Center for Advanced Intelligence Project

29 papers have been accepted at the International Conference on Machine Learning (ICML) 2025, a major conference on Artificial Intelligence (July 13-19, 2025, Vancouver, Canada).
For more details, please refer to the link below.

[Website]https://icml.cc/Conferences/2025

[Acceptance rate]
There are 3,260 accepted papers from 12,107 submissions, leading to a 26.9% percent acceptance rate.
*Spotlight Posters:2.6% (313/12,107)

Spotlight (2 papers)

• Parallel Simulation for Sampling under Isoperimetry and Score-based Diffusion Models
  Huanjian Zhou (The University of Tokyo / RIKEN AIP)
  Masashi Sugiyama (RIKEN AIP /The University of Tokyo)
• Statistical Test for Feature Selection Pipelines by Selective Inference
  Tomohiro Shiraishi* (Nagoya University, RIKEN AIP)
  Tatsuya Matsukawa* (Nagoya University)
  Shuichi Nishino (Nagoya University, RIKEN AIP)
  Ichiro Takeuchi (Nagoya University, RIKEN AIP)
  *Equal Contribution

Posters (27 papers)

• Adaptive Localization of Knowledge Negation for Continual LLM Unlearning
  Abudukelimu Wuerkaixi (Tsinghua University / RIKEN AIP)+
  Qizhou Wang (Hong Kong Baptist University) +
  Sen Cui (Tsinghua University)++
  Wutong Xu (Tsinghua University)
  Bo Han (Hong Kong Baptist University / RIKEN AIP)
  Gang Niu (RIKEN AIP)
  Masashi Sugiyama (RIKEN AIP / The University of Tokyo)
  Changshui Zhang (Tsinghua University)
• Can Diffusion Models Learn Hidden Inter-Feature Rules Behind Images?
  Yujin Han (HKU)
  Andi Han (RIKEN AIP)
  Wei Huang (RIKEN AIP)
  Chaochao Lu (Shanghai AI Lab)
  Difan Zou (HKU)
• Direct Density Ratio Optimization: A Statistically Consistent Approach to Aligning Large Language Models
  Rei Higuchi (The University of Tokyo / RIKEN AIP)
  Taiji Suzuki (The University of Tokyo / RIKEN AIP)
• Distributionally Robust Active Learning for Gaussian Process Regression
  Shion Takeno (Nagoya University)
  Yoshito Okura (Nagoya University)
  Yu Inatsu (Nagoya Institute of Technology)
  Aoyama Tatsuya (Nagoya University)
  Tomonari Tanaka (Nagoya University)
  Akahane Satoshi (Nagoya University)
  Hiroyuki Hanada (Nagoya University)
  Noriaki Hashimoto (RIKEN AIP)
  Taro Murayama (DENSO)
  Hanju Lee (DENSO)
  Shinya Kojima (DENSO)
  Ichiro Takeuchi (Nagoya University / RIKEN AIP)
• Efficient optimization with orthogonality constraint: a randomized Riemannian submanifold method.
  Andi Han (RIKEN AIP)
  Pierre-Louis Poirion (RIKEN AIP)
  Akiko Takeda (University of Tokyo / RIKEN AIP)
• Geometric Resampling in Nearly Linear Time for Follow-the-Perturbed-Leader with Best-of-Both-Worlds Guarantee in Bandit Problems
  Botao Chen* (Kyoto University)
  Jongyeong Lee* (Seoul National University)
  Junya Honda (Kyoto University, RIKEN AIP)
  * Equal contribution
• Gradual Transition from Bellman Optimality Operator to Bellman Operator in Online Reinforcement Learning
  Motoki Omura (The University of Tokyo)
  Kazuki Ota (The University of Tokyo)
  Takayuki Osa (RIKEN AIP)
  Yusuke Mukuta (The University of Tokyo)
  Tatsuya Harada (The University of Tokyo / RIKEN AIP)
• Heavy-Tailed Linear Bandits: Huber Regression with One-Pass Update
