Neo4j Live: Entity Architecture for Efficient RAG on Graphs

Join us live as we dive into the Entity Architecture for efficient Retrieval-Augmented Generation (RAG) on Knowledge Graphs. The architecture organizes RAG workflows into three distinct layers: an input layer for data ingestion, a middle layer for knowledge graph representation and reasoning, and an output layer for generating contextually accurate AI-driven results. Learn how this innovative approach leverages fixed entities to enhance data retrieval, improve contextual understanding and boost AI performance. Guest: Irina Adamchic, PhD From Local to Global: https://bit.ly/40tc4Ov Blog: https://medium.com/@irina.karkkanen/three-layer-fixed-entity-architecture-for-efficient-rag-on-graphs-787c70e3151a 0:00 - Welcome and introduction 0:32 - Overview of today's topic: Entity Architecture for Efficient RAG on Graphs 1:37 - Irina Adamchic - Introduction 3:37 - Discovering graph solutions for RAG use cases 6:05 - How GenAI projects evolved with graph technology 9:57 - Fixed Entity Architecture: Origin and challenges it solves 16:53 - Building the ontology layer: A practical approach 24:43 - Three-layer architecture and how it enhances scalability 31:04 - Comparing Microsoft GraphRAG vs. Fixed Entity Architecture 46:15 - Closing remarks, upcoming events, and resources #genai #graphrag #entity #knowledgegraph #neo4j