Abstract: With the background of intelligent and digital education, we construct a diversified AI teaching assistant platform of optics based on large language models and knowledge bases in order to meet the needs of students’ learning and exploring. The platform starts from the OBE concept and leads to three parts: knowledge graph, AI questions and answers module, and MOOC learning platform. Based on these parts, we construct a back-end platform to analyze the learning data. In terms of content, knowledge graph can construct teaching frameworks and judge mastery levels. For the AI questions and answers module, it introduces tool plugins with the characteristics of the optics course and gives professional replies. Additionally, the MOOC learning platform focuses on integrating learning resources such as videos and courseware. The learning data of the students, including learning progress and mastery levels, can be analyzed in the back-end of AI teaching assistant platform. The application of the AI teaching assistant platform in teaching can satisfy the needs of students for fragmented learning and may promote the intelligent and digital development of course teaching.

Key words: AI teaching assistant, diversified platform, AI questions and answers, intelligent