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

2026-08-04 至 2026-08-04 共收录 5
2608.01428 2026-08-04 cs.RO cs.AI cs.LG 新提交

When Replanning Becomes the Bottleneck: Budgeted Replanning for Embodied Agents

当重规划成为瓶颈:具身智能体的预算型重规划

Shuaijun Liu, Feiyang You, Xingwei Chen, Ningxin Su

AI总结 针对具身智能体重规划延迟超标的瓶颈问题,提出搭载E-RECAP渐进式token剪枝方法的BRACE控制器,通过预算控制循环减少token消耗、降低SLO违反率,在多平台任务中提升了重规划的实时性与成功率。

Comments 20 pages total: 9 pages main text, 3 pages references, and 8 pages appendix; 18 figures and 32 tables. Accepted at ICML 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.01400 2026-08-04 cs.LG cs.AI 新提交

TabDPT-Turbo: Efficient In-Context Learning for Tabular Prediction

TabDPT-Turbo:面向表格预测的高效上下文学习方法

Rasa Hosseinzadeh, Alex Labach, Zexin Xue, Shuyi Han, Valentin Thomas, Anthony L. Caterini

AI总结 本文提出TabDPT-Turbo模型,通过结合基于行的注意力、长上下文预训练及SSL预训练,在保持与TabDPT v1.1相当性能的同时大幅提升推理速度,是当前最快的表格预测基础模型。

Comments Presented as a poster at the non-archival ICML workshop on Foundation Models for Structured Data (FMSD)

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.01287 2026-08-04 cs.DS cs.LG 新提交

Active Regression for Single-Index Models with Unknown Link Functions

未知链接函数下单索引模型的主动回归

Chansophea Wathanak In, Yi Li, Wai Ming Tai, Xuan Wu

AI总结 本文针对一般p≥1下未知链接函数的单索引模型主动ℓₚ回归问题,提出非自适应采样算法并建立几乎紧下界,缩小了该领域的理论差距。

Comments Earlier version accepted to ICML 2026; the lower bound has been extended to adaptive queries

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.00540 2026-08-04 cs.CV 新提交

DiffuseAgent-MI: Distributionally-Grounded,Tool-Integrated Self-Evolving Agents for Faithful Visual Reasoning

DiffuseAgent-MI:用于忠实视觉推理的基于分布、集成工具的自进化智能体

An Lanji, Dawei Liu, Jin Li, Haoran Xu, Mei Chen, Yu Tian

AI总结 DiffuseAgent-MI是基于分布、集成工具的自进化视觉推理智能体,通过KL最小能量模型与轨迹级验证器提升忠实性,在多数据集上准确率及相关指标表现优于现有模型。

Comments 11 pages, 9 figures, accepted by ICML 2026 manitrack

详情

展开后加载摘要…

URL PDF HTML 收藏
2608.01582 2026-08-04 cond-mat.stat-mech cond-mat.dis-nn cs.LG math-ph math.MP 新提交

LieStoNet: Learning Lie Symmetries from Spatiotemporal Data for Stochastic Dynamical Systems

LieStoNet:从时空数据中学习随机动力系统的李对称

Shida Liu, Abhishek Gupta, Sumit Sinha, L. Mahadevan

AI总结 LieStoNet是端到端无模板框架,可从时空轨迹发现SDE的李点对称性,在已知解析对称的典型SDE上恢复出与真实对称代数一致的生成元,为含噪动力学提供可解释对称性发现方案。

Comments 25 Pages, 7 figures. Accepted to the International Conference on Machine Learning (ICML 2026)

详情

展开后加载摘要…

URL PDF HTML 收藏