DrAgent: empowering Large Language Models as medical agents for multi-hop medical reasoning

Conference item

Although large language models (LLMs) have demonstrated outperforming human experts in medical examinations, it remains challenging to adopt LLMs in real-world clinical decisionmaking that typically involves multi-hop medical reasoning. Common practices include prompting commercial LLMs and fine-tuning LLMs on medical data. However, in the clinical domain, using commercial LLMs raises privacy concerns regarding sensitive patient data. Finetuning competitive medical LLMs for different tasks usually requires extensive data and computing resources, which are difficult to acquire, especially in medical institutions with limited infrastructure. We propose DrAgent, which can build LLMs as agents to deliver accurate medical decision-making and reasoning. In implementation, we take a lightweight LLM as the backbone to collaborate with diverse clinical tools. To make efficient use of data, DrAgent introduces recursive curriculum learning to optimize the LLM in an easy-to-hard progression. The results show that our approach achieves competitive performance on diverse datasets.

Authors: Liu, F; Li, Z; Zhou, H; Yin, Q; Yang, J

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: ACL Anthology
• Host title: Findings of the Association for Computational Linguistics: EMNLP 2025
• Pages: 15656–15668
• Publication date: 2025-11-01
• Acceptance date: 2025-08-21
• Event title: 30th Conference on Empirical Methods in Natural Language Processing (EMNLP 2025)
• Event location: Suzhou, China
• Event website: https://2025.emnlp.org/
• Event start date: 2025-11-04
• Event end date: 2025-11-09
• DOI: 10.18653/v1/2025.findings-emnlp.848
• Language: English