arXivDaily arXiv每日学术速递 周一至周五更新

期刊&会议

International Conference on Machine Learning · 会议 · Machine Learning

2026-01-21 至 2026-01-21 共收录 12
2601.14238 2026-01-21 cs.LG

Spatiotemporal Wildfire Prediction and Reinforcement Learning for Helitack Suppression

时空野火预测与直升机灭火强化学习

Shaurya Mathur, Shreyas Bellary Manjunath, Nitin Kulkarni, Alina Vereshchaka

机构 * Department of Computer Science(计算机科学系) Engineering University at Buffalo Buffalo, New York, USA(布法罗大学工程学院)

AI总结 FireCastRL结合深度学习和强化学习,通过时空预测和物理模拟实现主动野火预测与灭火策略优化。

Comments 6 pages, 5 figures (two of them in tables), Conference: IEEE International Conference on Machine Learning and Applications 2025 (ICMLA 2025): https://www.icmla-conference.org/icmla25/

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.14228 2026-01-21 cs.LG

Attention-Based Offline Reinforcement Learning and Clustering for Interpretable Sepsis Treatment

基于注意力机制的离线强化学习与聚类用于可解释的脓毒症治疗

Punit Kumar, Vaibhav Saran, Divyesh Patel, Nitin Kulkarni, Alina Vereshchaka

机构 * Department of Computer Science(计算机科学系) Engineering University at Buffalo Buffalo, New York, USA(布法罗大学工程学院)

AI总结 本文提出基于注意力机制的离线强化学习与聚类方法,用于可解释的脓毒症治疗决策支持,通过多模块整合提升治疗准确性和可解释性。

Comments 8 pages, 6 figures, Conference: IEEE International Conference on Machine Learning and Applications 2025 (ICMLA 2025): https://www.icmla-conference.org/icmla25/

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.14208 2026-01-21 cs.CV cs.GR cs.LG

Rig-Aware 3D Reconstruction of Vehicle Undercarriages using Gaussian Splatting

基于相机 rig 的车辆底盘 3D 重建使用高斯溅射

Nitin Kulkarni, Akhil Devarashetti, Charlie Cluss, Livio Forte, Dan Buckmaster, Philip Schneider, Chunming Qiao, Alina Vereshchaka

机构 * University at Buffalo(布法罗大学)

AI总结 提出一种基于相机 rig 的 3D 重建方法,利用高斯溅射生成逼真的车辆底盘模型,提升检查效率和买家信任度。

Comments 8 pages, 9 figures, Conference: IEEE International Conference on Machine Learning and Applications 2025 (ICMLA 2025): https://www.icmla-conference.org/icmla25/

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.13776 2026-01-21 cs.LG stat.ML

Orthogonium : A Unified, Efficient Library of Orthogonal and 1-Lipschitz Building Blocks

正交性:统一、高效的正交和1-利普希茨构建块库

Thibaut Boissin, Franck Mamalet, Valentin Lafargue, Mathieu Serrurier

机构 * Natural Intelligence Toulouse Institute, France(法国图卢兹自然智能研究所) INRIA, Bordeaux, France(法国波尔多国家信息与自动化研究所)

AI总结 Orthogonium 是一个统一高效的PyTorch库,提供正交和1-利普希茨层,以提升深度学习模型的鲁棒性和稳定性。

Journal ref ICML 2025 Workshop on Championing Open- source Development in Machine Learning (CODEML '25), Jul 2025, Vancouver, France

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.13645 2026-01-21 cs.LG cs.AI cs.CV

Quadratic Upper Bound for Boosting Robustness

二次上界用于提升鲁棒性

Euijin You, Hyang-Won Lee

机构 * Department of Computer Science and Engineering, Konkuk University, Seoul, South Korea(计算机科学与工程系,konkuk大学,首尔,韩国)

AI总结 本文提出二次上界损失函数,用于提升对抗训练的鲁棒性,实验表明该方法能有效改善模型的鲁棒性。

Comments Accepted at ICML 2025. Published in PMLR 267:72656-72676

Journal ref Proceedings of the 42nd International Conference on Machine Learning (ICML 2025), Proceedings of Machine Learning Research (PMLR), vol. 267, pp. 72656-72676, 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2305.14543 2026-01-21 stat.ML cs.LG

Deep Functional Factor Models: Forecasting High-Dimensional Functional Time Series via Bayesian Nonparametric Factorization

深度函数因子模型:通过贝叶斯非参数因子化进行高维函数时间序列的预测

Yirui Liu, Xinghao Qiao, Yulong Pei, Liying Wang

机构 * London School of Economics(伦敦经济学院) Faculty of Business(商学院) Economics, The University of Hong Kong(经济学,香港大学) Management School, University of Liverpool(利物浦大学管理学院)

