Federated Continual Learning Goes Online: Uncertainty-Aware Memory Management for Vision Tasks and Beyond
Comments Accepted at ICLR 2025: https://openreview.net/forum?id=f65RuQgVlp
期刊&会议
ICLR
International Conference on Learning Representations · 会议 · Machine Learning
PolyNet: Learning Diverse Solution Strategies for Neural Combinatorial Optimization
机构 * Bielefeld University（比勒菲尔德大学） ； Boston University（波士顿大学）
Comments Accepted at ICLR 2025
Spatial CAPTCHA: Generatively Benchmarking Spatial Reasoning for Human-Machine Differentiation
Comments Submitted to ICLR 2026
EvoEngineer: Mastering Automated CUDA Kernel Code Evolution with Large Language Models
机构 * Department of Computer Science（计算机科学系） ； City University of Hong Kong（香港城市大学）
Comments Under Review of ICLR 2026
AgentBench: Evaluating LLMs as Agents
机构 * Tsinghua University（清华大学） ； The Ohio State University（俄亥俄州立大学） ； UC Berkeley（加州大学伯克利分校）
Comments Published in ICLR 2024
On amortizing convex conjugates for optimal transport
机构 * Meta AI
Comments ICLR 2023
Q-FSRU: Quantum-Augmented Frequency-Spectral For Medical Visual Question Answering
机构 * Amity University（阿米蒂大学）
Comments 12 pages (9 main + 2 references/appendix), 2 figures, conference paper submitted to ICLR 2026
Diffusion Posterior Sampling for General Noisy Inverse Problems
机构 * KAIST（韩国科学技术院） ； Los Alamos National Laboratory（洛斯阿拉莫斯国家实验室）
Comments ICLR 2023 spotlight
Journal ref The Eleventh International Conference on Learning Representations (ICLR) 2023
Fine-tuning with RAG for Improving LLM Learning of New Skills
Comments Under review at ICLR 2026
EMR-AGENT: Automating Cohort and Feature Extraction from EMR Databases
机构 * AITRICS ； Korea Advanced Institute of Science and Technology（韩国科学技术院）
Comments currently under submission to ICLR 2026
NeRAF: 3D Scene Infused Neural Radiance and Acoustic Fields
机构 * Center for Robotics, Mines Paris - PSL University Paris, France（机器人中心，巴黎 Mines Paris - PSL 大学）
Comments ICLR 2025 (Poster). Camera ready version. Project Page: https://amandinebtto.github.io/NeRAF; 24 pages, 13 figures
Journal ref The Thirteenth International Conference on Learning Representations, 2025
SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration
机构 * Dept. of Comp. Sci. & Tech.（计算机科学与技术系） ； Institute for AI（人工智能研究院） ； BNRist Center（BNRist中心） ； Tsinghua-Bosch Joint ML Center（清华大学-博世联合机器学习中心） ； THBI Lab（THBI实验室） ； Tsinghua University（清华大学）
Comments @inproceedings{zhang2025sageattention, title={SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration}, author={Zhang, Jintao and Wei, Jia and Zhang, Pengle and Zhu, Jun and Chen, Jianfei}, booktitle={International Conference on Learning Representations (ICLR)}, year={2025} }
Journal ref The Thirteenth International Conference on Learning Representations (ICLR 2025)
RAE: A Neural Network Dimensionality Reduction Method for Nearest Neighbors Preservation in Vector Search
机构 * University of Washington（华盛顿大学）
Comments submitted to ICLR 2026
Blueprint-Bench: Comparing spatial intelligence of LLMs, agents and image models
机构 * Andon Labs（安登实验室）
Comments 9 pages, 8 figures, submitted for ICLR 2026
InfiAgent: Self-Evolving Pyramid Agent Framework for Infinite Scenarios
机构 * The Hong Kong University（香港大学） ； The Hong Kong Polytechnic University（香港理工大学）
Comments 9 pages of main content and 32 pages of others, 2 figures, under review as a conference paper at ICLR 2026
A Model Zoo on Phase Transitions in Neural Networks
机构 * Department of Computer Science University of St. Gallen, Switzerland（斯德哥尔摩大学计算机科学系） ； Department of Computer Science Dartmouth College, USA（达特茅斯学院计算机科学系） ； Department of Electrical and Computer Engineering National University of Singapore, Singapore（新加坡国立大学电子与计算机工程系）
Comments Accepted at the ICLR Workshop on Neural Network Weights as a New Data Modality 2025; Accepted in the Journal of Data-centric Machine Learning Research (DMLR) 2025; Dataset and code accessible at https://phasetransitions.modelzoos.cc/
