In the realm of LLM-powered applications, Retrieval Augmented Generation (RAG) consolidated as established itself as a leading framework. It is based on the idea of retrieving relevant context from custom knowledge base via embeddings, that are vector representation of texts.
Over the last months, we witnessed the rise of numerous variants of traditional RAG, and one of the most prominent is GraphRAG, based on graph-like organization of the knowledge base.
In this session, we explore the idea behind GraphRAG with a hands-on implementation with LangChain.