arXivDaily arXiv每日学术速递 周一至周五更新

高校专区

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

共收录 2407
2601.11254 2026-01-19 cs.CV

FTDMamba: Frequency-Assisted Temporal Dilation Mamba for Unmanned Aerial Vehicle Video Anomaly Detection

FTDMamba:基于频率的时域扩张Mamba用于无人机视频异常检测

Cheng-Zhuang Liu, Si-Bao Chen, Qing-Ling Shu, Chris Ding, Jin Tang, Bin Luo

机构 * Information Materials and Intelligent Sensing Laboratory of Anhui Province(安徽省信息材料与智能感知实验室) Anhui Provincial Key Laboratory of Multimodal Cognitive Computation(安徽省多模态认知计算重点实验室) School of Computer Science and Technology, Anhui University(安徽大学计算机科学与技术学院) Department of Computer Science and Engineering, The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳)计算机科学与工程系)

AI总结 FTDMamba通过频率辅助的时域扩张Mamba网络,解决动态背景下无人机视频异常检测的挑战,提出两个核心模块并构建大规模数据集,实现最先进的检测性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.11002 2026-01-19 cs.CL

Redefining Machine Simultaneous Interpretation: From Incremental Translation to Human-Like Strategies

重新定义机器同声传译:从逐步翻译到人样策略

Qianen Zhang, Zeyu Yang, Satoshi Nakamura

机构 * The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳)) Nara Institute of Science and Technology(奈良研究所)

AI总结 本文提出通过扩展SiMT的操作空间,引入四种自适应操作以提升翻译质量与实时性,实验证明在语义指标和延迟方面优于现有方法。

Comments arXiv admin note: substantial text overlap with arXiv:2509.21801

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.23759 2026-01-19 eess.IV cs.CV

Spatio-Temporal Representation Decoupling and Enhancement for Federated Instrument Segmentation in Surgical Videos

联邦学习中手术视频仪器分割的时空表示解耦与增强

Zheng Fang, Xiaoming Qi, Chun-Mei Feng, Jialun Pei, Weixin Si, Yueming Jin

机构 * Department of Electrical and Computer Engineering, NUS, Singapore(新加坡国立大学电子与计算机工程系) Department of Biomedical Engineering, NUS, Singapore(新加坡国立大学生物医学工程系) Institute of High Performance Computing, A*STAR, Singapore(新加坡A*STAR高性能计算研究所) Department of Computer Science and Engineering, The Chinese University of Hong Kong, HKSAR, China(香港中文大学(深圳)计算机科学与工程系) Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China(中国科学院深圳先进技术研究所)

AI总结 本文提出FedST方法,通过解耦和增强时空表示,提升联邦学习中手术视频仪器分割的性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2503.18197 2026-01-19 cs.LG cs.AI

FROG: Fair Removal on Graphs

FROG: 图上的公平移除

Ziheng Chen, Jiali Cheng, Hadi Amiri, Kaushiki Nag, Lu Lin, Sijia Liu, Xiangguo Sun, Gabriele Tolomei

机构 * University of Massachusetts Lowell, USA(马萨诸塞大学洛约拉分校) Pennsylvania State University, USA(宾夕法尼亚州立大学) Michigan State University, USA(密歇根州立大学) The Chinese University of Hong Kong, China(香港中文大学) Sapienza University of Rome, Italy(罗马萨皮恩扎大学)

AI总结 FROG通过联合优化图结构和模型,实现图上的公平无学习,通过移除冗余边和增强边来保持公平性,并引入最坏情况评估机制提升鲁棒性。

Comments CIKM 2025; v2 fixes author list

详情

展开后加载摘要…

URL PDF HTML 收藏
1712.06080 2026-01-16 cs.CV

Spatial As Deep: Spatial CNN for Traffic Scene Understanding

空间作为深度:用于交通场景理解的空间CNN

Xingang Pan, Xiaohang Zhan, Jianping Shi, Ping Luo, Xiaogang Wang, Xiaoou Tang

机构 * The Chinese University of Hong Kong(香港中文大学) SenseTime Group Limited(时光集团)

AI总结 本文提出空间CNN用于交通场景理解,通过改进卷积结构提升车道检测性能,实现8.7%和4.6%的提升,并在挑战赛中取得第一名。

Comments Accepted to AAAI 2018

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02737 2026-01-16 cs.CV

Unveiling and Bridging the Functional Perception Gap in MLLMs: Atomic Visual Alignment and Hierarchical Evaluation via PET-Bench