From An LLM Swarm To A PDDL-Empowered HIVE: Planning Self-Executed Instructions In A Multi-Modal Jungle
机构 * Huawei Technologies Ltd., UK（华为技术有限公司，英国） ； University of Edinburgh, UK（爱丁堡大学，英国）
Comments Published as a conference paper at ICLR 2025
Rare-to-Frequent: Unlocking Compositional Generation Power of Diffusion Models on Rare Concepts with LLM Guidance
机构 * KRAFTON ； Seoul National University（首尔国立大学） ； University of Wisconsin-Madison（威斯康星大学麦迪逊分校）
Comments ICLR 2025 (spotlight)
Neural Context Flows for Meta-Learning of Dynamical Systems
机构 * School of Computer Science University of Bristol（计算机科学学院英国布里斯托尔大学） ； School of Engineering Mathematics and Technology University of Bristol（工程数学与技术学院英国布里斯托尔大学）
Comments Accepted as a conference paper at ICLR 2025
AnveshanaAI: A Multimodal Platform for Adaptive AI/ML Education through Automated Question Generation and Interactive Assessment
Comments 11 pages, 12 figures. Under review as a conference paper at ICLR 2026. Preprint version posted on arXiv
Unsupervised Conformal Inference: Bootstrapping and Alignment to Control LLM Uncertainty
机构 * Department of Statistics, University of California, Davis（加州大学戴维斯分校统计系） ； Department of Applied Mathematics and Statistics, Johns Hopkins University（约翰霍普金斯大学应用数学与统计学系）
Comments 26 pages including appendix; 3 figures and 5 tables. Under review for ICLR 2026
Global Convergence in Neural ODEs: Impact of Activation Functions
机构 * DePaul University（德保罗大学） ； Iowa State University（爱荷华州立大学）
Comments ICLR 2025 (Oral)
Learning More with Less: A Dynamic Dual-Level Down-Sampling Framework for Efficient Policy Optimization
机构 * Tsinghua University（清华大学） ； WeChat, Tencent（微信、腾讯） ； University of California, Davis（加州大学戴维斯分校）
Comments 18 pages, 5 figures, Under review as a conference paper at ICLR 2026
FastGRPO: Accelerating Policy Optimization via Concurrency-aware Speculative Decoding and Online Draft Learning
机构 * Lanzhou University（兰州大学） ； The University of Hong Kong（香港大学） ； National University of Singapore（新加坡国立大学）
Comments Submitted to ICLR 2026
Function Spaces Without Kernels: Learning Compact Hilbert Space Representations
机构 * University of Texas at Austin（德克萨斯大学奥斯汀分校） ； Brigham Young University
Comments Submitted to ICLR 2026
Learning Greens Operators through Hierarchical Neural Networks Inspired by the Fast Multipole Method
机构 * Department of Computer Science（计算机科学系） ； University College London（伦敦大学学院） ； Department of Medical Physics and Biomedical Engineering（医学物理与生物医学工程系）
Comments Previously under review at ICLR 2025, originally submitted on the 12th of May 2025. The OpenReview page can be found at: http://openreview.net/forum?id=QHJb2Q7ONx
Understanding Optimization in Deep Learning with Central Flows
Comments First two authors contributed equally; author order determined by coin flip. This is the full version of a paper published at ICLR 2025. We encourage readers to explore the blog version of this paper, with animated optimization trajectories, at https://centralflows.github.io. Our code can be found at https://github.com/centralflows/centralflows
Dynamic Parameter Memory: Temporary LoRA-Enhanced LLM for Long-Sequence Emotion Recognition in Conversation
机构 * South China University of Technology, China（华南理工大学） ； MiniMax, China（MiniMax） ； The Chinese University of Hong Kong, China（香港中文大学）
Comments submitted to ICLR 2026
ColorBlindnessEval: Can Vision-Language Models Pass Color Blindness Tests?
机构 * Apply U United Kingdom（英国Apply大学）
Comments Accepted at the Open Science for Foundation Models (SCI-FM) Workshop at ICLR 2025
BRAID: Input-Driven Nonlinear Dynamical Modeling of Neural-Behavioral Data
机构 * Electrical and Computer Engineering, 2 Computer Science, 3 Biomedical Engineering（电气与计算机工程、计算机科学、生物医学工程） ； Viterbi School of Engineering, University of Southern California (USC)（维特比工程学院，南加州大学）
Comments Published at the International Conference on Learning Representations (ICLR) 2025. Code is available at GitHub https://github.com/ShanechiLab/BRAID
Journal ref ICLR 2025
