arXivDaily arXiv每日学术速递 周一至周五更新

高校专区

University of Chinese Academy of Sciences(中国科学院大学)

共收录 1960
2507.19024 2026-01-13 cs.CV

A Survey of Multimodal Hallucination Evaluation and Detection

多模态幻觉评估与检测综述

Zhiyuan Chen, Yuecong Min, Jie Zhang, Bei Yan, Jiahao Wang, Xiaozhen Wang, Shiguang Shan

机构 * State Key Laboratory of AI Safety(人工智能安全国家重点实验室) Institute of Computing Technology, Chinese Academy of Sciences (CAS)(中国科学院计算技术研究所) University of Chinese Academy of Sciences(中国科学院大学) Trustworthy Technology and Engineering Laboratory(可信技术与工程实验室) Huawei(华为公司)

AI总结 本文综述了多模态幻觉评估与检测方法,分析了幻觉分类、评估基准、检测技术及未来研究方向。

Comments 40 pages, 5 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2501.09705 2026-01-13 cs.CV cs.AI cs.LG

Practical Continual Forgetting for Pre-trained Vision Models

实用的预训练视觉模型持续遗忘

Hongbo Zhao, Fei Zhu, Bolin Ni, Feng Zhu, Gaofeng Meng, Zhaoxiang Zhang

机构 * State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences(多模态人工智能系统国家重点实验室,自动化研究所,中国科学院) University of Chinese Academy of Sciences(中国科学院大学) Centre for Artificial Intelligence and Robotics, Hong Kong Institute of Science & Innovation, Chinese Academy of Sciences(人工智能与机器人中心,香港科学创新研究院,中国科学院) SenseTime Research(商汤科技研究院)

AI总结 该研究提出GS-LoRA方法,通过组稀疏正则化和原型信息监督,实现预训练视觉模型的持续遗忘,有效删除特定类别信息同时最小化对其他类别的影响。

Comments Accepted by TPAMI. arXiv admin note: substantial text overlap with arXiv:2403.11530

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.05693 2026-01-12 cs.AI

Circular Reasoning: Understanding Self-Reinforcing Loops in Large Reasoning Models

循环推理:理解大推理模型中的自增强循环

Zenghao Duan, Liang Pang, Zihao Wei, Wenbin Duan, Yuxin Tian, Shicheng Xu, Jingcheng Deng, Zhiyi Yin, Xueqi Cheng

机构 * State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences(中国科学院人工智能安全国家重点实验室) University of Chinese Academy of Sciences(中国科学院大学) People’s Public Security University of China(中国人民公安大学)

AI总结 本文提出LoopBench数据集和CUSUM算法,用于识别和预测大型推理模型中的自增强循环问题,揭示循环推理的机理并提升模型稳定性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04992 2026-01-12 cs.CL

Learning from Mistakes: Negative Reasoning Samples Enhance Out-of-Domain Generalization

从错误中学习:负推理样本增强领域外泛化

Xueyun Tian, Minghua Ma, Bingbing Xu, Nuoyan Lyu, Wei Li, Heng Dong, Zheng Chu, Yuanzhuo Wang, Huawei Shen

机构 * CAS Key Laboratory of AI Safety, Institute of Computing Technology, CAS, Beijing, China(中国科学院人工智能安全重点实验室,计算技术研究所,中国科学院,北京,中国) Harbin Institute of Technology, Harbin, China(哈尔滨工业大学,哈尔滨,中国) University of Chinese Academy of Sciences, Beijing, China(中国科学院大学,北京,中国) Tsinghua University, Beijing, China(清华大学,北京,中国)

AI总结 通过引入负例推理样本,GLOW方法有效提升了大语言模型在领域外任务中的泛化能力,通过调节损失下降和提升策略熵来减少过拟合并促进探索。

Comments Code and data are available at https://github.com/Eureka-Maggie/GLOW

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.21541 2026-01-12 cs.CV

Video Generation Models Are Good Latent Reward Models

视频生成模型是良好的潜在奖励模型

Xiaoyue Mi, Wenqing Yu, Jiesong Lian, Shibo Jie, Ruizhe Zhong, Zijun Liu, Guozhen Zhang, Zixiang Zhou, Zhiyong Xu, Yuan Zhou, Qinglin Lu, Fan Tang

机构 * University of Chinese Academy of Sciences(中国科学院大学) Tencent Hunyuan(腾讯文元) Huazhong University of Science and Technology(华中科技大学) Peking University(北京大学) Shanghai Jiao Tong University(上海交通大学) Tsinghua University(清华大学) Nanjing University(南京大学)