揭示和弥合MLLMs中的功能感知差距:通过PET-Bench实现原子视觉对齐和分层评估

Zanting Ye, Xiaolong Niu, Xuanbin Wu, Xu Han, Shengyuan Liu, Jing Hao, Zhihao Peng, Hao Sun, Jieqin Lv, Fanghu Wang, Yanchao Huang, Hubing Wu, Yixuan Yuan, Habib Zaidi, Arman Rahmim, Yefeng Zheng, Lijun Lu

机构 * School of Biomedical Engineering, Southern Medical University(生物医学工程学院,南方医科大学) School of Biomedical Engineering, Shanghai Jiaotong University(生物医学工程学院,上海交通大学) Department of Electronic Engineering, Chinese University of Hong Kong(电子工程系,中国香港大学) Faculty of Dentistry, The University of Hong Kong(牙科学院,香港大学) Department of Nuclear Medicine, The Second Affiliated Hospital of Guangzhou University of Chinese Medicine(核医学科,广州中医药大学第二附属医院) PET Center, Department of Nuclear Medicine, Guangdong Provincial People’s Hospital, Southern Medical University(PET中心,核医学科,广东省人民医院,南方医科大学) Department of Nuclear Medicine, Nanfang Hospital, Southern Medical University(核医学科,南芳医院,南方医科大学) Division of Nuclear Medicine and Molecular Imaging, Geneva University Hospitals(核医学与分子影像学部,日内瓦大学医院) Departments of Radiology, Physics, and Biomedical Engineering, The University of British Columbia(放射学、物理和生物医学工程系,不列颠哥伦比亚大学) Medical Artificial Intelligence Laboratory, Westlake University(医学人工智能实验室,西湖大学)

AI总结 本文提出AVA方法,通过原子视觉对齐解决MLLMs在功能成像中的感知差距,提升诊断准确性14.83%。

Comments 9 pages, 6 figures, 6 tables

详情

展开后加载摘要…

URL PDF HTML 收藏
2407.07720 2026-01-15 eess.IV cs.CV

Exploiting Scale-Variant Attention for Segmenting Small Medical Objects

利用尺度变异注意力进行小医学物体分割

Wei Dai, Rui Liu, Zixuan Wu, Tianyi Wu, Min Wang, Junxian Zhou, Yixuan Yuan, Jun Liu

机构 * Centre for Robotics and Automation, City University of Hong Kong(香港城市大学机器人与自动化中心) Department of Electronic Engineering, The Chinese University of Hong Kong(香港中文大学电子工程系)

AI总结 本文提出SvANet,通过引入尺度变异注意力等模块,提升医学图像中小物体的分割精度。

Comments 14 pages, 9 figures, under review

Journal ref IEEE Transactions on Neural Networks and Learning Systems, 1-18 (2026)

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.08440 2026-01-14 cs.CV

Incentivizing Cardiologist-Like Reasoning in MLLMs for Interpretable Echocardiographic Diagnosis

激励MLLMs实现类似心内科医生的推理以实现可解释的超声心动图诊断

Yi Qin, Lehan Wang, Chenxu Zhao, Alex P. W. Lee, Xiaomeng Li

机构 * The Hong Kong University of Science and Technology(香港科学与技术大学) The Chinese University of Hong Kong(香港中文大学)

AI总结 本文提出CRT和CardiacMind,通过引入心内科医生的思维模式,提升MLLMs在超声心动图诊断中的推理能力,实现48%的诊断性能提升。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.07972 2026-01-14 cs.CL

Knowing But Not Doing: Convergent Morality and Divergent Action in LLMs

知晓而不做:在大语言模型中收敛的道德与发散的行为

Jen-tse Huang, Jiantong Qin, Xueli Qiu, Sharon Levy, Michelle R. Kaufman, Mark Dredze

机构 * Johns Hopkins University(约翰霍普金斯大学) Chinese University of Hong Kong(香港中文大学) Rutgers University(罗格斯大学)

AI总结 研究揭示了大语言模型在价值对齐训练下仍存在知识与行为之间的不一致,通过ValAct-15k数据集发现模型在情境决策上高度一致,但自我报告与实际行为关联较弱。

Comments 9 pages, 7 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.20002 2026-01-14 cs.LG

