, Hadas Orgad, Nikhil Prakash, Anja Reusch, Aruna Sankaranarayanan, Shun Shao, Alessandro Stolfo, Martin Tutek, Amir Zur, David Bau, Yonatan Belinkov

Comments Accepted to ICML 2025. Project website at https://mib-bench.github.io

Bipartite Ranking From Multiple Labels: On Loss Versus Label Aggregation

Michal Lukasik, Lin Chen, Harikrishna Narasimhan, Aditya Krishna Menon, Wittawat Jitkrittum, Felix X. Yu, Sashank J. Reddi, Gang Fu, Mohammadhossein Bateni, Sanjiv Kumar

机构 * Google Research（谷歌研究） 纠错

Comments Accepted by ICML 2025

InfoSEM: A Deep Generative Model with Informative Priors for Gene Regulatory Network Inference

Tianyu Cui, Song-Jun Xu, Artem Moskalev, Shuwei Li, Tommaso Mansi, Mangal Prakash, Rui Liao

Comments ICML 2025

Generalized Interpolating Discrete Diffusion

Dimitri von Rütte, Janis Fluri, Yuhui Ding, Antonio Orvieto, Bernhard Schölkopf, Thomas Hofmann

Comments Published at ICML 2025; Code available at https://github.com/dvruette/gidd

PhantomWiki: On-Demand Datasets for Reasoning and Retrieval Evaluation

Albert Gong, Kamilė Stankevičiūtė, Chao Wan, Anmol Kabra, Raphael Thesmar, Johann Lee, Julius Klenke, Carla P. Gomes, Kilian Q. Weinberger

Comments Accepted to ICML 2025

AMPO: Active Multi-Preference Optimization for Self-play Preference Selection

Taneesh Gupta, Rahul Madhavan, Xuchao Zhang, Chetan Bansal, Saravan Rajmohan

机构 * Microsoft（微软） 纠错

Comments Accepted at ICML 2025

ParallelComp: Parallel Long-Context Compressor for Length Extrapolation

Jing Xiong, Jianghan Shen, Chuanyang Zheng, Zhongwei Wan, Chenyang Zhao, Chiwun Yang, Fanghua Ye, Hongxia Yang, Lingpeng Kong, Ngai Wong

Comments This paper has been accepted by ICML 2025

AdaSplash: Adaptive Sparse Flash Attention

Nuno Gonçalves, Marcos Treviso, André F. T. Martins

Comments Accepted as spotlight in ICML 2025

Minerva: A Programmable Memory Test Benchmark for Language Models

Menglin Xia, Victor Ruehle, Saravan Rajmohan, Reza Shokri

Comments ICML 2025

RepLoRA: Reparameterizing Low-Rank Adaptation via the Perspective of Mixture of Experts

Tuan Truong, Chau Nguyen, Huy Nguyen, Minh Le, Trung Le, Nhat Ho

Comments ICML 2025 (Poster)

Trajectory World Models for Heterogeneous Environments

Shaofeng Yin, Jialong Wu, Siqiao Huang, Xingjian Su, Xu He, Jianye Hao, Mingsheng Long

Comments ICML 2025. Code is available at https://github.com/thuml/TrajWorld

Scalable Sobolev IPM for Probability Measures on a Graph

Tam Le, Truyen Nguyen, Hideitsu Hino, Kenji Fukumizu

Comments To appear in International Conference on Machine Learning (ICML), 2025

ExLM: Rethinking the Impact of [MASK] Tokens in Masked Language Models

Kangjie Zheng, Junwei Yang, Siyue Liang, Bin Feng, Zequn Liu, Wei Ju, Zhiping Xiao, Ming Zhang

Comments 30 pages, 12 figures; ICML 2025

A Cognac Shot To Forget Bad Memories: Corrective Unlearning for Graph Neural Networks

Varshita Kolipaka, Akshit Sinha, Debangan Mishra, Sumit Kumar, Arvindh Arun, Shashwat Goel, Ponnurangam Kumaraguru

Comments In Proceedings of ICML 2025

Tilted Sharpness-Aware Minimization

Tian Li, Tianyi Zhou, Jeffrey A. Bilmes

Comments Accepted by ICML 2025

Theoretical Limitations of Ensembles in the Age of Overparameterization

Niclas Dern, John P. Cunningham, Geoff Pleiss

Comments Accepted for publication at ICML 2025. 33 pages, 17 figures

Parameter-Efficient Fine-Tuning of State Space Models

Kevin Galim, Wonjun Kang, Yuchen Zeng, Hyung Il Koo, Kangwook Lee

Comments Accepted at ICML 2025. Code is available at https://github.com/furiosa-ai/ssm-peft