arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~

高校专区

University of Pennsylvania(宾夕法尼亚大学)

2026-03-04 至 2026-03-04 共收录 5
2603.03278 2026-03-04 cs.RO cs.AI cs.CV

Tether: Autonomous Functional Play with Correspondence-Driven Trajectory Warping

Tether: 基于对应驱动轨迹扭曲的自主功能性玩耍

William Liang, Sam Wang, Hung-Ju Wang, Osbert Bastani, Yecheng Jason Ma, Dinesh Jayaraman

机构 * University of Pennsylvania(宾夕法尼亚大学) University of California, Berkeley(加州大学伯克利分校) Dyna Robotics(Dyna机器人技术)

AI总结 Tether通过基于对应驱动的轨迹扭曲,实现从少量演示开始的自主多任务玩耍,生成高质量数据集并提升模仿策略性能。

Comments International Conference on Learning Representations (ICLR), 2026. Project website and code: https://tether-research.github.io

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.03265 2026-03-04 cs.CV

DuoMo: Dual Motion Diffusion for World-Space Human Reconstruction

DuoMo:用于世界空间人体重建的双动差分

Yufu Wang, Evonne Ng, Soyong Shin, Rawal Khirodkar, Yuan Dong, Zhaoen Su, Jinhyung Park, Kris Kitani, Alexander Richard, Fabian Prada, Michael Zollhofer

机构 * Meta Reality Labs University of Pennsylvania(宾夕法尼亚大学) Carnegie Mellon University(卡内基梅隆大学)

AI总结 DuoMo通过双扩散模型实现世界空间人体运动重建,有效处理噪声和不完整输入,提升重建精度。

Comments CVPR 2026. Project page: https://yufu-wang.github.io/duomo/

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.08666 2026-03-04 cs.CL cs.AI

Hallucination, Monofacts, and Miscalibration: An Empirical Investigation

幻觉、单事实与校准偏差:一项实证研究

Miranda Muqing Miao, Michael Kearns

机构 * University of Pennsylvania(宾夕法尼亚大学) Department of Computer and Information Science(计算机与信息科学系)

AI总结 本文通过实验证明,选择性加权可有效降低大型语言模型的幻觉,同时保持准确性,揭示了优化目标间的内在矛盾。

Comments Code available at https://github.com/mmiao2/Hallucination.git

详情

展开后加载摘要…

URL PDF HTML 收藏
2501.12477 2026-03-04 eess.IV cs.CV

Slot-BERT: Self-supervised Object Discovery in Surgical Video

Slot-BERT: 在手术视频中实现自监督的对象发现

Guiqiu Liao, Matjaz Jogan, Marcel Hussing, Kenta Nakahashi, Kazuhiro Yasufuku, Amin Madani, Eric Eaton, Daniel A. Hashimoto

机构 * Outcomes Laboratory, Department of Surgery, University of Pennsylvania, Philadelphia, PA, USA. Department of Computer Information Science, University of Pennsylvania, Philadelphia, PA, USA. Division of Thoracic Surgery, Toronto General Hospital, University Health Network, Toronto, Ontario, Canada. Surgical Artificial Intelligence Research Academy, University Health Network, Toronto, ON, Canada.

AI总结 Slot-BERT通过双向长距离模型在手术视频中实现自监督的对象发现,提升时间一致性和跨领域适应能力。

Comments Accepted to Medical Image Analysis Journal, 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2302.01976 2026-03-04 cs.LG

SPARLING: Learning Latent Representations with Extremely Sparse Activations

SPARLING:通过极稀疏激活学习潜在表示

Kavi Gupta, Osbert Bastani, Armando Solar-Lezama

机构 * Department of Electrical Engineering and Computer Science(电气工程与计算机科学系) Massachusetts Institute of Technology(麻省理工学院) Department of Computer and Information Science(计算机与信息科学系) University of Pennsylvania(宾夕法尼亚大学)

AI总结 SPARLING通过极稀疏激活学习潜在表示,利用信息瓶颈技术实现高效中间状态建模,实验证明极端稀疏性对建模效果至关重要。

Comments 10 pages, 6 figures

详情

展开后加载摘要…

URL PDF HTML 收藏