  Jing Wang (Nanjing University)
  Yu-Jie Zhang (RIKEN AIP)
  Peng Zhao (Nanjing University)
  Zhi-Hua Zhou (Nanjing University)
• Learning without Isolation: Pathway Protection for Continual Learning
  Zhikang Chen* (Tsinghua University / RIKEN AIP)
  Abudukelimu Wuerkaixi* (Tsinghua University/RIKEN AIP)
  Sen Cui* (Tsinghua University)
  Haoxuan Li (Peking University)
  Ding Li (Tsinghua University)
  Jingfeng Zhang (The University of Auckland / RIKEN AIP)
  Bo Han (Hong Kong Baptist University / RIKEN AIP)
  Gang Niu (RIKEN AIP)
  Houfang Liu (Tsinghua University)
  Yi Yang (Tsinghua University)
  Sifan Yang (Tsinghua University)
  Changshui Zhang† (Tsinghua University)
  Tianling Ren† (Tsinghua University)
  * Equal contribution
  † Corresponding author
• Low-Rank Tensor Transitions (LoRT) for Transferable Tensor Regression
  Andong Wang (RIKEN AIP)
  Yuning Qiu (RIKEN AIP)
  Zhong Jin (China University of Petroleum-Beijing at Karamay)
  Guoxu Zhou (Guangdong University of Technology)
  Qibin Zhao (RIKEN AIP)
• Metastable Dynamics of Chain-of-Thought Reasoning: Provable Benefits of Search, RL and Distillation
  Juno Kim (The University of Tokyo / RIKEN AIP)
  Denny Wu (New York University / Flatiron Institute)
  Jason D. Lee (UC Berkeley)
  Taiji Suzuki (The University of Tokyo / RIKEN AIP)
• Mixture of Experts Provably Detect and Learn the Latent Cluster Structure in Gradient-Based Learning
  Ryotaro Kawata (The University of Tokyo / RIKEN AIP)
  Kohsei Matsutani (The University of Tokyo / RIKEN AIP)
  Yuri Kinoshita (The University of Tokyo)
  Naoki Nishikawa (The University of Tokyo / RIKEN AIP)
  Taiji Suzuki (The University of Tokyo / RIKEN AIP)
• Modified K-means Method with Local Optimality Guarantees
  Mingyi Li (University of Tokyo)
  Michael R. Metel (Huawei Noah’s Ark Lab)
  Akiko Takeda (University of Tokyo / RIKEN AIP)
• Nonlinear transformers can perform inference-time feature learning: a case study of in-context learning on single-index models
  Naoki Nishikawa (The University of Tokyo / RIKEN AIP)
  Yujin Song (The University of Tokyo / RIKEN AIP)
  Kazusato Oko (UC Berkeley / RIKEN AIP)
  Denny Wu (New York University / Flatiron Institute)
  Taiji Suzuki (The University of Tokyo / RIKEN AIP)
• Non-stationary Online Learning for Curved Losses: Improved Dynamic Regret via Mixability
  Yu-Jie Zhang (RIKEN AIP)
  Peng Zhao (Nanjing University)
  Masashi Sugiyama (RIKEN AIP)
• On the Role of Label Noise in the Feature Learning Process.
  Andi Han (RIKEN AIP)
  Wei Huang (RIKEN AIP)
  Zhanpeng Zhou (SJTU)
  Gang Niu (RIKEN AIP)
  Wuyang Chen (Simon Fraser University)
  Junchi Yan (SJTU)
  Akiko Takeda (University of Tokyo / RIKEN AIP)
  Taiji Suzuki (University of Tokyo / RIKEN AIP).
• PAC-Bayes Analysis for Recalibration in Classification
  Masahiro Fujisawa* (The University of Osaka / RIKEN AIP),
  Futoshi Futami* (The University of Osaka / RIKEN AIP)
  * Equal contribution
• Propagation of Chaos for Mean-Field Langevin Dynamics and its Application to Model Ensemble
  Atsushi Nitanda (A*STAR / Nanyang Technological University)
