视觉大模型 / VLM

What's Missing in Vision-Language Models? Probing Their Struggles with Causal Order Reasoning

视觉-语言模型缺少什么？探究它们在因果顺序推理中的困难

Zhaotian Weng, Haoxuan Li, Xin Eric Wang, Kuan-Hao Huang, Jieyu Zhao

机构 * University of Southern California（南加州大学） ； University of California, Santa Barbara（加州大学圣芭芭拉分校） ； Texas A&M University（德克萨斯A&M大学） 纠错

专题命中 视觉问答 ：vision-language model(title,abstract)；VLM(summary_cn,abstract_cn)

AI总结 针对VLM在因果推理中的不足，提出VQA-Causal和VCR-Causal基准，发现模型在因果任务上仅略优于随机猜测，归因于训练数据缺乏因果表达，并通过难负例微调改善。

Comments 13 pages

Grounding Multi-Hop Reasoning in Structural Causal Models via Group Relative Policy Optimization

通过群体相对策略优化在结构因果模型中 grounding 多跳推理

Yunhan Bu, Quan Zhang, Huaping Zhang, Guotong Geng, Chunxiao Gao, Askar Hamdulla, Juan Wang, Qiuchi Li, Baohua Zhang, Shuai Lei, Yunbo Cao, Zhunchen Luo

机构 * School of Computer Science and Technology, Xinjiang University, Urumqi, China（新疆大学计算机科学与技术学院，乌鲁木齐，中国） ； Beijing Institute of Technology, Beijing, China（北京理工大学，北京，中国） ； Military Science Information Research Center, Academy of Military Science, Beijing, China（军事科学院军事科学信息研究中心，北京，中国） 纠错

专题命中 视觉定位与Grounding ：grounding(title,title_cn)；分类 cs.AI

AI总结 本文提出基于结构因果模型的多跳事实验证框架，通过群体相对策略优化解决推理链长度与准确率的平衡问题，实验表明其在HoVer和EX-FEVER数据集上优于现有方法。

MedP-CLIP: Medical CLIP with Region-Aware Prompt Integration

MedP-CLIP：具有区域感知提示集成的医学CLIP

Jiahui Peng, He Yao, Jingwen Li, Yanzhou Su, Sibo Ju, Yujie Lu, Jin Ye, Hongchun Lu, Xue Li, Lincheng Jiang, Min Zhu, Junlong Cheng

机构 * Sichuan University（四川大学） ； Xinjiang University（新疆大学） ； Alibaba Group（阿里巴巴集团） ； Fuzhou University（福州大学） ； Shanghai AI Laboratory（上海人工智能实验室） ； Southwest Jiaotong University（西南交通大学） 纠错

专题命中 视觉定位与Grounding ：VLM(abstract,abstract_cn)；vision-language model(abstract)；multimodal large language model(abstract)；分类 cs.CV

AI总结 提出MedP-CLIP，一种区域感知医学视觉语言模型，通过特征级区域提示集成机制，支持多种提示形式，在保持全局上下文的同时聚焦局部区域，在零样本识别、交互式分割等任务中显著优于基线方法。

Comments Accepted by Medical Image Analysis (MedIA)

Journal ref Medical Image Analysis, 113 (2026), 104193

MSPL: Multi-Step Pseudo-Labeling for Open-Vocabulary Object Detection

MSPL: 用于开放词汇目标检测的多步伪标签方法

Hojun Choi, Youngsun Lim, Jaeyo Shin, Hyunjung Shim

机构 * KAIST AI（韩国韩世大学AI研究所） ； Boston University（波士顿大学） 纠错

专题命中 视觉定位与Grounding ：vision-language model(abstract)；visual reasoning(abstract)；grounding(abstract)；分类 cs.CV

AI总结 提出MSPL框架，通过多步视觉推理（目标定位、类别识别、背景确认）生成伪标签，解决开放词汇检测中单步图像-文本匹配忽略中间推理步骤的问题，在OV-COCO和OV-LVIS上取得最优性能。

Comments This paper has been accepted by ECCV 2026

Business as Rulesual: A Benchmark and Framework for Business Rule Flow Modeling with LLMs

业务即规则：面向LLM的业务规则流建模基准与框架

Chen Yang, Ruping Xu, Ruizhe Li, Bin Cao, Jing Fan

机构 * Zhejiang University of Technology（浙江工业大学） ； Zhejiang Key Laboratory of Visual Information Intelligent Processing（浙江省视觉信息智能处理重点实验室） ； University of Aberdeen（阿伯丁大学） ； University of Birmingham（伯明翰大学） 纠错

专题命中 视觉定位与Grounding ：grounding(abstract)

AI总结 提出BREX基准（409份真实业务文档与2855条专家标注规则）和ExIde框架，通过可执行伪代码生成显式建模规则依赖，在13个LLM上验证了可执行接地作为归纳偏置的有效性。