AI总结 本文提出PRFL框架,利用预训练视频生成模型在噪声潜在空间中进行奖励建模,实现高效去噪和降低训练成本。

详情

展开后加载摘要…

URL PDF HTML 收藏
2406.03505 2026-01-12 cs.LG cs.AI

Dynamic and Adaptive Feature Generation with LLM

动态和自适应特征生成与大语言模型

Xinhao Zhang, Jinghan Zhang, Banafsheh Rekabdar, Yuanchun Zhou, Pengfei Wang, Kunpeng Liu

机构 * Portland State University(波特兰州立大学) Computer Network Information Center, Chinese Academy of Sciences(中国科学院计算机网络信息中心) University of Chinese Academy of Sciences, Chinese Academy of Sciences(中国科学院大学)

AI总结 本文提出利用大语言模型和特征生成提示,实现动态和自适应的特征生成方法,以提高特征生成的可解释性、适用性和灵活性。

Comments Accepted by IJCAI 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.05124 2026-01-09 cs.CV

Re-Align: Structured Reasoning-guided Alignment for In-Context Image Generation and Editing

Re-Align:基于结构化推理的对齐方法用于上下文图像生成与编辑

Runze He, Yiji Cheng, Tiankai Hang, Zhimin Li, Yu Xu, Zijin Yin, Shiyi Zhang, Wenxun Dai, Penghui Du, Ao Ma, Chunyu Wang, Qinglin Lu, Jizhong Han, Jiao Dai

机构 * Hunyuan, Tencent(腾讯文库) IIE, CAS(中国科学院信息工程研究所) UCAS Project Page(中国科学技术大学)

AI总结 Re-Align通过结构化推理引导的对齐方法,提升上下文图像生成与编辑任务的性能。

Comments 13 pages, 9 figures, project page: https://github.com/hrz2000/realign

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04734 2026-01-09 cs.CV

AIVD: Adaptive Edge-Cloud Collaboration for Accurate and Efficient Industrial Visual Detection

AIVD:自适应边缘-云协作用于准确高效的工业视觉检测

Yunqing Hu, Zheming Yang, Chang Zhao, Qi Guo, Meng Gao, Pengcheng Li, Wen Ji

机构 * Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所) Institute of AI for Industries, Chinese Academy of Sciences(中国科学院工业人工智能研究所) University of Chinese Academy of Sciences(中国科学院大学)

AI总结 AIVD通过边缘-云协作提升工业视觉检测的精度与效率,采用轻量边缘检测与云MLLM协同,结合高效微调策略和动态调度算法,实现高吞吐低延迟的资源优化。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04698 2026-01-09 cs.AI cs.CL cs.LG

TourPlanner: A Competitive Consensus Framework with Constraint-Gated Reinforcement Learning for Travel Planning

TourPlanner: 一种基于约束门控强化学习的竞争共识框架用于旅行规划

Yinuo Wang, Mining Tan, Wenxiang Jiao, Xiaoxi Li, Hao Wang, Xuanyu Zhang, Yuan Lu, Weiming Dong

机构 * Xiaohongshu Inc.(小红书公司) Renmin University of China(中国人民大学) School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院) MAIS, Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所MAIS)

AI总结 TourPlanner通过多路径推理和约束门控强化学习,解决旅行规划中候选点修剪、探索能力限制和约束优化难题,实现更高效的行程生成。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04199 2026-01-09 cs.LG cs.AI cs.CL

The Forgotten Shield: Safety Grafting in Parameter-Space for Medical MLLMs

被遗忘的盾牌:参数空间中的医疗大语言模型安全性移植

Jiale Zhao, Xing Mou, Jinlin Wu, Hongyuan Yu, Mingrui Sun, Yang Shi, Xuanwu Yin, Zhen Chen, Zhen Lei, Yaohua Wang

机构 * National University of Defense Technology(国防科技大学) Multimodal Artificial Intelligence Systems (MAIS), Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所多模态人工智能系统(MAIS)) Multimedia Department, Xiaomi Inc(小米公司多媒体部门) Centre for Artificial Intelligence and Robotics, Hong Kong Institute of Science and Innovation, Chinese Academy of Sciences, Hong Kong(香港科学院人工智能与机器人中心) School of Artificial Intelligence, University of Chinese Academy of Sciences, UCAS(中国科学院大学人工智能学院)

