摘要
针对斜视细粒度诊断中现有深度学习缺乏透明推理及大视觉语言模型易产生幻觉的问题，本文提出 MAGIS 框架。该框架将黑盒生成转化为结构化诊断流程，包含假设生成、双重证据约束上下文及基于证据的校正验证机制。通过整合视觉证据与临床规则，MAGIS 显著提升了诊断准确率（加权 F1 从 72.0% 升至 91.3%）及报告的可解释性与临床可靠性，为构建精准、可信的医疗诊断系统提供了有效方案。
AI 推荐理由
论文核心提出基于证据的多智能体推理框架，解决医疗诊断中的幻觉问题，强调可解释推理过程。
研究机构
School of Information and Software Engineering, University of Electronic Science and Technology of China, Chengdu, 611731, China
Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China, Shenzhen, 518000, China
Joint Shantou International Eye Center of Shantou University and The Chinese University of Hong Kong, Shantou, 515041, China
School of Artificial Intelligence, Guangzhou City Polytechnic, Guangzhou, 510655, China
Medical College, Shantou University, Shantou, 515041, China
College of Engineering, Shantou University, Shantou, 515063, China
Department of Ophthalmology, Xinhua Hospital Affiliated to Shanghai Jiaotong University School of Medicine, Shanghai, 200092, China
Shenzhen Loop Area Institute, Shenzhen, 518048, China
论文信息