DISC-MedLLM: Bridging General Large Language Models and Real-World Medical Consultation

TL;DR

DISC-MedLLM is a new large language model designed specifically for accurate, truthful, and conversational medical responses, constructed using innovative data strategies and outperforming existing models in medical consultations.

Contribution

The paper introduces DISC-MedLLM, a novel medical LLM trained with high-quality datasets created through knowledge-graphs, dialogue reconstruction, and human preferences, advancing medical AI capabilities.

Findings

Outperforms existing medical LLMs in consultation scenarios

Effective bridging of general LLMs and real-world medical responses

Public release of dataset and model for further research

Abstract

We propose DISC-MedLLM, a comprehensive solution that leverages Large Language Models (LLMs) to provide accurate and truthful medical response in end-to-end conversational healthcare services. To construct high-quality Supervised Fine-Tuning (SFT) datasets, we employ three strategies: utilizing medical knowledge-graphs, reconstructing real-world dialogues, and incorporating human-guided preference rephrasing. These datasets are instrumental in training DISC-MedLLM, surpassing existing medical LLMs in both single-turn and multi-turn consultation scenarios. Extensive experimental results demonstrate the effectiveness of the proposed model in bridging the gap between general language models and real-world medical consultation. Additionally, we release the constructed dataset and model weights to further contribute to research and development. Further details and resources can be found at https://github.com/FudanDISC/DISC-MedLLM