AI总结 本文提出参数空间干预方法,通过提取原始模型的安全知识并注入目标模型,提升医疗大语言模型的安全性,同时保持医疗性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04118 2026-01-09 cs.CV

GeoReason: Aligning Thinking And Answering In Remote Sensing Vision-Language Models Via Logical Consistency Reinforcement Learning

GeoReason: 通过逻辑一致性强化学习对遥感视觉-语言模型中的思考与回答进行对齐

Wenshuai Li, Xiantai Xiang, Zixiao Wen, Guangyao Zhou, Ben Niu, Feng Wang, Lijia Huang, Qiantong Wang, Yuxin Hu

机构 * Aerospace Information Research Institute, Chinese Academy of Sciences(中国科学院航空信息研究所) Key Laboratory of Target Cognition and Application Technology, Chinese Academy of Sciences(中国科学院目标认知与应用技术重点实验室) University of Chinese Academy of Sciences(中国科学院大学)

AI总结 GeoReason通过逻辑一致性强化学习提升遥感视觉-语言模型的推理可靠性与可解释性,实现思考与决策的同步。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.03714 2026-01-09 cs.CL cs.CV

Visual Merit or Linguistic Crutch? A Close Look at DeepSeek-OCR

视觉价值还是语言依赖?对DeepSeek-OCR的深入考察

Yunhao Liang, Ruixuan Ying, Bo Li, Hong Li, Kai Yan, Qingwen Li, Min Yang, Okamoto Satoshi, Zhe Cui, Shiwen Ni

机构 * Chengdu Institute of Computer Applications, CAS(成都计算机应用研究所,中国科学院) University of Chinese Academy of Sciences(中国科学院大学) IMRAM, Tohoku University(东大理工学院IMRAM) China Tower Corporation Limited(中国塔公司) CSIS, Tohoku University(东大理工学院CSIS) National Institute for Materials Science(国家材料科学研究所) Shenzhen Institutes of Advanced Technology, CAS(深圳先进技术研究所,中国科学院) Artificial Intelligence Research Institute, Shenzhen University of Advanced Technology(深圳先进技术大学人工智能研究院)

AI总结 DeepSeek-OCR通过光学映射实现高比率视觉-文本压缩,但其性能依赖于语言先验,缺乏语言支持时性能大幅下降,揭示了当前压缩技术可能加剧长上下文瓶颈的问题。

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.04358 2026-01-09 cs.CV

MAFNet:Multi-frequency Adaptive Fusion Network for Real-time Stereo Matching

MAFNet:多频率自适应融合网络用于实时立体匹配

Ao Xu, Rujin Zhao, Xiong Xu, Boceng Huang, Yujia Jia, Hongfeng Long, Fuxuan Chen, Zilong Cao, Fangyuan Chen

机构 * College of Surveying and Geo-Informatics, Tongji University(同济大学测绘与地理信息学院) Shanghai Research Institute for Intelligent Autonomous Systems, Tongji University(上海智能自主系统研究院,同济大学) Institute of Optics and Electronics, Chinese Academy of Sciences(中国科学院光学精密工程研究所) Shanghai Integrated Innovation Center for Manned Lunar Exploration(上海载人月球探测创新中心) Shanghai Key Laboratory for Planetary Mapping and Remote Sensing for Deep Space Exploration(上海行星测绘与深空探测遥感重点实验室) University of Chinese Academy of Sciences(中国科学院大学)

AI总结 MAFNet通过多频率自适应融合机制,利用高效2D卷积实现高质量实时立体匹配,优于现有方法。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.04035 2026-01-08 cs.AI

MobileDreamer: Generative Sketch World Model for GUI Agent

MobileDreamer: 用于GUI代理的生成式草图世界模型

Yilin Cao, Yufeng Zhong, Zhixiong Zeng, Liming Zheng, Jing Huang, Haibo Qiu, Peng Shi, Wenji Mao, Wan Guanglu

机构 * State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences(多模态人工智能系统国家重点实验室,自动化研究所,中国科学院) School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院) Meituan(美团)

AI总结 MobileDreamer通过生成式草图世界模型和rollout想象策略,提升GUI代理在长周期任务中的决策能力,任务成功率提升5.25%。

详情

展开后加载摘要…

URL PDF HTML 收藏
2412.04666 2026-01-08 cs.CV

PhysDepth: Plug-and-Play Physical Refinement for Monocular Depth Estimation in Challenging Environments

