ICLR

Is Large-Scale Pretraining the Secret to Good Domain Generalization?

Piotr Teterwak, Kuniaki Saito, Theodoros Tsiligkaridis, Bryan A. Plummer, Kate Saenko

机构 * Boston University（波士顿大学） ； MIT Lincoln Laboratory（麻省理工学院林肯实验室） 纠错

Comments Accepted at ICLR 2025

Non-Adversarial Inverse Reinforcement Learning via Successor Feature Matching

Arnav Kumar Jain, Harley Wiltzer, Jesse Farebrother, Irina Rish, Glen Berseth, Sanjiban Choudhury

机构 * Mila – Québec AI Institute（魁北克AI研究所） ； Université de Montréal（蒙特利尔大学） ； McGill University（麦吉尔大学） ； Cornell University（康奈尔大学） 纠错

Comments Accepted to ICLR 2025

AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models

Junfeng Fang, Houcheng Jiang, Kun Wang, Yunshan Ma, Shi Jie, Xiang Wang, Xiangnan He, Tat-seng Chua

机构 * University of Science and Technology of China（中国科学技术大学） ； National University of Singapore（新加坡国立大学） 纠错

Journal ref 13th International Conference on Learning Representations (ICLR 2025 Oral)

On the Price of Differential Privacy for Hierarchical Clustering

Chengyuan Deng, Jie Gao, Jalaj Upadhyay, Chen Wang, Samson Zhou

Comments ICLR 2025

InstaRevive: One-Step Image Enhancement via Dynamic Score Matching

Yixuan Zhu, Haolin Wang, Ao Li, Wenliang Zhao, Yansong Tang, Jingxuan Niu, Lei Chen, Jie Zhou, Jiwen Lu

机构 * Department of Automation, Tsinghua University（自动化系，清华大学） ； Tsinghua Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院，清华大学） 纠错

Comments Accepted by ICLR 2025

Manifold Induced Biases for Zero-shot and Few-shot Detection of Generated Images

Jonathan Brokman, Amit Giloni, Omer Hofman, Roman Vainshtein, Hisashi Kojima, Guy Gilboa

机构 * Fujitsu（富士通） 纠错

Comments Accepted to ICLR 2025 (The International Conference on Learning Representations)

Improving Learning to Optimize Using Parameter Symmetries

Guy Zamir, Aryan Dokania, Bo Zhao, Rose Yu

机构 * University of Wisconsin-Madison（威斯康星大学麦迪逊分校） ； University of California, San Diego（加州大学圣地亚哥分校） 纠错

Comments Accepted at the ICLR Workshop on Neural Network Weights as a New Data Modality 2025

Solving New Tasks by Adapting Internet Video Knowledge

Calvin Luo, Zilai Zeng, Yilun Du, Chen Sun

Comments ICLR 2025. Project Webpage: https://diffusion-supervision.github.io/adapt2act/

Vertical Federated Learning with Missing Features During Training and Inference

Pedro Valdeira, Shiqiang Wang, Yuejie Chi

机构 * Department of Electrical and Computer Engineering, Carnegie Mellon University（卡内基梅隆大学电气与计算机工程系） ； Institute for Systems and Robotics, Instituto Superior Técnico（系统与机器人研究所，技术高等学院） ； IBM T. J. Watson Research Center（IBM沃森研究中心） ； CMU（卡内基梅隆大学） ； IBM Research（IBM研究） 纠错

Comments Accepted to ICLR 2025

An Undetectable Watermark for Generative Image Models

Sam Gunn, Xuandong Zhao, Dawn Song

机构 * UC Berkeley（加州大学伯克利分校） 纠错

Comments ICLR 2025

AutoDAN-Turbo: A Lifelong Agent for Strategy Self-Exploration to Jailbreak LLMs

Xiaogeng Liu, Peiran Li, Edward Suh, Yevgeniy Vorobeychik, Zhuoqing Mao, Somesh Jha, Patrick McDaniel, Huan Sun, Bo Li, Chaowei Xiao

