Large Language Models and Knowledge Graphs Synergistically Enhancing Personalized Learning

The rapid development of artificial intelligence technology has promoted the intelligence of the education field. The collaborative work of large language models (LLMs) and knowledge graphs (KGs) can effectively develop personalized learning paths. Personalized learning in programming courses is crucial for meeting students' different needs and proficiency levels. This article proposes a method for personalized teaching of programming courses using LLMs and KGs. The LLMs is used to analyze students' learning status, provide real-time feedback, and generate customized learning materials, while the KGs provides a structured knowledge base to plan essential skills and learning progress, highly integrating the advantages of the LLMs and the logical connections of knowledge from various disciplines. The application results show that designing a knowledge system ontology through disciplinary knowledge analysis, retraining and fine-tuning the LLMs, forms the final personalized knowledge system, improves the accuracy of learning path recommendations and the relevance of generated exercises, and can be easily promoted to other disciplinary fields.