PhysDepth:用于在恶劣环境中单目深度估计的即插即用物理细化

Kebin Peng, Haotang Li, Zhenyu Qi, Huashan Chen, Zi Wang, Wei Zhang, Sen He, Huanrui Yang, Qing Guo

机构 * Department of Computer Science, East Carolina University(东卡罗来纳大学计算机科学系) Department of Electrical and Computer Engineering, The University of Arizona(亚利桑那大学电气与计算机工程系) Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所) School of Cyber Security, University of Chinese Academy of Sciences(中国科学院大学网络安全学院) School of Computer and Cyber Sciences, Augusta University(奥古斯塔大学计算机与网络安全学院) VCIP, CS, Nankai University(南开大学)

AI总结 PhysDepth通过引入物理先验信息,提升了在恶劣环境中单目深度估计的鲁棒性和准确性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.03783 2026-01-08 cs.CL

HearSay Benchmark: Do Audio LLMs Leak What They Hear?

HearSay基准:音频大语言模型是否泄露所听内容?

Jin Wang, Liang Lin, Kaiwen Luo, Weiliu Wang, Yitian Chen, Moayad Aloqaily, Xuehai Tang, Zhenhong Zhou, Kun Wang, Li Sun, Qingsong Wen

机构 * XDU(北华大学) NTU(国立台湾大学) NCEPU(南京工程大学) BUPT(北京邮电大学) SHU(上海大学) UAEU(阿联酋大学) UCAS-IIE(中国科学院大学国际学院) Squirrel AI

AI总结 HearSay基准研究发现音频大语言模型通过声纹泄露隐私,揭示其固有隐私风险及安全机制不足。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.03713 2026-01-08 cs.CV

BREATH-VL: Vision-Language-Guided 6-DoF Bronchoscopy Localization via Semantic-Geometric Fusion

BREATH-VL:基于视觉-语言引导的6自由度支气管镜定位:通过语义-几何融合

Qingyao Tian, Bingyu Yang, Huai Liao, Xinyan Huang, Junyong Li, Dong Yi, Hongbin Liu

机构 * State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences(多模态人工智能系统国家重点实验室,自动化研究所) School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院) Department of Pulmonary and Critical Care Medicine, The First Affiliated Hospital of Sun Yat-sen University(中山大学附属第一医院呼吸与危重症医学科) Centre of AI and Robotics, Hong Kong Institute of Science & Innovation, Chinese Academy of Sciences(香港科学院人工智能与机器人中心) School of Biomedical Engineering and Imaging Sciences, King’s College London(伦敦国王学院生物医学工程与影像科学学院)

AI总结 BREATH-VL通过融合语义和几何信息,实现高精度的6自由度支气管镜定位,减少定位误差并提升效率。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.03577 2026-01-08 cs.LG

Variational Inference, Entropy, and Orthogonality: A Unified Theory of Mixture-of-Experts

变分推断、熵与正交性:混合专家模型的统一理论

Ye Su, Yong Liu

机构 * University of Chinese Academy of Sciences, Beijing, China(中国科学院大学)

AI总结 本文从贝叶斯和信息论视角构建混合专家模型的统一理论框架,推导路由机制为最优稀疏后验近似,并证明正交性可缩小全局最优与贪心近似间的差距。

Comments 27 pages, 3 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.03301 2026-01-08 cs.MA cs.AI

PC2P: Multi-Agent Path Finding via Personalized-Enhanced Communication and Crowd Perception

PC2P:通过个性化增强通信与人群感知进行多智能体路径寻找

Guotao Li, Shaoyun Xu, Yuexing Hao, Yang Wang, Yuhui Sun

机构 * Institute of Microelectronics of the Chinese Academy of Sciences(中国科学院微电子研究所) University of Chinese Academy of Sciences(中国科学院大学)

AI总结 PC2P通过个性化增强通信与人群感知方法,提升多智能体路径寻找在复杂环境中的协同与扩展能力。

Comments 8 pages,7 figures,3 tables,Accepted to IROS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.03261 2026-01-08 cs.CL cs.AI

DeepResearch-Slice: Bridging the Retrieval-Utilization Gap via Explicit Text Slicing

DeepResearch-Slice: 通过显式文本切片弥合检索-利用差距

Shuo Lu, Yinuo Xu, Jianjie Cheng, Lingxiao He, Meng Wang, Jian Liang

机构 * NLPR & MAIS CASIA(CASIA NLPR与MAIS) School of AI UCAS(UCAS人工智能学院) Meituan Inc.(美团公司)