机构 * University of Wisconsin–Madison（威斯康星大学麦迪逊分校） ； NVIDIA（NVIDIA公司） ； Cornell University（康奈尔大学） ； Washington University, St. Louis（圣路易斯华盛顿大学） ； University of Michigan, Ann Arbor（安娜堡大学） ； The Ohio State University（俄亥俄州立大学） ； UIUC（伊利诺伊大学厄巴纳-香槟分校） 纠错

Comments ICLR 2025 Spotlight. Project Page: https://autodans.github.io/AutoDAN-Turbo Code: https://github.com/SaFoLab-WISC/AutoDAN-Turbo

Certifying Counterfactual Bias in LLMs

Isha Chaudhary, Qian Hu, Manoj Kumar, Morteza Ziyadi, Rahul Gupta, Gagandeep Singh

机构 * UIUC（伊利诺伊大学香槟分校） ； Amazon（亚马逊） ； Oracle Health（Oracle健康） 纠错

Comments Published at ICLR 2025

Ambient Diffusion Posterior Sampling: Solving Inverse Problems with Diffusion Models Trained on Corrupted Data

Asad Aali, Giannis Daras, Brett Levac, Sidharth Kumar, Alexandros G. Dimakis, Jonathan I. Tamir

Journal ref ICLR, 2025

Compute-Optimal LLMs Provably Generalize Better With Scale

Marc Finzi, Sanyam Kapoor, Diego Granziol, Anming Gu, Christopher De Sa, J. Zico Kolter, Andrew Gordon Wilson

机构 * Carnegie Mellon University（卡内基梅隆大学） ； New York University（纽约大学） ； PureStrength AI ； Boston University（波士顿大学） ； Cornell University（康奈尔大学） 纠错

Comments ICLR 2025

Latent Bayesian Optimization via Autoregressive Normalizing Flows

Seunghun Lee, Jinyoung Park, Jaewon Chu, Minseo Yoon, Hyunwoo J. Kim

机构 * Department of Computer Science and Engineering, Korea University（韩国大学计算机科学与工程系） ； School of Computing, KAIST（韩国科学技术院计算机科学学院） 纠错

Comments ICLR 2025

Dynamic Contrastive Skill Learning with State-Transition Based Skill Clustering and Dynamic Length Adjustment

Jinwoo Choi, Seung-Woo Seo

机构 * Department of Electrical and Computer Engineering（电气与计算机工程系） ； Seoul National University（首尔国立大学） 纠错

Comments ICLR 2025; 23 pages, 12 figures

Mitigating Parameter Interference in Model Merging via Sharpness-Aware Fine-Tuning

Yeoreum Lee, Jinwook Jung, Sungyong Baik

机构 * Dept. of Artificial Intelligence（人工智能系） ； Dept. of Data Science（数据科学系） 纠错

Comments ICLR 2025

MP-Mat: A 3D-and-Instance-Aware Human Matting and Editing Framework with Multiplane Representation

Siyi Jiao, Wenzheng Zeng, Yerong Li, Huayu Zhang, Changxin Gao, Nong Sang, Mike Zheng Shou

机构 * Key Laboratory of Image Processing and Intelligent Control, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology（图像处理与智能控制重点实验室，人工智能与自动化学院，华中科技大学） ； Show Lab, National University of Singapore（Show实验室，新加坡国立大学） 纠错

Comments Accepted by ICLR 2025

Using street view imagery and deep generative modeling for estimating the health of urban forests

Akshit Gupta, Remko Uijlenhoet

机构 * TU Delft（代尔夫特理工大学） 纠错

Comments Accepted at ICLR 2025 Workshop

Aioli: A Unified Optimization Framework for Language Model Data Mixing

Mayee F. Chen, Michael Y. Hu, Nicholas Lourie, Kyunghyun Cho, Christopher Ré

机构 * Department of Computer Science, Stanford University（斯坦福大学计算机科学系） ； Center for Data Science, New York University（纽约大学数据科学中心） ； Computer Science Department, New York University（纽约大学计算机科学系） ； Prescient Design, Genentech（基因泰克预见设计） 纠错

Comments ICLR 2025 Camera Ready

Inference Optimal VLMs Need Fewer Visual Tokens and More Parameters

Kevin Y. Li, Sachin Goyal, Joao D. Semedo, J. Zico Kolter

机构 * Carnegie Mellon University（卡内基梅隆大学） ； Bosch Center for Artificial Intelligence（博世人工智能中心） 纠错

Comments Published at ICLR 2025

How Does Critical Batch Size Scale in Pre-training?

