UR$^{2}$-MLLM: Uncertainty-aware Revisit Reasoning in Multimodal Large Language Models for Radiology Report Generation
UR²-MLLM:面向放射科报告生成的多模态大语言模型中基于不确定性感知的重访推理
机构 * MedVisAI Lab(MedVisAI实验室) ; Lee Kong Chian School of Medicine, Nanyang Technological University(南洋理工大学李光前医学院) ; Centre of AI in Medicine, Singapore(新加坡医学人工智能中心) ; AI Thrust, The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)AI方向) ; Ruijin Hospital, Shanghai Jiao Tong University School of Medicine(上海交通大学医学院附属瑞金医院)
专题命中 视觉推理 :MLLM(title,title_cn);multimodal large language model(title,abstract);grounding(abstract,abstract_cn);分类 cs.CV
AI总结 该研究针对放射科报告生成任务,提出UR²-MLLM框架,通过动态重访不确定区域的机制实现最优性能,提升报告的可靠性与临床适配性。
Comments EMNLP 2026 Findings