arXivDaily arXiv每日学术速递 周一至周五更新

高校专区

University of Texas at Austin(得克萨斯大学奥斯汀分校)

2026-01-28 至 2026-01-28 共收录 6
2601.19202 2026-01-28 cs.CL

Do Images Speak Louder than Words? Investigating the Effect of Textual Misinformation in VLMs

图像胜过言语吗?探讨文本误导在VLMs中的影响

Chi Zhang, Wenxuan Ding, Jiale Liu, Mingrui Wu, Qingyun Wu, Ray Mooney

机构 * The University of Texas at Austin(德克萨斯大学奥斯汀分校) Pennsylvania State University(宾夕法尼亚州立大学) New York University(纽约大学) University of Chinese Academy of Sciences(中国科学院大学) AG2ai, Inc.(AG2ai公司)

AI总结 研究探讨了文本误导对视觉-语言模型(VLMs)的影响,发现模型易受误导性文本提示影响,导致性能显著下降。

Comments 24 pages, 10 figures. Accepted at EACL 2026 (main conference)

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.17756 2026-01-28 cs.CV cs.AI cs.GR

MV-S2V: Multi-View Subject-Consistent Video Generation

MV-S2V:多视图主体一致视频生成

Ziyang Song, Xinyu Gong, Bangya Liu, Zelin Zhao

机构 * The Hong Kong Polytechnic University(香港理工大学) The University of Texas at Austin(德克萨斯大学) University of Wisconsin-Madison(威斯康星大学麦迪逊分校) Georgia Institute of Technology(佐治亚理工学院)

AI总结 本文提出MV-S2V方法,通过多视角参考生成一致的3D主体视频,解决单视角限制并提升视频生成质量。

Comments 13 pages, 9 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.20772 2026-01-28 cs.AI cs.LG

Damper-B-PINN: Damper Characteristics-Based Bayesian Physics-Informed Neural Network for Vehicle State Estimation

基于阻尼特性的人工神经网络:用于车辆状态估计的贝叶斯物理信息神经网络

Tianyi Zeng, Tianyi Wang, Zimo Zeng, Feiyang Zhang, Jiseop Byeon, Yujin Wang, Yajie Zou, Yangyang Wang, Junfeng Jiao, Christian Claudel, Xinbo Chen

机构 * School of Automation and Intelligent Sensing, Shanghai Jiao Tong University, Shanghai 200240, China(上海交通大学自动化与智能感知学院) Department of Civil, Architectural, and Environmental Engineering, The University of Texas at Austin(德克萨斯大学奥斯汀分校土木、建筑与环境工程系) College of Electrical Engineering, Zhejiang University(浙江大学电气工程学院) School of Automotive Studies, Tongji University(同济大学汽车学院) Key Laboratory of Road and Traffic Engineering of Ministry of Education, Tongji University(同济大学交通工程教育部长三角交通工程重点实验室) School of Architecture, The University of Texas at Austin(德克萨斯大学奥斯汀分校建筑系)

AI总结 本文提出基于阻尼特性的贝叶斯物理信息神经网络,用于提高车辆动态车轮载荷估计的精度和鲁棒性。

Comments 8 pages, 8 figures, 3 tables, accepted for IEEE Intelligent Vehicles (IV) Symposium 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.14780 2026-01-28 cs.CL cs.AI cs.CV

ReVision: A Dataset and Baseline VLM for Privacy-Preserving Task-Oriented Visual Instruction Rewriting

ReVision:一个用于隐私保护任务导向视觉指令重写的数据集和基线VLM

Abhijit Mishra, Mingda Li, Hsiang Fu, Richard Noh, Minji Kim

机构 * School of Information(信息学院) The University of Texas at Austin(德克萨斯大学奥斯汀分校) Department of Statistics and Data Science(统计与数据科学系) Yale University(耶鲁大学) School of Computing and Augmented Intelligence(计算与增强智能学院)

AI总结 ReVision提出一个数据集和基线VLM,通过将多模态指令转换为纯文本命令,实现轻量级设备端指令重写,提升隐私保护的多模态AI应用能力。

Comments In Proceedings of the IJCNLP-AACL 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.09851 2026-01-28 cs.CV cs.AI cs.HC

ViSIL: Unified Evaluation of Information Loss in Multimodal Video Captioning

ViSIL:多模态视频描述信息损失的统一评估

Po-han Li, Shenghui Chen, Ufuk Topcu, Sandeep Chinchali

机构 * The University of Texas at Austin, Texas, USA(德克萨斯大学奥斯汀分校)

AI总结 ViSIL通过信息论框架量化多模态视频摘要的信息损失,实现跨格式的统一评估,并在VQA任务中提升准确率7%。

详情

展开后加载摘要…

URL PDF HTML 收藏
2507.02834 2026-01-28 cs.LG cs.CL

ExPO: Unlocking Hard Reasoning with Self-Explanation-Guided Reinforcement Learning

ExPO: 通过自我解释引导的强化学习解锁复杂推理

Ruiyang Zhou, Shuozhe Li, Amy Zhang, Liu Leqi

机构 * University of Texas at Austin(德克萨斯大学奥斯汀分校)

AI总结 ExPO通过自我解释引导强化学习,提升模型在复杂推理任务中的学习效率和性能。

Comments Accepted to NeurIPS 2025 (Poster). Code available at https://github.com/HumainLab/ExPO_rl_reasoning_by_explanation

详情

展开后加载摘要…

URL PDF HTML 收藏