Graph RAG: Bringing together graph and vector search to empower retrieval
summary
Claims that combining graph and vector retrieval yields more factual RAG outcomes and shows this on an Interstellar knowledge base using a hybrid pipeline. Indexing and serving split is explicit: chunked text goes into a vector store while extracted triples populate a graph; retrieval concatenates top-k vectors with graph triples and a ranker. Implementation details include Kùzu as the embedded graph, LanceDB as the vector store, an LLM text-to-Cypher step to translate queries, and an evaluated hybrid RAG that improves factual recall.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:45Graphs and graph databases
- 7:43RAG and Graph RAG history
- 15:27Graph RAG retrieval
- 18:35Graph RAG indexing and serving
- 21:26Demo: Graph RAG on movies
- 27:45Key takeaways
related talks
27mWhen Vectors Aren’t Enough: Unlocking AI’s Hidden Connections with Graphs
39mEnhancing RAG with Neo4j Knowledge Graph
33mVector Databases in the Age of AI: Embracing RAG for Enhanced Search Capabilities
01h 17mGraph Database and Retrieval Augmented Generation
42mIntegrating Graph Context With OpenSearch for Enhanced RAG Application Performance
27mAdaptive Knowledge Graphs for Dynamic RAG Systems
40mRAG Pipeline Evaluation
06mWhat makes vector databases essential for RAG and AI apps?