  Anzelle Lee (National University of Singapore / A*STAR)
  Damian Tan Xing Kai (Nanyang Technological University / A*STAR)
  Mizuki Sakaguchi (Kyushu Institute of Technology)
  Taiji Suzuki (The University of Tokyo / RIKEN AIP)
• Provable In-Context Vector Arithmetic via Retrieving Task Concepts.
  Dake Bu (CityUHK / RIKEN AIP)
  Wei Huang (RIKEN AIP)
  Andi Han (RIKEN AIP)
  Atsushi Nitanda (NTU /A*STAR CFAR)
  Taiji Suzuki (University of Tokyo/RIKEN AIP)
  Qingfu Zhang (CityUHK)
  Hau-San Wong (CityUHK)
• Quantifying Memory Utilization with Effective State-Size
  Rom Parnichkun (The University of Tokyo / Liquid AI)
  Neehal Tumma (Liquid AI)
  Armin W Thomas (Liquid AI)
  Alessandro Moro (The University of Tokyo)
  Qi An (The University of Tokyo)
  Taiji Suzuki (The University of Tokyo / RIKEN AIP)
  Atsushi Yamashita (The University of Tokyo)
  Michael Poli (Stanford University / Liquid AI)
  Stefano Massaroli (RIKEN AIP / Liquid AI)
• Ridgelet Transform and Unified Universality Theorem for Deep and Shallow Join-Group-Equivariant Machines
  Sho Sonoda (RIKEN AIP)
  Yuka Hashimoto (NTT Corporation / RIKEN AIP)
  Isao Ishikawa (Kyoto University / RIKEN AIP)
  Masahiro Ikeda (The University of Osaka / RIKEN AIP)
• Scalable Sobolev IPM for Probability Measures on a Graph
  Tam Le* (ISM / RIKEN AIP)
  Truyen Nguyen* (The University of Akron)
  Hideitsu Hino (ISM)
  Kenji Fukumizu (ISM)
  * equal contribution
• Self-Supervised Learning of Intertwined Content and Positional Features for Object
  Detection
  Kang-Jun Liu (Tohoku University / RIKEN AIP)
  Masanori Suganuma (Tohoku University / RIKEN AIP)
  Takayuki Okatani (Tohoku University / RIKEN AIP)
• Tensor Decomposition Based Memory-Efficient Incremental Learning
  Yuhang Li (Guangdong University of Technology)
  Guoxu Zhou (Guangdong University of Technology)
  Zhenhao Huang (Guangdong University of Technology)
  Xinqi Chen (Guangdong University of Technology)
  Yuning Qiu (RIKEN AIP)
  Qibin Zhao (RIKEN AIP)
• TreeLoRA: Efficient Continual Learning via Layer-Wise LoRAs Guided by a Hierarchical Gradient-Similarity Tree
  Yu-Yang Qian (Nanjing University)
  Yuan-Ze Xu (Nanjing University)
  Zhen-Yu Zhang (RIKEN AIP)
  Peng Zhao (Nanjing University)
  Zhi-Hua Zhou (Nanjing University)
• Tree-Sliced Wasserstein Distance: A Geometric Perspective
  Hoang V. Tran* (National University of Singapore)
  Trang Pham* (Qualcomm AI, Vietnam)
  Tho Tran (National University of Singapore)
  Khoi Nguyen (FPT, Vietnam)
  Thanh Chu (National University of Singapore)
  Tam Le** (ISM / RIKEN AIP)
  Tan M. Nguyen** (National University of Singapore)
  *equal contribution, **co-last author
• Tree-Sliced Wasserstein Distance with Nonlinear Projection
  Thanh Tran* (VinUniversity, Vietnam)
  Hoang V. Tran* (National University of Singapore)
  Thanh Chu (National University of Singapore)
  Trang Pham (Qualcomm AI, Vietnam)
  Laurent El Ghaoui** (VinUniversity, Vietnam)
  Tam Le** (ISM / RIKEN AIP)
  Tan M. Nguyen** (National University of Singapore)
  * equal contribution, ** co-last author

+Past student Trainee of RIKEN AIP
++ Past interns of RIKEN AIP

Related Laboratories