Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Learn how to enhance retrieval-augmented generation (RAG) systems through entity-resolved knowledge graphs in this 31-minute conference talk. Explore the challenges of building accurate and reliable large language model applications, particularly the issue of hallucinations that plague LLM performance. Discover how traditional RAG approaches using vector databases can be significantly improved by representing data in knowledge graph formats, and understand why many knowledge graph implementations fail due to entity resolution problems. Examine real-world examples of entity duplication issues, from simple cases like "My Company Inc." versus "My Company, Inc." to complex variations such as "Liz Smith," "Elizabeth Conner-Smith," and "Dr. L. Conner-Smith" that cannot be resolved with basic string matching. Master sophisticated entity resolution techniques that incorporate multiple disparate data sources to create entity-resolved knowledge graphs (ERKGs), where duplicate entities are consolidated into single entities while preserving all associated information. See practical demonstrations using real-world data that showcase improvements in both graph data science tasks and LLM accuracy, and gain actionable insights for implementing ERKGs in your own RAG systems to achieve transformative performance enhancements.
