通知：联邦节假日调休：哥伦布日 / 原住民日（Columbus Day / Indigenous Peoples' Day），今日（13日）无更新。通知：联邦节假日调休：哥伦布日 / 原住民日（Columbus Day / Indigenous Peoples' Day），今日（13日）无更新。
arXiv 2511.19220cs.CVcs.AI
大视觉语言模型真的在医学图像上具有基础性吗？来自意大利临床视觉问答的证据
Are Large Vision Language Models Truly Grounded in Medical Images? Evidence from Italian Clinical Visual Question Answering
SIIAM；NSBProject；Dept. of Life Sciences & Public Health, UCSC（生命科学与公共卫生系，UCSC）；ASL RM 4；UCSC；Univ. Paris Cité（巴黎Cité大学）；Univ. of Pisa（比萨大学）
Federico Felizzi, Olivia Riccomi, Michele Ferramola, Francesco Andrea Causio, Manuel Del Medico, Vittorio De Vita, Lorenzo De Mori, Alessandra Piscitelli, Pietr… 展开作者
Federico Felizzi, Olivia Riccomi, Michele Ferramola, Francesco Andrea Causio, Manuel Del Medico, Vittorio De Vita, Lorenzo De Mori, Alessandra Piscitelli, Pietro Eric Risuleo, Bianca Destro Castaniti, Antonio Cristiano, Alessia Longo, Luigi De Angelis, Mariapia Vassalli, Marcello Di Pumpo
备注：
Accepted at the Workshop on Multimodal Representation Learning for Healthcare (MMRL4H), EurIPS 2025
Large vision language models (VLMs) have achieved impressive performance on medical visual question answering benchmarks, yet their reliance on visual information remains unclear. We investigate whether frontier VLMs demonstrate genuine visual grounding when answering Italian medical questions by testing four state-of-the-art models: Claude Sonnet 4.5, GPT-4o, GPT-5-mini, and Gemini 2.0 flash exp. Using 60 questions from the EuropeMedQA Italian dataset that explicitly require image interpretation, we substitute correct medical images with blank placeholders to test whether models truly integrate visual and textual information. Our results reveal striking variability in visual dependency: GPT-4o shows the strongest visual grounding with a 27.9pp accuracy drop (83.2% [74.6%, 91.7%] to 55.3% [44.1%, 66.6%]), while GPT-5-mini, Gemini, and Claude maintain high accuracy with modest drops of 8.5pp, 2.4pp, and 5.6pp respectively. Analysis of model-generated reasoning reveals confident explanations for fabricated visual interpretations across all models, suggesting varying degrees of reliance on textual shortcuts versus genuine visual analysis. These findings highlight critical differences in model robustness and the need for rigorous evaluation before clinical deployment.
Truly Assessing Fluid Intelligence of Large Language Models through Dynamic Reasoning Evaluation
Shanghai Jiao Tong University（上海交通大学）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；Harbin Institute of Technology（哈尔滨工业大学）；The University of Hong Kong（香港大学）
Watch Before You Answer: Learning from Visually Grounded Post-Training
University of British Columbia（不列颠哥伦比亚大学）；Vector Institute（向量研究所）；Etude AI；Kolors Team, Kuaishou Technology（快手科技Kolors团队）；University of Toronto（多伦多大学）；University of Waterloo（滑铁卢大学）；University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
GAOKAO-Eval: Does high scores truly reflect strong capabilities in LLMs?
Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；East China Normal University（华东师范大学）；Harbin Institute of Technology（哈尔滨工业大学）；The Chinese University of Hong Kong（香港中文大学）
From Judgments to Issues: Structured Extraction of Legal Reasoning with Citation-Hallucination Control
Quantitative and Digital Law Laboratory（量化与数字法律实验室）；CIRSFID - Alma AI, Faculty of Law University of Bologna（CIRSFID - 阿尔玛AI，博洛尼亚大学法学院）；Robert Schuman Centre European University Institute（罗伯特·施曼中心欧洲大学研究所）
Do Vision-Language Models Truly Perform Vision Reasoning? A Rigorous Study of the Modality Gap
College of Computing and Data Science, Nanyang Technological University, Singapore（南洋理工大学计算机与数据科学学院，新加坡）；Alibaba-NTU Global e-Sustainability CorpLab (ANGEL)（阿里巴巴-国立大学全球可持续发展公司实验室（ANGEL））；Tongyi Lab, Alibaba Group, China（阿里云实验室，阿里巴巴集团，中国）
Proactive Rejection and Grounded Execution: A Dual-Stage Intent Analysis Paradigm for Safe and Efficient AIoT Smart Homes
School of Information and Electronic Engineering (Sussex Artificial Intelligence Institute), Zhejiang Gongshang University（信息与电子工程学院（Sussex人工智能研究院），浙江工商大学）；Sussex Artificial Intelligent Institute, Zhejiang Gongshang University（Sussex人工智能研究院，浙江工商大学）；Department of Engineering and Design, University of Sussex（工程与设计系， Sussex大学）；Artificial Intelligence Research Institute, Shenzhen MSU-BIT University（人工智能研究院，深圳MSU-BIT大学）；College of Computer Science and Software Engineering, Shenzhen University（计算机科学与软件工程学院，深圳大学）；Department of Electrical and Computer Engineering, The University of British Columbia（电气与计算机工程系，不列颠哥伦比亚大学）
CogToM: A Comprehensive Theory of Mind Benchmark inspired by Human Cognition for Large Language Models
BrainCog Lab, Institute of Automation, Chinese Academy of Sciences（脑认知实验室，自动化研究所，中国科学院）；Beijing Institute of AI Safety and Governance (Beijing-AISI)（北京人工智能安全与治理研究所）；Beijing Key Laboratory of Safe AI and Superalignment（北京安全人工智能与超对齐重点实验室）；School of Artificial Intelligence, UCAS（人工智能学院，中国科学院大学）；Long-term AI（长期人工智能）
A Sentiment Consolidation Framework for Meta-Review Generation
School of Computing and Information Systems, The University of Melbourne（墨尔本大学计算与信息系统学院）；Language Technologies Institute, Carnegie Mellon University（卡内基梅隆大学语言技术研究所）