Cosmos 3: Omnimodal World Models for Physical AI

Cosmos 3：面向物理AI的全模态世界模型

NVIDIA, :, Aditi, Niket Agarwal, Arslan Ali, Jon Allen, Martin Antolini, Adeline Aubame, Alisson Azzolini, Junjie Bai, Maciej Bala, Yogesh Balaji, Josh Bapst, Aarti Basant, Mukesh Beladiya, Mohammad Qazim Bhat, Zaid Pervaiz Bhat, Dan Blick, Vanni Brighella, Han Cai, Tiffany Cai, Eric Cameracci, Jiaxin Cao, Yulong Cao, Mark Carlson, Carlos Casanova, Ting-Yun Chang, Yan Chang, Yu-Wei Chao, Prithvijit Chattopadhyay, Roshan Chaudhari, Chieh-Yun Chen, Junyu Chen, Ke Chen, Qizhi Chen, Wenkai Chen, Xiaotong Chen, Yu Chen, An-Chieh Cheng, Click Cheng, Xiu Chia, Jeana Choi, Chaeyeon Chung, Wenyan Cong, Yin Cui, Magdalena Dadela, Nalin Dadhich, Wenliang Dai, Joyjit Daw, Alperen Degirmenci, Rodrigo Vieira Del Monte, Robert Denomme, Sameer Dharur, Marco Di Lucca, Ke Ding, Wenhao Ding, Yifan Ding, Yuzhu Dong, Nicole Drumheller, Yilun Du, Aigul Dzhumamuratova, Aleksandr Efitorov, Hamid Eghbalzadeh, Naomi Eigbe, Imad El Hanafi, Hassan Eslami, Benedikt Falk, Jiaojiao Fan, Jim Fan, Amol Fasale, Sergiy Fefilatyev, Liang Feng, Francesco Ferroni, Sanja Fidler, Xiao Fu, Vikram Fugro, Prashant Gaikwad, TJ Galda, Katelyn Gao, Yihuai Gao, Wenhang Ge, Sreyan Ghosh, Arushi Goel, Vivek Goel, Akash Gokul, Rama Govindaraju, Jinwei Gu, Miguel Guerrero, Elfie Guo, Aryaman Gupta, Siddharth Gururani, Hugo Hadfield, Song Han, Ankur Handa, Zekun Hao, Mohammad Harrim, Ali Hassani, Nathan Hayes-Roth, Yufan He, Chris Helvig, Cyrus Hogg, Madison Huang, Michael Huang, Sophia Huang, Yufan Huang, Jacob Huffman, DeLesley Hutchins, Suneel Indupuru, Boris Ivanovic, Arihant Jain, Joel Jang, Ryan Ji, Yanan Jian, Dongfu Jiang, Jingyi Jin, Atharva Joshi, Nikhilesh Joshi, Pranjali Joshi, Andy Ju, Jaehun Jung, Weiwei Kang, Scott Kassekert, Jan Kautz, Ashna Khetan, Julia Kiczka, Slawek Kierat, Gwanghyun Kim, Kuno Kim, Sunny Kim, Kezhi Kong, Xin Kong, Zhifeng Kong, Tomasz Kornuta, Egor Krivov, Hui Kuang, Saurav Kumar, Chia-Wen Kuo, George Kurian, Wojciech Kutak, JF Lafleche, Himangshu Lahkar, Omar Laymoun, Jayjun Lee, Sanggil Lee, Gabriele Leone, Boyi Li, Freya Li, Jiajun Li, Jinfeng Li, Ling Li, Pengcheng Li, Shangru Li, Tingle Li, Xiaolong Li, Xuan Li, Zhaoshuo Li, Zhiqi Li, Hao Liang, Maosheng Liao, Chen-Hsuan Lin, Tsung-Yi Lin, Ming-Yu Liu, Sifei Liu, Zihan Liu, Hai Loc Lu, Xiangyu Lu, Alice Luo, Ruipu Luo, Wenjie Luo, Jiangran Lyu, Martin Ding Ma, Nic Ma, Qianli Ma, Dawid Majchrowski, Louis Marcoux, Miguel Martin, Qing Miao, Ashkan Mirzaei, Shreyas Misra, Kaichun Mo, Durra Mohsin, Hyejin Moon, Pawel Morkisz, Saeid Motiian, Kirill Motkov, Seungjun Nah, Yashraj Narang, Deepak Narayanan, Thabang Ngazimbi, Julian Ouyang, Shubham Pachori, David Page, Yatian Pang, Sehwi Park, Mahesh Patekar, Mostofa Patwary, Marco Pavone, Trung Pham, Wei Ping, Soha Pouya, Shrimai Prabhumoye, Varun Praveen, Delin Qu, Hesam Rabeti, Morteza Ramezanali, Marilyn Reeb, Xuanchi Ren, Kristen Rumley, Wojciech Rymer, Jun Saito, Yeongho Seol, John Shao, Piyush Shekdar, Tianwei Shen, Humphrey Shi, Min Shi, Stella Shi, Kevin Shih, Mohammad Shoeybi, Mateusz Sieniawski, Shuran Song, Alexander Sotelo, Amir Sotoodeh, Sunil Srinivasa, Vignesh Srinivasakumar, Bartosz Stefaniak, Rahul Heinrich Steiger, Shangkun Sun, Jiaxiang Tang, Shitao Tang, Yangyang Tang, Yue Tang, Tolou Tavakkoli, Kayley Ting, Krzysztof Tomala, Wei-Cheng Tseng, Jibin Varghese, Sergei Vasilev, Thomas Volk, Raju Wagwani, Roger Waleffe, Andrew Z. Wang, Boxiang Wang, Haoxiang Wang, Qiao Wang, Shihao Wang, Shijie Wang, Ting-Chun Wang, Yan Wang, Yu Wang, Rohit Watve, David Wehr, Fangyin Wei, Xinshuo Weng, Jay Zhangjie Wu, Kedi Wu, Hongchi Xia, Summer Xiao, Tianjun Xiao, Kevin Xie, Daguang Xu, Jiashu Xu, Mengyao Xu, Ruqing Xu, Xingqian Xu, Yao Xu, Dinghao Yang, Dong Yang, Hans Yang, Xiaodong Yang, Xuning Yang, Yichu Yang, Yurong You, Zhiding Yu, Hao Yuan, Simon Yuen, Xiaohui Zeng, Pengcuo Zeren, Cindy Zha, Haotian Zhang, Jenny Zhang, Jing Zhang, Liangkai Zhang, Paris Zhang, Shun Zhang, Xuanmeng Zhang, Zhizheng Zhang, Ann Zhao, Yilin Zhao, Yuliya Zhautouskaya, Charles Zhou, Fengzhe Zhou, Shilin Zhu, Yuke Zhu, Dima Zhylko, Artur Zolkowski

