多模态多智能体强化学习在放射学报告生成中的应用
Multi-Modal Multi-Agent Reinforcement Learning for Radiology Report Generation
AI总结：
本文提出MARL-Rad框架，通过多智能体强化学习提升放射学报告生成的临床效果，通过联合优化区域特定智能体和全局整合智能体，提高报告的准确性和一致性。
摘要：
我们提出MARL-Rad，一种多模态多智能体强化学习框架，用于放射学报告生成，该框架在部署的放射学工作流中训练整个智能体系统。MARL-Rad解决了事后智能体化限制，即固定LLMs被组织成手工设计的智能体工作流，而未针对其分配的角色进行优化。我们的框架将胸部X光解读分解为区域特定智能体和一个全局整合智能体，并使用临床可验证的奖励共同优化它们。在MIMIC-CXR和IU X光数据集上的实验表明，MARL-Rad一致提高了临床效能指标，如RadGraph、CheXbert和GREEN分数，实现了最先进的临床效能性能。进一步分析显示，MARL-Rad提高了侧向一致性，并生成了更准确和详细的报告。盲审临床医生评估进一步表明，MARL-Rad生成的报告在临床上与真实报告相当。
英文摘要：
We propose MARL-Rad, a multi-modal multi-agent reinforcement learning framework for radiology report generation that trains the entire agentic system on policy within its deployed radiology workflow. MARL-Rad addresses the limitation of post-hoc agentization, where fixed LLMs are organized into hand-designed agentic workflows without being optimized for their assigned roles. Our framework decomposes chest X-ray interpretation into region-specific agents and a global integrating agent, and jointly optimizes them using clinically verifiable rewards. Experiments on the MIMIC-CXR and IU X-ray datasets show that MARL-Rad consistently improves clinical efficacy metrics such as RadGraph, CheXbert, and GREEN scores, achieving state-of-the-art clinical efficacy performance. Further analyses show that MARL-Rad improves laterality consistency and produces more accurate and detailed reports. A blinded clinician evaluation further suggests that MARL-Rad produces reports clinically comparable to ground-truth reports.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