LoFT-LLM: Low-Frequency Time-Series Forecasting with Large Language Models

LoFT-LLM: 利用大语言模型进行低频时间序列预测

Jiacheng You, Jingcheng Yang, Yuhang Xie, Zhongxuan Wu, Xiucheng Li, Feng Li, Pengjie Wang, Jian Xu, Bo Zheng, Xinyang Chen

机构 * School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen)(计算机科学与技术学院,哈尔滨工业大学(深圳)) The Chinese University of Hong Kong(香港中文大学)

AI总结 LoFT-LLM通过整合大语言模型与低频学习,提升时间序列预测的准确性、鲁棒性和可解释性。

Comments This submission is withdrawn due to internal review and compliance considerations

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.10035 2026-01-14 cs.GR cs.AI

FastFLUX: Pruning FLUX with Block-wise Replacement and Sandwich Training

FastFLUX: 通过块级替换和 sandwich 训练进行 FLUX 剪枝

Fuhan Cai, Yong Guo, Jie Li, Wenbo Li, Jian Chen, Xiangzhong Fang

机构 * Shanghai Jiao Tong University(上海交通大学) Max Planck Institute for Informatics(马克斯·普朗克研究所(信息学)) South China University of Technology(华南理工大学) Chinese University of Hong Kong(香港中文大学)

AI总结 FastFLUX 通过块级替换和 sandwich 训练方法,提高 FLUX 模型的推理效率和图像质量。

Comments 14 pages

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.05280 2026-01-14 cs.CV

Unifying Appearance Codes and Bilateral Grids for Driving Scene Gaussian Splatting

统一外观代码与双侧网格用于驾驶场景高斯点漂浮

Nan Wang, Yuantao Chen, Lixing Xiao, Weiqing Xiao, Bohan Li, Zhaoxi Chen, Chongjie Ye, Shaocong Xu, Saining Zhang, Ziyang Yan, Pierre Merriaux, Lei Lei, Tianfan Xue, Hao Zhao

机构 * BAAI(北京人工智能研究院) AIR, THU(清华大学人工智能研究院) SJTU(上海交通大学) EIT(Ningbo)(宁波工程学院) CUHK(香港大学) LeddarTech

AI总结 本文提出了一种多尺度双侧网格方法,统一了外观代码与双侧网格,提升了自动驾驶场景中的几何重建精度。

Comments Accepted to NeurIPS 2025 ; Project page: https://bigcileng.github.io/bilateral-driving ; Code: https://github.com/BigCiLeng/bilateral-driving

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.07779 2026-01-13 cs.MA cs.AI cs.CL cs.CV cs.HC

OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent

OS-Symphony: 一个用于鲁棒且通用计算机使用代理的综合框架

Bowen Yang, Kaiming Jin, Zhenyu Wu, Zhaoyang Liu, Qiushi Sun, Zehao Li, JingJing Xie, Zhoumianze Liu, Fangzhi Xu, Kanzhi Cheng, Qingyun Li, Yian Wang, Yu Qiao, Zun Wang, Zichen Ding

机构 * University of Science and Technology of China(中国科学技术大学) Shanghai AI Laboratory(上海人工智能实验室) National University of Singapore(新加坡国立大学) The Hong Kong University of Science and Technology(香港科学与技术大学) The University of Hong Kong(香港大学) CUHK MMLab(香港中文大学MMLab) Xi’an Jiaotong University(西安交通大学) Nanjing University(南京大学) Harbin Institute of Technology(哈尔滨工业大学)

AI总结 OS-Symphony通过反思记忆代理和多功能工具代理,提升计算机使用代理在长周期任务和新领域中的鲁棒性和泛化能力。

Comments 31 pages, 11 figures, 12 tables

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.07411 2026-01-13 cs.LG cs.AI cs.CL

SCALPEL: Selective Capability Ablation via Low-rank Parameter Editing for Large Language Model Interpretability Analysis

SCALPEL:通过低秩参数编辑实现大语言模型可解释性分析中的选择性能力消融

Zihao Fu, Xufeng Duan, Zhenguang G. Cai

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

AI总结 SCALPEL通过低秩参数编辑实现大语言模型能力的细粒度分析与选择性消融,揭示能力在参数空间中的分布结构。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.06922 2026-01-13 cs.CL

TreePS-RAG: Tree-based Process Supervision for Reinforcement Learning in Agentic RAG

TreePS-RAG: 基于树的强化学习过程监督在代理RAG中的应用

Tianhua Zhang, Kun Li, Junan Li, Yunxiang Li, Hongyin Luo, Xixin Wu, James Glass, Helen Meng