机构 * NVIDIA 纠错

专题命中 GUI与屏幕智能体 ：vision-language model(abstract)；分类 cs.CV、cs.AI、cs.LG

AI总结 提出基于统一混合Transformer架构的全模态世界模型Cosmos 3，联合处理语言、图像、视频、音频和动作序列，在理解和生成任务上达到新最优，为具身智能体提供可扩展的通用骨干。

MOCHA: Multi-modal Objects-aware Cross-arcHitecture Alignment

MOCHA: 多模态对象感知的跨架构对齐

Elena Camuffo, Francesco Barbato, Mete Ozay, Simone Milani, Umberto Michieli

机构 * University of Padova（帕多瓦大学） 纠错

专题命中 VLM训练与架构 ：VLM(summary_cn,abstract)；vision-language model(abstract)；分类 cs.CV、cs.AI、cs.LG

AI总结 提出MOCHA蒸馏框架，将冻结VLM教师的多模态区域知识迁移至轻量视觉检测器，通过双重损失实现局部对齐与全局关系一致性，在少样本个性化检测中平均提升10.1%。

Comments 18 pages main paper, 10 pages supplementary material

Journal ref Proceedings of ECCV 2026

When Language Overwrites Vision: Over-Alignment and Geometric Debiasing in Vision-Language Models

语言覆盖视觉：视觉语言模型中的过度对齐与几何去偏

Harshvardhan Saini, Samyak Jha, Yiming Tang, Dianbo Liu

机构 * Indian Institute of Technology Dhanbad（印度理工学院丹巴德分校） ； National University of Singapore（新加坡国立大学） 纠错

专题命中 VLM训练与架构 ：vision-language model(title,abstract)；分类 cs.CV、cs.AI

AI总结 本文研究视觉语言模型中过度对齐导致的幻觉问题，提出几何去偏方法，通过投影消除文本子空间影响，减少幻觉并提升长文本生成性能。

HiPath: Hierarchical Vision-Language Alignment for Structured Pathology Report Prediction

HiPath: 用于结构化病理报告预测的分层视觉-语言对齐

Ruicheng Yuan, Zhenxuan Zhang, Anbang Wang, Liwei Hu, Xiangqian Hua, Yaya Peng, Jiawei Luo, Guang Yang

机构 * College of Computer Science and Electronic Engineering, Hunan University（湖南大学计算机科学与电子工程学院） ； Department of Bioengineering and Imperial-X, Imperial College London（帝国理工学院伦敦校区生物工程系） ； Department of Pathology, Xiangtan Maternal and Child Health Hospital（湘潭 maternal and child health hospital pathology department） ； Department of Pathology, The First People’s Hospital of Xiangtan City（湘潭市第一人民医院病理科） 纠错

专题命中 VLM训练与架构 ：VLM(abstract,abstract_cn)；vision-language model(abstract)；分类 cs.CV、cs.AI、cs.LG

AI总结 提出HiPath框架，通过分层补丁聚合器、对比学习和槽位掩码诊断预测，在冻结UNI2和Qwen3骨干上实现结构化病理报告预测，准确率达68.9%，安全率97.3%。

Comments 10 pages, 1 figures, 3 tables