AI Powered Knowledge Graph Generator for Video Learning Using RAG

The rapid growth of educational video content has transformed modern learning, yet it in- troduces challenges such as knowledge fragmentation, poor retention, and inefficient retrieval. Existing tools such as video summarizers and note-taking systems fail to provide a persistent and interconnected learning experience. This paper presents an AI-powered knowledge graph gener- ation system that transforms passive video consumption into an active and queryable knowledge base. The proposed system processes YouTube videos by extracting transcripts and perform- ing multi-dimensional analysis using large language models, generating structured outputs such as summaries, quizzes, flashcards, and key insights. To enable efficient retrieval, the system encodes content into high-dimensional semantic embeddings stored in a vector database. A Retrieval-Augmented Generation (RAG) framework allows users to interact with stored knowl- edge through natural language queries, providing context-aware and source-grounded responses. Additionally, an automated knowledge graph captures relationships between concepts across multiple videos, enabling visual exploration of interconnected knowledge. Experimental evaluation demonstrates improved knowledge retention and retrieval efficiency compared to traditional approaches. The proposed system provides a scalable and intelligent solution for long-term learning from video-based content.