arXivDaily arXiv每日学术速递 周一至周五更新

期刊&会议

International Conference on Learning Representations · 会议 · Machine Learning

2026-05-12 至 2026-05-12 共收录 32
2605.08202 2026-05-12 cs.LG cs.AI

Beyond Penalization: Diffusion-based Out-of-Distribution Detection and Selective Regularization in Offline Reinforcement Learning

超越惩罚:基于扩散的分布外检测与选择性正则化在离线强化学习中

Qingjun Wang, Hongtu Zhou, Hang Yu, Junqiao Zhao, Yanping Zhao, Chen Ye, Ziqiao Wang, Guang Chen

机构 * School of Computer Science and Technology, Tongji University, Shanghai, China(同济大学计算机科学与技术学院,上海,中国) MOE Key Lab of Embedded System and Service Computing, Tongji University, Shanghai, China(教育部嵌入式系统与服务计算重点实验室,同济大学,上海,中国) Shanghai Innovation Institute(上海创新研究院)

AI总结 本文提出DOSER框架,通过扩散模型捕捉行为策略和状态分布,利用单步去噪重构误差检测分布外动作,并通过评估预测转移区分有益与有害动作,提供理论保证和实验验证。

Comments 10 pages, 5 figures. Accepted to ICLR 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.08200 2026-05-12 cs.AI cs.CV cs.LG

Where Reliability Lives in Vision-Language Models: A Mechanistic Study of Attention, Hidden States, and Causal Circuits

在视觉-语言模型中可靠性在哪里存在:注意力、隐藏状态和因果回路的机制研究

Logan Mann, Ajit Saravanan, Ishan Dave, Shikhar Shiromani, Saadullah Ismail, Yi Xia, Emily Huang

机构 * UC Santa Barbara(加州大学圣巴巴拉分校) UC Berkeley(加州大学伯克利分校) NVIDIA(英伟达) Algoverse AI Research(Algoverse人工智能研究) Brown University(布朗大学)

AI总结 本文通过机制性研究发现,视觉-语言模型的可靠性主要体现在隐藏状态几何、分层边际形成和稀疏晚层回路,而非注意力图的锐度。

Comments 15 pages, 4 figures, 10 tables. Accepted at the ICLR 2026 Workshop on Multimodal Reasoning. Code and probe-training pipelines: https://github.com/itsloganmann/VLM-Reliability-Probe

详情

展开后加载摘要…

URL PDF HTML 收藏