ICML

Emergent Response Planning in LLMs

Zhichen Dong, Zhanhui Zhou, Zhixuan Liu, Chao Yang, Chaochao Lu

机构 * Shanghai Artificial Intelligence Laboratory（上海人工智能实验室） 纠错

Comments Published at ICML 2025. Code available at: https://github.com/niconi19/Emergent-Response-Planning-in-LLMs

Journal ref Proceedings of the 42nd International Conference on Machine Learning (ICML 2025)

Speak Easy: Eliciting Harmful Jailbreaks from LLMs with Simple Interactions

Yik Siu Chan, Narutatsu Ri, Yuxin Xiao, Marzyeh Ghassemi

机构 * Brown University（布朗大学） ； Columbia University（哥伦比亚大学） ； Massachusetts Institute of Technology（麻省理工学院） 纠错

Comments ICML 2025

Core Context Aware Transformers for Long Context Language Modeling

Yaofo Chen, Zeng You, Shuhai Zhang, Haokun Li, Yirui Li, Yaowei Wang, Mingkui Tan

机构 * South China University of Technology（华南理工大学） ； Peng Cheng Laboratory（鹏城实验室） ； Pazhou Laboratory（琶洲实验室） ； Harbin Institute of Technology（哈尔滨工业大学） ； Key Laboratory of Big Data and Intelligent Robot, Ministry of Education（教育部大数据与智能机器人重点实验室） 纠错

Comments Accepted for publication at ICML 2025

Machines and Mathematical Mutations: Using GNNs to Characterize Quiver Mutation Classes

Jesse He, Helen Jenne, Herman Chau, Davis Brown, Mark Raugas, Sara Billey, Henry Kvinge

机构 * Halıcıoğlu Data Science Institute, University of California San Diego, San Diego, CA, USA（哈利奇奥格鲁数据科学研究所，加州大学圣地亚哥分校） ； Pacific Northwest National Laboratory, Richland, WA, USA（太平洋西北国家实验室） ； Department of Mathematics, University of Washington, Seattle, WA, USA（数学系，华盛顿大学） 纠错

Comments ICML 2025. v3: Corrected typo in references

HiPPO-Prophecy: State-Space Models can Provably Learn Dynamical Systems in Context

Federico Arangath Joseph, Kilian Konstantin Haefeli, Noah Liniger, Caglar Gulcehre

机构 * ETH Zurich（苏黎世联邦理工学院） ； EPFL（苏黎世联邦理工学院） 纠错

Comments ICML 2024, Next Generation Sequence Modeling Architectures Workshop

Theoretically Unmasking Inference Attacks Against LDP-Protected Clients in Federated Vision Models

Quan Nguyen, Minh N. Vu, Truc Nguyen, My T. Thai

机构 * Department of CISE, University of Florida, USA.（佛罗里达大学计算机与信息科学系） ； Center for Nonlinear Studies, Los Alamos National Lab, USA（洛斯阿拉莫斯国家实验室非线性研究中心） ； Computational Science Center, National Renewable Energy Laboratory, USA（国家可再生能源实验室计算科学中心） 纠错

Comments Accepted to ICML 2025

Unable to Forget: Proactive Interference Reveals Working Memory Limits in LLMs Beyond Context Length

Chupei Wang, Jiaqiu Vince Sun

Comments Accepted at ICML 2025 Workshop on Long Context Foundation Models (ICFM). Code: https://github.com/zhuangziGiantfish/Unable-to-Forget

Neutral Residues: Revisiting Adapters for Model Extension

Franck Signe Talla, Edouard Grave, Hervé Jégou

Comments Accepted at ICML 2025

MetaAgent: Automatically Constructing Multi-Agent Systems Based on Finite State Machines

Yaolun Zhang, Xiaogeng Liu, Chaowei Xiao

机构 * University of Wisconsin - Madison（威斯康星大学麦迪逊分校） 纠错

Comments ICML 2025

FGFP: A Fractional Gaussian Filter and Pruning for Deep Neural Networks Compression

Kuan-Ting Tu, Po-Hsien Yu, Yu-Syuan Tseng, Shao-Yi Chien

机构 * Graduate School of Advanced Technology, National Taiwan University, Taipei, Taiwan（国家台湾大学高级科技研究所） ； Graduate Institute of Electronics Engineering, National Taiwan University, Taipei, Taiwan（国家台湾大学电子工程研究所） 纠错

Comments 8 pages, 2 figures, 4 tables, Accepted by ICML 2025 Workshop (TTODLer-FM)