机构 * The Chinese University of Hong Kong, Hong Kong SAR, China(香港中文大学) Massachusetts Institute of Technology, Cambridge MA, USA(麻省理工学院)

AI总结 TreePS-RAG通过基于树的在线强化学习框架,在代理RAG中实现细粒度过程监督,提升多跳和通用问答任务性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04770 2026-01-13 cs.AI cs.DB

SciIF: Benchmarking Scientific Instruction Following Towards Rigorous Scientific Intelligence

SciIF: 科学指令遵循基准测试以实现严谨的科学智能

Encheng Su, Jianyu Wu, Chen Tang, Lintao Wang, Pengze Li, Aoran Wang, Jinouwen Zhang, Yizhou Wang, Yuan Meng, Xinzhu Ma, Shixiang Tang, Houqiang Li

机构 * Shanghai AI Laboratory(上海人工智能实验室) University of Science and Technology of China(中国科学技术大学) Shanghai Jiao Tong University(上海交通大学) The Chinese University of Hong Kong(香港中文大学) University of Sydney(悉尼大学) Fudan University(复旦大学) Tsinghua University(清华大学) Beihang University(北航大学)

AI总结 SciIF是一个评估模型在科学问题解决中严格遵守约束条件能力的多学科基准测试,通过显式证据验证科学有效性,提升LLM在科学逻辑框架中的可靠性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.09195 2026-01-13 cs.CV

Towards Trustworthy Dermatology MLLMs: A Benchmark and Multimodal Evaluator for Diagnostic Narratives

迈向可信的皮肤科多模态大语言模型:一种基准和多模态评估器用于诊断叙述

Yuhao Shen, Jiahe Qian, Shuping Zhang, Zhangtianyi Chen, Tao Lu, Juexiao Zhou

机构 * School of Data Science, The Chinese University of Hong Kong, Shenzhen(数据科学学院,香港中文大学(深圳)) Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所) Department of Dermatology, The First Affiliated Hospital, Shantou University Medical College(皮肤科,汕头大学医学院第一附属医院)

AI总结 本文提出DermBench和DermEval用于评估皮肤科多模态大语言模型的诊断叙述,通过结合基准和自动评估器,实现临床意义的可重复评估,验证模型性能与专家评分的一致性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.09572 2026-01-13 cs.LG math.LO math.OC stat.ML

SAD Neural Networks: Divergent Gradient Flows and Asymptotic Optimality via o-minimal Structures

SAD神经网络:通过o-最小结构实现的分歧梯度流和渐近最优性

Julian Kranz, Davide Gallon, Steffen Dereich, Arnulf Jentzen

机构 * Department of Information Systems, University of Münster, Germany(慕尼黑大学信息系统系) Applied Mathematics: Institute for Analysis and Numerics, University of Münster, Germany(慕尼黑大学应用数学系) RiskLab Switzerland, ETH Zürich, Switzerland(苏黎世联邦理工学院瑞士风险实验室) Applied Mathematics: Institute for Mathematical Stochastics, University of Münster, Germany(慕尼黑大学应用数学系) School of Data Science and School of Artificial Intelligence, The Chinese University of Hong Kong, Shenzhen (CUHK-Shenzhen), China(香港中文大学(深圳)数据科学学院和人工智能学院)

AI总结 SAD神经网络通过o-最小结构的几何特性,证明了梯度流在特定条件下会发散到无穷大,揭示了神经网络损失优化的渐近最优性。

Comments Accepted for NeurIPS 2025, 30 pages, 6 figures. The result about continuous data distributions now has an additional assumption since a gap was identified in a previous version of the proof

详情

展开后加载摘要…

URL PDF HTML 收藏
2503.05096 2026-01-13 cs.CL

AdaSpec: Adaptive Speculative Decoding for Fast, SLO-Aware Large Language Model Serving

AdaSpec: 一种适应性推测解码用于快速、面向SLO的大型语言模型服务

Kaiyu Huang, Hao Wu, Zhubo Shi, Han Zou, Minchen Yu, Qingjiang Shi

机构 * Tongji University(同济大学) Shenzhen Research Institute of Big Data, The Chinese University of Hong Kong, Shenzhen(深圳大数据研究院,香港中文大学(深圳)) Huazhong University of Science and Technology(华中科技大学) School of Data Science, The Chinese University of Hong Kong, Shenzhen(数据科学学院,香港中文大学(深圳))

