In this session, Leann Chen will introduce GraphRAG, a method that integrates knowledge graphs with large language models (LLMs) to enhance Retrieval-Augmented Generation (RAG) systems. Graph RAG can address challenges like hallucinations and limited explainability in LLM-based systems.

We will walk through the process of building a Graph RAG, using examples to show how this approach improves accuracy, relevance, and transparency by combining the strengths of both vector-based and graph-based retrieval methods.

The session will also examine the pros and cons of vector-only and graph-only RAG systems, demonstrating how Graph RAG can effectively merge the best aspects of both.