Hanlin Zhang, Depen Morwani, Nikhil Vyas, Jingfeng Wu, Difan Zou, Udaya Ghai, Dean Foster, Sham Kakade

机构 * Harvard University（哈佛大学） ； Kempner Institute at Harvard University（哈佛大学凯普勒研究所） ； University of California, Berkeley（加州大学伯克利分校） ； The University of Hong Kong（香港大学） ； Amazon（亚马逊公司） 纠错

Comments ICLR 2025, Blog post: https://kempnerinstitute.harvard.edu/research/deeper-learning/how-does-critical-batch-size-scale-in-pre-training-decoupling-data-and-model-size

Metamizer: a versatile neural optimizer for fast and accurate physics simulations

Nils Wandel, Stefan Schulz, Reinhard Klein

机构 * Department of Computer Science University of Bonn（计算机科学系 波恩大学） 纠错

Comments to be published at International Conference on Learning Representations (ICLR) 2025

Fine-Grained Verifiers: Preference Modeling as Next-token Prediction in Vision-Language Alignment

Chenhang Cui, An Zhang, Yiyang Zhou, Zhaorun Chen, Gelei Deng, Huaxiu Yao, Tat-Seng Chua

机构 * National University of Singapore（新加坡国立大学） ； UNC-Chapel Hill（北卡罗来纳大学教堂山分校） ； University of Chicago（芝加哥大学） ； Nanyang Technological University（南洋理工大学） 纠错

Comments 23 pages; Published as a conference paper at ICLR 2025

Preference Diffusion for Recommendation

Shuo Liu, An Zhang, Guoqing Hu, Hong Qian, Tat-seng Chua

Comments Accepted by ICLR 2025

Journal ref ICLR 2025

BlendRL: A Framework for Merging Symbolic and Neural Policy Learning

Hikaru Shindo, Quentin Delfosse, Devendra Singh Dhami, Kristian Kersting

Comments ICLR 2025 (Spotlight)

Context-Parametric Inversion: Why Instruction Finetuning Can Worsen Context Reliance

Sachin Goyal, Christina Baek, J. Zico Kolter, Aditi Raghunathan

机构 * Carnegie Mellon University（卡内基梅隆大学） 纠错

Comments Published at ICLR 2025 (Oral)

Adversarially Robust Out-of-Distribution Detection Using Lyapunov-Stabilized Embeddings

Hossein Mirzaei, Mackenzie W. Mathis

机构 * École Polytechnique Fédérale de Lausanne (EPFL)（瑞士联邦理工学院（EPFL）） 纠错

Comments Accepted at the International Conference on Learning Representations (ICLR) 2025. Code and pre-trained models are available at https://github.com/AdaptiveMotorControlLab/AROS

Journal ref ICLR 2025

LOKI: A Comprehensive Synthetic Data Detection Benchmark using Large Multimodal Models

Junyan Ye, Baichuan Zhou, Zilong Huang, Junan Zhang, Tianyi Bai, Hengrui Kang, Jun He, Honglin Lin, Zihao Wang, Tong Wu, Zhizheng Wu, Yiping Chen, Dahua Lin, Conghui He, Weijia Li

Comments ICLR 2025 SPOTLIGHT, 83 pages, 63 figures

How Do Large Language Models Understand Graph Patterns? A Benchmark for Graph Pattern Comprehension

Xinnan Dai, Haohao Qu, Yifen Shen, Bohang Zhang, Qihao Wen, Wenqi Fan, Dongsheng Li, Jiliang Tang, Caihua Shan

机构 * Michigan State University（密歇根州立大学） ； The Hong Kong Polytechnic University（香港理工大学） ； Microsoft Research（微软研究院） 纠错

Comments The paper is published in ICLR 2025