Leveraging Advanced Language Models for Personalized Education: Design and Evaluation of AI-assisted Course Tutoring
Traditional tutoring typically encompasses one-on-one interactions between students and their instructors. While this method affords the opportunity for detailed feedback, it presents significant challenges when attempting to reach larger student population and personalize the experience for each individual. With generative artificial intelligence joining forces with tutoring, the proposed project aims to develop an AI-assisted tutoring system based on two Large Language Models (LLM) for CUHKSZ students. The intelligent system is guided by course learning outcomes and comprises two components: an analyzer and a tutor. The analyzer accurately tracks students’ progress and identifies individual strengths and weaknesses, while the tutor provides personalized learning support and focused assistance at the precise points where students’ learning requires enhancement. Additionally, the AI tutor aids instructors in analyzing class-wide progress towards attaining learning outcomes. By maintaining comprehensive learning profiles, including assignments, exams, and classroom interactions, the AI-assisted tutoring system delivers tailored support to each student, beyond what can be accomplished by human instructors alone. The LLM implementation and deployment solutions developed in this project will be open-sourced to facilitate intelligent teaching in other courses.
