Abstract: This article is based on the teaching practice of graduate courses on neural networks and their applications, and proposes a research feedback teaching model to address the problem of traditional teaching lagging behind under the rapid development of artificial intelligence technology. This model is developed from four aspects: integrating scientific research content into the curriculum, project driven teaching design, dual mentor collaborative guidance, and diversified evaluation mechanism. It focuses on introducing real scientific research projects and industry cases into the classroom to enhance the application orientation and practical depth of the curriculum. This article combines specific teaching cases to systematically explain the curriculum design path, innovative teaching methods, and student development support, aiming to provide references for the reform of artificial intelligence courses.

Key words: Project driven, Neural network, Reform in teaching, Research feeds back teaching, Dual teacher collaboration, Health monitoring