Abstract:
The research study developed an adaptive learning system based on Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) technology to deliver customized educational content. The system differentiates from traditional LLM educational software by accepting complete lecture materials which ensures quiz responses and feedback match the specific content of the current course. The application retrieves dynamic relevant content from lecture slides to provide focused structured learning beyond standardized pre-trained responses. Pinecone serves as a vector database for semantic content retrieval and open provides GPT for natural language generation from the system architecture. The educational materials undergo SentenceTransformers processing to create semantic embeddings that enable both precise content retrieval as well as contextual adjustments. Specific course materials determine the alignment of quizzes and feedback through this method instead of using pre-existing knowledge as a basis. The first phase concentrates on implementing the system for a single subject to provide detailed refinement of retrieval techniques and quiz adjustment. The system's functionality verifies that it obtains lecture-specific content while producing structured quiz questions that create context-based feedback responses. The evaluation of the system's learning effectiveness through quantitative methods remains an unaddressed task. The forthcoming research tasks will analyse new retrieval approaches alongside optimized content selection methods and different retriever comparison to boost accuracy and adaptability rates. This research integrates RAG retrieval alongside LLM-based generation with direct course materials to demonstrate improved learning results over static and generic educational tools utilizing LLMs. Organized assessments will determine how well the approach improves student engagement while enhancing both understanding and customized learning efficiency.