AI总结 AdaSpec通过动态调整推测策略,提高大型语言模型服务的响应速度和SLO满足度,实验显示性能提升达66%。

Comments This paper is accepted by ACM SoCC 2025

Journal ref In ACM Symposium on Cloud Computing (SoCC '25), November 19-21, 2025, Online, USA

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.06101 2026-01-13 cs.CY cs.AI

How to Assess AI Literacy: Misalignment Between Self-Reported and Objective-Based Measures

如何评估人工智能素养:自我报告与基于客观的评估之间的不一致

Shan Zhang, Ruiwei Xiao, Anthony F. Botelho, Guanze Liao, Thomas K. F. Chiu, John Stamper, Kenneth R. Koedinger

机构 * University of Florida(佛罗里达大学) Carnegie Mellon University(卡内基梅隆大学) National Tsing Hua University(国立清华大学) The Chinese University of Hong Kong(香港中文大学)

AI总结 本研究开发并评估了教师人工智能素养的自我报告和基于客观的测量方法,揭示了两者在不同教师群体中的显著差异,并提出了适用于专业发展的诊断工具。

Comments 16 pages, 6 figures, LAK2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02228 2026-01-13 cs.CV

FMVP: Masked Flow Matching for Adversarial Video Purification

FMVP: 用于对抗视频净化的掩码流匹配

Duoxun Tang, Xueyi Zhang, Chak Hin Wang, Xi Xiao, Dasen Dai, Xinhang Jiang, Wentao Shi, Rui Li, Qing Li

机构 * Tsinghua University(清华大学) The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳)) University of New South Wales(新南威尔士大学) The Chinese University of Hong Kong(香港中文大学) Peng Cheng Laboratory(鹏城实验室)

AI总结 FMVP通过掩码策略和条件流匹配技术,有效净化对抗性视频,提升鲁棒性与检测能力。

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.11323 2026-01-13 stat.ML cs.LG

Convergence Rates of Constrained Expected Improvement

带约束期望改进的收敛速率

Haowei Wang, Jingyi Wang, Zhongxiang Dai, Nai-Yuan Chiang, Szu Hui Ng, Cosmin G. Petra

机构 * National University of Singapore(新加坡国立大学) Lawrence Livermore National Laboratory(劳伦斯利弗莫尔国家实验室) The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳))

AI总结 本文研究了带约束期望改进算法的收敛速率,证明了在不同核函数和高斯过程假设下,该算法在约束优化中的收敛性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.03731 2026-01-12 cs.SE cs.AI

From Laboratory to Real-World Applications: Benchmarking Agentic Code Reasoning at the Repository Level

从实验室到现实应用:在仓库级别评估代理代码推理

Jia Li, Yuxin Su, Michael R. Lyu

机构 * The Chinese University of Hong Kong(香港中文大学) Sun Yat-sen University(中山大学)

AI总结 RepoReason提出了一种白盒诊断基准测试,通过执行驱动的突变框架和动态程序切片,量化推理能力,揭示集成宽度是代理代码推理的主要瓶颈。

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.21272 2026-01-12 cs.CV

Co-Training Vision Language Models for Remote Sensing Multi-task Learning

联合训练遥感多任务学习的视觉语言模型

Qingyun Li, Shuran Ma, Junwei Luo, Yi Yu, Yue Zhou, Fengxiang Wang, Xudong Lu, Xiaoxing Wang, Xin He, Yushi Chen, Xue Yang

机构 * Harbin Institute of Technology(哈尔滨工业大学) Shanghai Jiao Tong University(上海交通大学) Xidian University(西安电子科技大学) East China Normal University(东华大学) Wuhan University(武汉大学) Southeast University(东南大学) National University of Defense Technology(国防科技大学) Chinese University of Hong Kong(香港中文大学)

AI总结 RSCoVLM通过联合训练视觉语言模型,提升遥感多任务学习的性能和灵活性。

Comments 14 pages, 6 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.05257 2026-01-12 cs.IR cs.AI

KP-Agent: Keyword Pruning in Sponsored Search Advertising via LLM-Powered Contextual Bandits

KP-Agent:通过LLM驱动的上下文老虎机实现关键词剪枝

Hou-Wan Long, Yicheng Song, Zidong Wang, Tianshu Sun

机构 * The Chinese University of Hong Kong(香港中文大学) University of Minnesota(明尼苏达大学) Cheung Kong Graduate School of Business(长江商学院)

