Application, Challenges and Prospects of Large Language Models in the Medical Field

Artificial intelligence has become a new trend in the new era. Within its field, large language models (LLM) have unique advantages in today's massive data due to their excellent text understanding and generation capabilities. This article briefly outlines the development background and architecture of LLM. In view of the rapid development of large language models, this article analyzes the practical applications, challenges and future development prospects of LLM from a medical perspective. The advantages of LLM in medical application were analyzed, and the application of LLM in various medical fields was reviewed by taking BioGPT, Zhongjing (CMLM-ZhongJing) large language model and HuatuoGPT as examples. Secondly, the paper analyzes the challenges that LLM currently faces in the medical field, such as the "hallucination" problem, the explainability problem, etc., and provides corresponding solutions. Finally, the future development prospects of LLM in the medical field are proposed, which is to transform from single modality to multimodality, from large-scale to lightweight, strengthen the integration with medical equipment, and better provide services for relevant medical staff, patients, and students in the medical field.