Visual Language Models as Zero-Shot Deepfake Detectors

Viacheslav Pirogov

Comments Accepted to the ICML 2025 Workshop on Reliable and Responsible Foundation Models

High-Resolution Live Fuel Moisture Content (LFMC) Maps for Wildfire Risk from Multimodal Earth Observation Data

Patrick Alan Johnson, Gabriel Tseng, Yawen Zhang, Heather Heward, Virginia Sjahli, Favyen Bastani, Joseph Redmon, Patrick Beukema

机构 * Mila -- Quebec AI Institute（魁北克人工智能研究所） ； McGill University（麦吉尔大学） ； Allen Institute for AI (Ai2)（人工智能研究所） ； University of Idaho（爱达荷大学） 纠错

Comments 10 pages, ICML 2025 (TerraBytes)

Don't Lag, RAG: Training-Free Adversarial Detection Using RAG

Roie Kazoom, Raz Lapid, Moshe Sipper, Ofer Hadar

机构 * Electrical and Computer Engineering, Ben Gurion University, Beer Sheba 84105, Israel（电子与计算机工程系，本· Gurion 大学） ； Computer Science, Ben Gurion University, Beer Sheba 84105, Israel（计算机科学系，本· Gurion 大学） 纠错

Comments Accepted at VecDB @ ICML 2025

Lightweight Online Adaption for Time Series Foundation Model Forecasts

Thomas L. Lee, William Toner, Rajkarn Singh, Artjom Joosen, Martin Asenov

机构 * intern（实习生） ； Huawei SIR Lab, Edinburgh Research Centre, UK（华为SIR实验室，爱丁堡研究中心，英国） ； School of Informatics, University of Edinburgh, UK（信息学院，爱丁堡大学，英国） 纠错

Comments 9 pages, Published at ICML 2025

CollabLLM: From Passive Responders to Active Collaborators

Shirley Wu, Michel Galley, Baolin Peng, Hao Cheng, Gavin Li, Yao Dou, Weixin Cai, James Zou, Jure Leskovec, Jianfeng Gao

机构 * stanford（斯坦福大学） ； microsoft（微软公司） 纠错

Comments Outstanding Paper Award at ICML 2025

Scaling and Data Saturation in Protein Language Models

Aviv Spinner, Erika DeBenedictis, Corey M. Hudson

Comments Presented at the GenBio Workshop, ICML 2025

Principled Curriculum Learning using Parameter Continuation Methods

Harsh Nilesh Pathak, Randy Paffenroth

机构 * Department of Data Science, Worcester Polytechnic Institute, USA（数据科学系，沃斯特理工学院，美国） ； Department of Mathematics, Computer and Data Science, Worcester Polytechnic Institute, USA（数学、计算机与数据科学系，沃斯特理工学院，美国） 纠错

Journal ref ICML 2025, Beyond first-order methods in ML systems

TypyBench: Evaluating LLM Type Inference for Untyped Python Repositories

Honghua Dong, Jiacheng Yang, Xun Deng, Yuhe Jiang, Gennady Pekhimenko, Fan Long, Xujie Si

机构 * University of Toronto（多伦多大学） ； Vector Institute（向量研究所） ； CIFAR AI Chair（CIFAR人工智能主席） 纠错

Journal ref Proceedings of the 42nd International Conference on Machine Learning, Vancouver, Canada. PMLR 267, 2025

Sample Complexity of Branch-length Estimation by Maximum Likelihood

David Clancy, Hanbaek Lyu, Sebastien Roch

Comments 22 pages, 2 figures

Journal ref ICML 2025

Half-life of Youtube News Videos: Diffusion Dynamics and Predictive Factors

Anahit Sargsyan, Hridoy Sankar Dutta, Juergen Pfeffer

Comments To be published in International Conference on Machine Learning, Optimization, and Data Science (LOD 2025)

INTEGRALBENCH: Benchmarking LLMs with Definite Integral Problems

Bintao Tang, Xin Yang, Yuhao Wang, Zixuan Qiu, Zimo Ji, Wenyuan Jiang

机构 * School of Software Engineering, Tongji University, Shanghai, China（同济大学软件工程学院） ； Polytechnic Institute, Zhejiang University, Zhejiang, China（浙江大学 polytechnic 院） ； ETH Zurich, Zurich, Switzerland（苏黎世联邦理工学院） ； Department of Computer Science（计算机科学系） ； Engineering, Hong Kong University of Science（科学大学工程学院） ； School of Mathematics（数学学院） ； Physics, Xi’an Jiaotong Liverpool University, Suzhou, China（西安交通大学利物浦大学物理学院） 纠错

