Multi-Modal Large Language Models (MLLMs), despite being successful, exhibit limited generality and often fall short when compared to specialized models. Recently, LLM-based agents have been developed to address these challenges by selecting appropriate specialized models as tools based on user inputs. However, such advancements have not been extensively explored within the medical domain. To bridge this gap, this paper introduces the first agent explicitly designed for the medical field, named \textbf{M}ulti-modal \textbf{Med}ical \textbf{Agent} (MMedAgent). We curate an instruction-tuning dataset comprising six medical tools solving seven tasks across five modalities, enabling the agent to choose the most suitable tools for a given task. Comprehensive experiments demonstrate that MMedAgent achieves superior performance across a variety of medical tasks compared to state-of-the-art open-source methods and even the closed-source model, GPT-4o. Furthermore, MMedAgent exhibits efficiency in updating and integrating new medical tools. Codes and models are all available.
@article{arxiv.2407.02483,
title = {MMedAgent: Learning to Use Medical Tools with Multi-modal Agent},
author = {Binxu Li and Tiankai Yan and Yuanting Pan and Jie Luo and Ruiyang Ji and Jiayuan Ding and Zhe Xu and Shilong Liu and Haoyu Dong and Zihao Lin and Yixin Wang},
journal= {arXiv preprint arXiv:2407.02483},
year = {2024}
}