AI总结 DeepResearch-Slice通过显式文本切片技术,有效解决检索与利用之间的差距问题,提升模型在嘈杂环境中的鲁棒性。

Comments Ongoing work

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.14435 2026-01-08 cs.CV cs.LG

MoTE: Mixture of Ternary Experts for Memory-efficient Large Multimodal Models

MoTE:混合三元专家用于内存高效的大型多模态模型

Hongyu Wang, Jiayu Xu, Ruiping Wang, Yan Feng, Yitao Zhai, Peng Pei, Xunliang Cai, Xilin Chen

机构 * Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences(中国科学院人工智能安全重点实验室,计算技术研究所) University of Chinese Academy of Sciences(中国科学院大学)

AI总结 MoTE通过训练更多低精度三元专家,实现内存高效的大规模多模态模型训练,提升端任务性能并降低内存需求。

Comments Work in progress

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02872 2026-01-07 cs.CL cs.AI

LongBench Pro: A More Realistic and Comprehensive Bilingual Long-Context Evaluation Benchmark

LongBench Pro: 一个更现实且全面的双语长上下文评估基准

Ziyang Chen, Xing Wu, Junlong Jia, Chaochen Gao, Qi Fu, Debing Zhang, Songlin Hu

机构 * School of Artificial Intelligence, Beihang University(北航人工智能学院) Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所) School of Cyber Security, University of Chinese Academy of Sciences(中国科学院大学网络安全学院)

AI总结 LongBench Pro通过人机协作构建,提供1500个双语长上下文样本,评估46个模型发现长上下文优化比参数扩展更有效,有效上下文长度较短且存在跨语言偏差,混合思考设计具有潜在优势。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02747 2026-01-07 cs.CV

D$^3$R-DETR: DETR with Dual-Domain Density Refinement for Tiny Object Detection in Aerial Images

D$^3$R-DETR:基于双域密度细化的DETR用于航空图像中的微小目标检测

Zixiao Wen, Zhen Yang, Xianjie Bao, Lei Zhang, Xiantai Xiang, Wenshuai Li, Yuhan Liu

机构 * Aerospace Information Research Institute, Chinese Academy of Sciences(中国科学院航空信息研究所) Key Laboratory of Technology in Geo-Spatial Information Processing and Application System, Chinese Academy of Sciences(中国科学院地理空间信息处理与应用系统重点实验室) Key Laboratory of Target Cognition and Application Technology, Chinese Academy of Sciences(中国科学院目标认知与应用技术重点实验室) School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences(中国科学院大学电子电气与通信工程学院)

AI总结 D$^3$R-DETR通过双域密度细化提升DETR模型在航空图像中微小目标检测的精度与效率

Comments This work has been submitted to the IEEE for possible publication

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02674 2026-01-07 cs.CL

Iterative Structured Pruning for Large Language Models with Multi-Domain Calibration

迭代式结构剪枝用于具有多领域校准的大语言模型

Guangxin Wu, Hao Zhang, Zhang Zhibin, Jiafeng Guo, Xueqi Cheng

机构 * Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所) University of Chinese Academy of Sciences(中国科学院大学) School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences(中国科学院大学先进交叉学科学院)

AI总结 本文提出了一种迭代式结构剪枝方法,通过多领域校准集和迭代策略实现高效模型压缩,减少计算开销和内存占用,同时保持性能稳定。

Comments 10 pages

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.24867 2026-01-07 cs.CL cs.AI

Encyclo-K: Evaluating LLMs with Dynamically Composed Knowledge Statements

Encyclo-K:通过动态组成的知识陈述评估LLM

Yiming Liang, Yizhi Li, Yantao Du, Ge Zhang, Jiayi Zhou, Yuchen Wu, Yinzhu Piao, Denghui Cao, Tong Sun, Ziniu Li, Li Du, Bo Lei, Jiaheng Liu, Chenghua Lin, Zhaoxiang Zhang, Wenhao Huang, Jiajun Zhang

机构 * School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院) Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所) Nanjing University(南京大学) The University of Manchester(曼彻斯特大学)

AI总结 Encyclo-K通过动态组合知识陈述,提出了一种评估LLM全面理解能力的基准测试框架,有效解决了现有基准测试的局限性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2403.11565 2026-01-07 math.OC cs.LG

Convergence of Decentralized Stochastic Subgradient-based Methods for Nonsmooth Nonconvex functions

