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

The Chinese University of Hong Kong(香港中文大学)

2026-03-04 至 2026-03-04 共收录 11
2603.03198 2026-03-04 cs.RO cs.CL cs.CV

ACE-Brain-0: Spatial Intelligence as a Shared Scaffold for Universal Embodiments

ACE-Brain-0:空间智能作为通用具身化体系的共享框架

Ziyang Gong, Zehang Luo, Anke Tang, Zhe Liu, Shi Fu, Zhi Hou, Ganlin Yang, Weiyun Wang, Xiaofeng Wang, Jianbo Liu, Gen Luo, Haolan Kang, Shuang Luo, Yue Zhou, Yong Luo, Li Shen, Xiaosong Jia, Yao Mu, Xue Yang, Chunxiao Liu, Junchi Yan, Hengshuang Zhao, Dacheng Tao, Xiaogang Wang

机构 * Shanghai Jiao Tong University(上海交通大学) Nanyang Technological University(南洋理工大学) The Chinese University of Hong Kong(香港中文大学) The University of Hong Kong(香港大学) University of Science(科学技术大学) Fudan University(复旦大学) Xiamen University(厦门大学) East China Normal University(华东师范大学) Wuhan University(武汉大学) Sun Yat-sen University(中山大学)

AI总结 ACE-Brain-0 通过空间智能作为共享框架,统一了自动驾驶、机器人和 UAVs 的具身化任务,采用 SSR 范式和 GRPO 方法实现跨领域泛化和领域精通的平衡。

Comments Code: https://github.com/ACE-BRAIN-Team/ACE-Brain-0 Hugging Face: https://huggingface.co/ACE-Brain/ACE-Brain-0-8B

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.02985 2026-03-04 cs.CV

The Dresden Dataset for 4D Reconstruction of Non-Rigid Abdominal Surgical Scenes

德里森数据集用于非刚性腹部手术场景的4D重建

Reuben Docea, Rayan Younis, Yonghao Long, Maxime Fleury, Jinjing Xu, Chenyang Li, André Schulze, Ann Wierick, Johannes Bender, Micha Pfeiffer, Qi Dou, Martin Wagner, Stefanie Speidel

机构 * Department of Translational Surgical Oncology, National Center for Tumor Diseases (NCT), NCT/UCC Dresden, a partnership between DKFZ, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, and Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Dresden, Germany(转化外科肿瘤中心部门,肿瘤疾病国家中心(NCT),NCT/UCC德累斯顿,由DKFZ、医学院和卡尔·戈斯瓦尔德·卡尔医院、德累斯顿技术大学以及德累斯顿-罗斯恩多夫赫尔姆霍兹中心(HZDR)组成的合作机构,德累斯顿,德国) Department of Translational Surgical Oncology, NCT/UCC Dresden, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology(转化外科肿瘤中心部门,NCT/UCC德累斯顿,医学院和卡尔·戈斯瓦尔德·卡尔医院,德累斯顿技术大学) Department for Visceral, Thoracic and Vascular Surgery, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden(visceral、胸腔和血管外科部门,医学院和卡尔·戈斯瓦尔德·卡尔医院,德累斯顿技术大学) Centre for Tactile Internet with Human-in-the-Loop (CeTI), Technical University Dresden(人机交互触觉互联网中心(CeTI),德累斯顿技术大学) Department of Computer Science and Engineering, The Chinese University of Hong Kong(计算机科学与工程部门,香港中文大学)

AI总结 D4D数据集通过提供高精度结构光几何与内窥镜视频,为非刚性腹部手术场景的4D重建和深度估计方法提供全面评估基准。

Comments 16 pages, 10 figures, accompanying data descriptor for dataset, submitted to Scientific Data

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.02907 2026-03-04 cs.CV

Harmonic Beltrami Signature Network: a Shape Prior Module in Deep Learning Framework

调和Beltrami签名网络:深度学习框架中的形状先验模块

Chenran Lin, Lok Ming Lui

机构 * Department of Mathematics, The Chinese University of Hong Kong(香港中文大学数学系)

AI总结 本文提出HBSN,一种用于提取形状先验信息的深度学习模块,能有效提升分割模型性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.05116 2026-03-04 cs.LG cs.CV

Value Gradient Guidance for Flow Matching Alignment

价值梯度指导下的流匹配对齐

Zhen Liu, Tim Z. Xiao, Carles Domingo-Enrich, Weiyang Liu, Dinghuai Zhang

机构 * The Chinese University of Hong Kong (Shenzhen)(香港中文大学(深圳)) University of Tübingen(图宾根大学) Microsoft Research(微软研究院) The Chinese University of Hong Kong(香港中文大学) Mila – Quebec AI Institute(魁北克AI研究所)

AI总结 VGG-Flow通过价值梯度指导实现流匹配模型的高效微调与先验保留对齐。

Comments Accepted at NeurIPS 2025; 26 pages, 20 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2510.12264 2026-03-04 cs.AI

