arXiv 2508.157462026-02-01cs.CLcs.AIcs.CV

端到端代理RAG系统训练用于可追溯的诊断推理

End-to-End Agentic RAG System Training for Traceable Diagnostic Reasoning

Shanghai Jiao Tong University（上海交通大学）；Shanghai AI Laboratory（上海人工智能实验室）；Xinhua Hospital affiliated to Shanghai Jiao Tong University School of Medicine（上海交通大学医学院附属新华医院）；Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine（上海交通大学医学院附属第九人民医院）；Shanghai Jiao Tong University School of Medicine（上海交通大学医学院）

Qiaoyu Zheng, Yuze Sun, Chaoyi Wu, Weike Zhao, Pengcheng Qiu, Yongguo Yu, Kun Sun, Jian Zhang, Yanfeng Wang, Ya Zhang, Weidi Xie

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

Deep-DxSearch通过端到端强化学习训练代理RAG系统，实现可追溯的诊断推理，显著提升诊断准确率。

摘要：

将大语言模型（LLM）整合到医疗健康领域受到知识限制、幻觉以及与循证医学（EBM）脱节的制约。虽然检索增强生成（RAG）提供了解决方案，但当前系统往往依赖静态工作流，忽略了临床医生的迭代、假说-演绎推理。为此，我们引入了Deep-DxSearch，一种通过强化学习（RL）端到端训练的代理RAG系统，用于可追溯的诊断推理。Deep-DxSearch充当一个主动调查者，在包含16000多个指南衍生疾病资料、150000多个病例推理患者记录以及超过2700万篇生物医学文档的环境中，将LLM作为代理。利用软可验证奖励，该模型学习制定查询、评估证据并细化搜索以缩小诊断差距。实验表明，我们的端到端RL框架在提示工程和无训练RAG方法上表现一致。在常见和罕见疾病的分布内（ID）和分布外（OOD）基准测试中，Deep-DxSearch超越了包括GPT-4o、DeepSeek-R1和医学专用框架在内的强大基线，平均准确率提升22.7%。在与150个真实世界案例的验证中，Deep-DxSearch将医生的平均诊断准确率从45.6%提升到69.1%。这些结果表明，进化代理系统以利用大规模医疗健康数据中的统计规律是可信诊断助手的关键。所有数据、代码和检查点均可在https://qiaoyu-zheng.github.io/Deep-DxSearch获取。

英文摘要：

The integration of Large Language Models (LLMs) into healthcare is constrained by knowledge limitations, hallucinations, and a disconnect from Evidence-Based Medicine (EBM). While Retrieval-Augmented Generation (RAG) offers a solution, current systems often rely on static workflows that miss the iterative, hypothetico-deductive reasoning of clinicians. To address this, we introduce Deep-DxSearch, an agentic RAG system trained end-to-end via reinforcement learning (RL) for traceable diagnostic reasoning. Deep-DxSearch acts as an active investigator, treating the LLM as an agent within an environment of 16,000+ guideline-derived disease profiles, 150,000+ patient records for case-based reasoning, and over 27 million biomedical documents. Using soft verifiable rewards that co-optimize retrieval and reasoning, the model learns to formulate queries, evaluate evidence, and refine searches to close diagnostic gaps. Experiments show our end-to-end RL framework consistently outperforms prompt-engineering and training-free RAG methods. On in-distribution (ID) and out-of-distribution (OOD) benchmarks for common and rare diseases, Deep-DxSearch surpasses strong baselines-including GPT-4o, DeepSeek-R1, and medical-specific frameworks-achieving an average accuracy gain of 22.7% over the second-best model. In validation with 150 real-world cases, Deep-DxSearch boosts physicians' average diagnostic accuracy from 45.6% to 69.1%. These results indicate that evolving agentic systems to leverage statistical regularities in large-scale healthcare data is key for trustworthy diagnostic assistants. All data, code, and checkpoints are available at https://qiaoyu-zheng.github.io/Deep-DxSearch.

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。