AI总结 KP-Agent通过LLM驱动的上下文老虎机框架,利用强化学习优化关键词剪枝,提升付费搜索广告的累计利润

详情

展开后加载摘要…

URL PDF HTML 收藏
2412.06931 2026-01-12 cs.RO

Non-Prehensile Tool-Object Manipulation by Integrating LLM-Based Planning and Manoeuvrability-Driven Controls

通过整合基于大语言模型的规划和机动性驱动的控制实现非抓取工具-物体 manipulation

Hoi-Yin Lee, Peng Zhou, Anqing Duan, Wanyu Ma, Chenguang Yang, David Navarro-Alarcon

机构 * organization= Department of Mechanical Engineering, The Hong Kong Polytechnic University , addressline= KLN, Hong Kong organization= School of Advanced Engineering, The Great Bay University , addressline= Dongguan, China organization= Department of Robotics, Mohamed Bin Zayed University of Artificial Intelligence , addressline= Abu Dhabi, UAE organization= Department of Surgery, The Chinese University of Hong Kong , addressline= NT, Hong Kong organization= Department of Computer Science, University of Liverpool , addressline= Liverpool, UK

AI总结 本文提出了一种结合大语言模型和机动性驱动控制的方法,用于实现非抓取工具-物体 manipulation任务。

详情

展开后加载摘要…

URL PDF HTML 收藏
2310.02954 2026-01-12 cs.CL

DQ-LoRe: Dual Queries with Low Rank Approximation Re-ranking for In-Context Learning

DQ-LoRe:双查询与低秩近似重排序用于上下文学习

Jing Xiong, Zixuan Li, Chuanyang Zheng, Zhijiang Guo, Yichun Yin, Enze Xie, Zhicheng Yang, Qingxing Cao, Haiming Wang, Xiongwei Han, Jing Tang, Chengming Li, Xiaodan Liang

机构 * Sun Yat-Sen University(中山大学) The Chinese University of Hong Kong(香港中文大学) Huawei Noah’s Ark Lab(华为诺亚实验室) The Hong Kong University of Science and Technology (Guangzhou)(香港科学与技术大学(广州)) MBZUAI The Hong Kong University of Science and Technology(香港科学与技术大学) Shenzhen MSU-BIT University(深圳MSU-BIT大学) DarkMatter AI Research(DarkMatter AI研究)

AI总结 DQ-LoRe通过双查询与低秩近似重排序方法提升GPT-4上下文学习性能,实现94.2%的性能提升。

Comments Accepted in ICLR 2024

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.05239 2026-01-09 cs.CV

Plenoptic Video Generation

光场视频生成

Xiao Fu, Shitao Tang, Min Shi, Xian Liu, Jinwei Gu, Ming-Yu Liu, Dahua Lin, Chen-Hsuan Lin

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

AI总结 PlenopticDreamer通过多输入单输出视频条件模型和相机引导检索策略,实现高质量多视角视频重绘,提升时空一致性和视角转换能力。

Comments Project Page: https://research.nvidia.com/labs/dir/plenopticdreamer/

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.05073 2026-01-09 cs.LG

Milestones over Outcome: Unlocking Geometric Reasoning with Sub-Goal Verifiable Reward

几何推理的里程碑:通过子目标可验证奖励解锁几何推理

Jianlong Chen, Daocheng Fu, Shengze Xu, Jiawei Chen, Yuan Feng, Yue Yang, Junchi Yan, Hongyuan Zha, Renqiu Xia

机构 * The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳)) Shanghai Jiao Tong University(上海交通大学) Fudan University(复旦大学) The Chinese University of Hong Kong(香港中文大学) University of Science and Technology Beijing(北京科技大学)

AI总结 本文提出SGVR框架,通过子目标可验证奖励提升多模态大语言模型在几何推理及泛化能力上的表现。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04973 2026-01-09 cs.AI cs.CL

ConMax: Confidence-Maximizing Compression for Efficient Chain-of-Thought Reasoning

ConMax:用于高效思维链推理的置信度最大化压缩

Minda Hu, Zexuan Qiu, Zenan Xu, Kun Li, Bo Zhou, Irwin King

机构 * The Chinese University of Hong Kong(香港中文大学) LLM Department, Tencent(腾讯机器学习部门)

AI总结 ConMax通过强化学习框架在保持推理模式的同时,高效压缩推理轨迹,提升大规模推理模型的效率与性能

详情

展开后加载摘要…

URL PDF HTML 收藏