Design of a Multi-Agent System for Personalized Training of Normal School Students' Teaching Skills

With the development of artificial intelligence technology, especially the emergence of generative AI, new opportunities have arisen for the training of normal school students' teaching skills. This paper discusses the design of intelligent agents and the current status of their application in the training of normal school students' teaching skills. It believes that the application of intelligent agent technology in this field is feasible and has a solid foundation. After analyzing the problems of traditional training modes, a four-layer, three-agent system framework is constructed, which includes the user layer, service layer, tool layer, and knowledge layer. The aim is to provide a comprehensive, personalized teaching skills training environment for normal school students, helping them to better master teaching skills and laying a solid foundation for their future teaching careers.