AI总结 本文提出DF2M模型,通过贝叶斯非参数因子化方法提升高维函数时间序列预测的可解释性和准确性。

Journal ref Proceedings of the 41st International Conference on Machine Learning 2024

详情

展开后加载摘要…

URL PDF HTML 收藏
2412.04134 2026-01-21 cs.LG

M2PDE: Compositional Generative Multiphysics and Multi-component PDE Simulation

M2PDE:基于扩散模型的组合生成多物理场与多组分PDE仿真

Tao Zhang, Zhenhai Liu, Feipeng Qi, Yongjun Jiao, Tailin Wu

机构 * State Key Laboratory of Advanced Nuclear Energy Technology, Nuclear Power Institute of China, China(先进核能技术国家重点实验室,中国核工业研究院) Department of Artificial Intelligence, Westlake University, China(人工智能学院,西湖大学)

AI总结 M2PDE通过基于扩散模型的生成方法,实现多物理场与多组分PDE仿真的高效准确模拟。

Comments 29pages,14 figures

Journal ref Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:75638-75666, 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.12008 2026-01-21 cs.LG

Extreme Value Policy Optimization for Safe Reinforcement Learning

极值政策优化用于安全强化学习

Shiqing Gao, Yihang Zhou, Shuai Shao, Haoyu Luo, Yiheng Bing, Jiaxin Ding, Luoyi Fu, Xinbing Wang

机构 * Shanghai Jiao Tong University, Shanghai, China.(上海交通大学)

AI总结 本文提出极值政策优化算法,通过极值理论建模和极值优先机制,有效减少约束违反并提升安全强化学习性能。

Comments Published in the 42nd International Conference on Machine Learning (ICML 2025)

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.11953 2026-01-21 cs.LG

Controlling Underestimation Bias in Constrained Reinforcement Learning for Safe Exploration

在安全探索中约束强化学习中的低估偏差控制

Shiqing Gao, Jiaxin Ding, Luoyi Fu, Xinbing Wang

机构 * Shanghai Jiao Tong University, Shanghai, China(上海交通大学)

AI总结 本文提出MICE方法,通过引入内在成本和偏差校正策略,有效控制约束强化学习中的低估偏差,减少约束违反并保持策略性能。

Comments Published in the 42nd International Conference on Machine Learning (ICML 2025, Oral)

详情

展开后加载摘要…

URL PDF HTML 收藏
2510.00332 2026-01-21 cs.AI cs.CE

When Hallucination Costs Millions: Benchmarking AI Agents in High-Stakes Adversarial Financial Markets

当幻觉成本百万:在高风险对抗性金融市场中基准测试AI代理

Zeshi Dai, Zimo Peng, Zerui Cheng, Ryan Yihe Li

机构 * Surf AI, Cybertino Lab(Surf AI,Cybertino 实验室) Princeton University(普林斯顿大学)

AI总结 CAIA基准测试揭示了AI在对抗性金融市场中的能力缺口,指出当前模型在面对虚假信息和不可逆决策时表现不佳,强调对抗鲁棒性对可信AI的重要性。

Comments 15 pages, 5 figures, 4 tables; Accepted to AAAI 2026 (AI-4-Finance Workshop - Oral, top 10%); In submission to ICML 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.18462 2026-01-21 cs.LG cs.AI

Scalable Equilibrium Sampling with Sequential Boltzmann Generators

可扩展的平衡采样与顺序玻尔兹曼生成器

Charlie B. Tan, Avishek Joey Bose, Chen Lin, Leon Klein, Michael M. Bronstein, Alexander Tong

AI总结 本文提出顺序玻尔兹曼生成器,通过高效归一化流和连续时间序列蒙特卡洛方法,在笛卡尔坐标上实现复杂肽系统的平衡采样。

Comments Presented at ICML 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2410.01643 2026-01-21 cs.LG cs.AI

Stable Offline Value Function Learning with Bisimulation-based Representations

基于仿射关系的稳定离线价值函数学习

Brahma S. Pavse, Yudong Chen, Qiaomin Xie, Josiah P. Hanna

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

AI总结 本文提出基于仿射关系的核表示算法,通过塑造状态-动作表示来稳定离线价值函数学习,提升评估的稳定性和准确性。

Comments Accepted at the International Conference on Machine Learning (ICML) 2025

Journal ref ICML 2025

详情

展开后加载摘要…

URL PDF HTML 收藏