非光滑非凸函数的去中心化随机子梯度法的收敛性

Siyuan Zhang, Nachuan Xiao, Xin Liu

机构 * State Key Laboratory of Scientific and Engineering Computing, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, and University of Chinese Academy of Sciences, China(国家科学工程计算重点实验室,数学系统科学学院,中国科学院,中国科学院大学)

AI总结 本文提出一个统一去中心化随机子梯度法的框架,为非光滑非凸函数的收敛性提供理论保证,并通过实验验证其在非光滑神经网络训练中的有效性。

Comments 35 pages

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02249 2026-01-06 cs.CV

SLGNet: Synergizing Structural Priors and Language-Guided Modulation for Multimodal Object Detection

SLGNet: 结合结构先验与语言引导调制的多模态目标检测

Xiantai Xiang, Guangyao Zhou, Zixiao Wen, Wenshuai Li, Ben Niu, Feng Wang, Lijia Huang, Qiantong Wang, Yuhan Liu, Zongxu Pan, Yuxin Hu

机构 * Aerospace Information Research Institute, Chinese Academy of Sciences(中国科学院航天信息研究所) Key Laboratory of Target Cognition and Application Technology, Chinese Academy of Sciences(中国科学院目标认知与应用技术重点实验室) School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences(中国科学院大学电子电气与通信工程学院) School of Software Engineering, Xi’an Jiaotong University(西安交通大学软件学院)

AI总结 SLGNet通过结合结构先验与语言引导调制,在冻结的ViT基础上实现高效多模态目标检测,提升环境适应性和检测性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02045 2026-01-06 cs.PL cs.AI cs.SE

The New Compiler Stack: A Survey on the Synergy of LLMs and Compilers

新的编译器栈:关于大型语言模型与编译器协同的综述

Shuoming Zhang, Jiacheng Zhao, Qiuchu Yu, Chunwei Xia, Zheng Wang, Xiaobing Feng, Huimin Cui

机构 * SKLP Institute of Computing Technology, CAS(计算技术研究所) University of Chinese Academy of Sciences(中国科学院大学) University of Leeds(利兹大学)

AI总结 本文综述了大型语言模型与编译器协同的新兴领域,提出了多维度分类法,揭示了LLMs在编译器开发中的三大优势,并指明了混合系统的发展路径。

Comments Accepted by CCF Transactions on High Performance Computing

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.02010 2026-01-06 q-bio.NC cs.AI cs.CL

A neural network for modeling human concept formation, understanding and communication

一种用于建模人类概念形成、理解和交流的神经网络

Liangxuan Guo, Haoyang Chen, Yang Chen, Yanchao Bi, Shan Yu

机构 * Laboratory of Brain Atlas and Brain-inspired Intelligence(脑图谱与脑启发智能实验室) Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所) School of Future Technology(未来技术学院) University of Chinese Academy of Sciences(中国科学院大学) State Key Laboratory of Brain Cognition and Brain-inspired Intelligence Technology(脑认知与脑启发智能技术国家重点实验室) School of Psychological and Cognitive Sciences & Beijing Key Laboratory of Behavior and Mental Health(心理学与认知科学学院及北京行为与心理健康重点实验室) IDG/McGovern Institute for Brain Research(IDG/麦戈文脑科学研究院) Institute for Artificial Intelligence(人工智能研究所) Key Laboratory of Machine Perception (Ministry of Education)(机器感知重点实验室)

AI总结 本文提出CATS Net,一种双模块神经网络,用于建模人类概念形成、理解和交流,通过概念抽象与任务解决模块实现跨网络知识转移。

Comments 6 main figures, 5 extended data figures and 4 supplementary figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.05753 2026-01-06 cs.AI cs.LG

A Fast Anti-Jamming Cognitive Radar Deployment Algorithm Based on Reinforcement Learning

基于强化学习的快速抗干扰认知雷达部署算法

Wencheng Cai, Xuchao Gao, Congying Han, Mingqiang Li, Tiande Guo

机构 * School of Mathematical Sciences University of Chinese Academy of Sciences(数学科学学院 中国科学院大学) Information Science Academy China Electronics Technology Group Corporation(信息科学院 中国电子科技集团)

AI总结 本文提出基于强化学习的快速抗干扰认知雷达部署算法FARDA,通过高效神经网络推理实现雷达部署速度提升7000倍,覆盖能力与进化算法相当。

详情

展开后加载摘要…

URL PDF HTML 收藏