Comments 19 pages, 5 figures

Journal ref 2nd AI for Math Workshop @ ICML 2025

UncertainSAM: Fast and Efficient Uncertainty Quantification of the Segment Anything Model

Timo Kaiser, Thomas Norrenbrock, Bodo Rosenhahn

机构 * Institute for Information Processing / L3S - Leibniz University Hannover, Germany（信息处理研究所 / L3S - 哈根大学） 纠错

Comments Accepted to ICML'25

KinDEL: DNA-Encoded Library Dataset for Kinase Inhibitors

Benson Chen, Tomasz Danel, Gabriel H. S. Dreiman, Patrick J. McEnaney, Nikhil Jain, Kirill Novikov, Spurti Umesh Akki, Joshua L. Turnbull, Virja Atul Pandya, Boris P. Belotserkovskii, Jared Bryce Weaver, Ankita Biswas, Dat Nguyen, Kent Gorday, Mohammad Sultan, Nathaniel Stanley, Daniel M Whalen, Divya Kanichar, Christoph Klein, Emily Fox, R. Edward Watts

Journal ref Forty-Second International Conference on Machine Learning (2025)

The pitfalls of next-token prediction

Gregor Bachmann, Vaishnavh Nagarajan

机构 * ETH Zürich, Switzerland（苏黎世联邦理工学院） ； Google Research, US（谷歌研究院） 纠错

Comments ICML 2024

Finetuning Stellar Spectra Foundation Models with LoRA

Xiaosheng Zhao, Yuan-Sen Ting, Alexander S. Szalay, Yang Huang

Comments 7 pages, 2 figures. Accepted to the Machine Learning for Astrophysics (ML4Astro) Colocated Workshop at ICML 2025. Presented as a spotlight talk

Modeling User Behavior from Adaptive Surveys with Supplemental Context

Aman Shukla, Daniel Patrick Scantlebury, Rishabh Kumar

机构 * Resonate Networks, Inc.（Resonate Networks公司） 纠错

Comments Best Paper, NewInML @ ICML 2025

Set-based Implicit Likelihood Inference of Galaxy Cluster Mass

Bonny Y. Wang, Leander Thiele

机构 * McWilliams Center for Cosmology and Astrophysics, Department of Physics, Carnegie Mellon University（麦卡威廉宇宙学与天体物理学中心，物理系，卡内基梅隆大学） ； Entertainment Technology Center, Carnegie Mellon University（娱乐技术中心，卡内基梅隆大学） ； Kavli IPMU (WPI), UTIAS, The University of Tokyo（Kavli IPMU（WPI），UTIAS，东京大学） ； Center for Data-Driven Discovery, Kavli IPMU (WPI), Japan（数据驱动发现中心，Kavli IPMU（WPI），日本） 纠错

Comments 5 pages, 4 figures; accepted as a spotlight talk at ICML-colocated ML4Astro 2025 workshop

Do Not Mimic My Voice: Speaker Identity Unlearning for Zero-Shot Text-to-Speech

Taesoo Kim, Jinju Kim, Dongchan Kim, Jong Hwan Ko, Gyeong-Moon Park

机构 * Department of Electrical and Computer Engineering, Sungkyunkwan University（电气与计算机工程系，全北大学） ； KT Corporation（KT公司） ； Visiting fellow at Carnegie Mellon University（卡内基梅隆大学访问学者） ； Department of Artificial Intelligence, Korea University（人工智能系，韩国大学） 纠错

Comments Proceedings of the 42nd International Conference on Machine Learning (ICML 2025), Vancouver, Canada. PMLR 267, 2025. Authors Jinju Kim and Taesoo Kim contributed equally

Implementing Adaptations for Vision AutoRegressive Model

Kaif Shaikh, Franziska Boenisch, Adam Dziedzic

机构 * cispa（CISPA信息安全研究中心） 纠错

Comments Accepted at DIG-BUGS: Data in Generative Models Workshop @ ICML 2025

Come Together, But Not Right Now: A Progressive Strategy to Boost Low-Rank Adaptation

Zhan Zhuang, Xiequn Wang, Wei Li, Yulong Zhang, Qiushi Huang, Shuhao Chen, Xuehao Wang, Yanbin Wei, Yuhe Nie, Kede Ma, Yu Zhang, Ying Wei

机构 * Southern University of Science（南方科技大学） ； City University of Hong Kong（香港城市大学） ； Zhejiang University（浙江大学） ； Hong Kong University of Science（香港科学大学） ； University of Surrey（萨里大学） ； New York University（纽约大学） 纠错

Comments Accepted by ICML 2025. Code link: https://github.com/zwebzone/coto