Reducing Belief Deviation in Reinforcement Learning for Active Reasoning

减少强化学习中的信念偏差以实现主动推理

Deyu Zou, Yongqiang Chen, Jianxiang Wang, Haochen Yang, Mufei Li, James Cheng, Pan Li, Yu Gong

机构 * The Chinese University of Hong Kong(香港中文大学) ByteDance(字节跳动) Georgia Institute of Technology(佐治亚理工学院)

AI总结 本文提出$\mathbf{T^3}$方法,通过跟踪信念偏差并截断训练轨迹,提升LLM在主动推理中的训练稳定性与性能表现。

Comments Published as a conference paper at ICLR 2026 (Oral)

详情

展开后加载摘要…

URL PDF HTML 收藏
2502.16894 2026-03-04 cs.CL

Make LoRA Great Again: Boosting LoRA with Adaptive Singular Values and Mixture-of-Experts Optimization Alignment

让LoRA更出色:通过自适应奇异值和专家混合优化对齐提升LoRA

Chenghao Fan, Zhenyi Lu, Sichen Liu, Chengfeng Gu, Xiaoye Qu, Wei Wei, Yu Cheng

机构 * School of Computer Science \& Technology, Huazhong University of Science The Chinese University of Hong Kong Zhejiang University

AI总结 GOAT通过自适应奇异值和专家混合优化对齐提升LoRA性能,实现在多个任务上的最优表现。

Comments Accepted by ICML 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.02615 2026-03-04 cs.CL

Think, But Don't Overthink: Reproducing Recursive Language Models

思考,但不要过度思考:重现递归语言模型

Daren Wang

机构 * The Chinese University of Hong Kong(香港中文大学)

AI总结 本研究通过评估不同递归深度对RLM性能的影响,发现更深层次递归会导致模型过度思考,从而降低性能并增加计算成本。

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.02576 2026-03-04 cs.LG

Wasserstein Proximal Policy Gradient

基于Wasserstein几何的近端策略梯度方法

Zhaoyu Zhu, Shuhan Zhang, Rui Gao, Shuang Li

机构 * Zhiyuan College, Shanghai Jiao Tong University, Shanghai 200240, China(上海交通大学紫阳学院) School of Data Science, The Chinese University of Hong Kong, Shenzhen, Guangdong, China(香港中文大学(深圳)数据科学学院) McCombs School of Business, The University of Texas at Austin, Austin, TX, USA(德克萨斯大学奥斯汀分校麦克拉姆商学院)

AI总结 WPPG通过Wasserstein几何视角提出一种无需计算策略对数密度或梯度的策略梯度方法,实现连续动作空间下的高效强化学习

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.00976 2026-03-04 cs.CV

PreciseCache: Precise Feature Caching for Efficient and High-fidelity Video Generation

PreciseCache: 用于高效且高保真的视频生成的精确特征缓存

Jiangshan Wang, Kang Zhao, Jiayi Guo, Jiayu Wang, Hang Guo, Chenyang Zhu, Xiu Li, Xiangyu Yue

机构 * MMLab, CUHK(香港中文大学MMLab) Tsinghua University(清华大学) Tongyi Lab, Alibaba(阿里巴巴通义实验室)

AI总结 PreciseCache通过精确检测和跳过冗余计算,实现视频生成的高效且高质量推理。

Comments ICLR 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.18833 2026-03-04 cs.SD cs.CV eess.AS eess.IV

PrismAudio: Decomposed Chain-of-Thoughts and Multi-dimensional Rewards for Video-to-Audio Generation

PrismAudio:分解的思维链与多维奖励用于视频到音频生成

Huadai Liu, Kaicheng Luo, Wen Wang, Qian Chen, Peiwen Sun, Rongjie Huang, Xiangang Li, Jieping Ye, Wei Xue

机构 * Hong Kong University of Science and Technology (HKUST)(香港科技大学) Tongyi Fun Team, Alibaba Group(通义Fun团队,阿里巴巴集团) The Chinese University of Hong Kong (CUHK)(香港中文大学)

AI总结 PrismAudio通过分解思维链与多维奖励,解决视频到音频生成中的目标纠缠问题,实现更高质量的生成效果。

Comments ICLR 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2510.03215 2026-03-04 cs.CL cs.LG

Cache-to-Cache: Direct Semantic Communication Between Large Language Models

缓存到缓存:大型语言模型之间的直接语义通信

Tianyu Fu, Zihan Min, Hanling Zhang, Jichao Yan, Guohao Dai, Wanli Ouyang, Yu Wang

机构 * Tsinghua University(清华大学) Infinigence AI The Chinese University of Hong Kong(香港中文大学) Shanghai Jiao Tong University(上海交通大学) SLAI Shanghai AI Laboratory(上海人工智能实验室)

AI总结 缓存到缓存通过直接语义通信提升大型语言模型的性能和效率,实现比文本通信更高的准确率和更低的延迟。

Comments Published in ICLR'26

详情

展开后加载摘要…

URL PDF HTML 收藏