Empowering Lifelong Learning: Transforming Data into Insights With GraphRAG
summary
Built a knowledge graph linking jobs, skills, courses and providers and populated it with 40–50k nodes, over 1,000 SSOC occupations and ~700,000 relationships; entity linking uses proprietary ML, labor‑market signals use a sliding‑window weighted algorithm while expert annotations remain fixed. Exposed the graph via a prototype web portal + parameterized Cypher APIs for analysts and a free‑text RAG chatbot (trial Telegram bot) combining vector+graph traversal via Neo4j GraphRAG; pilots are promising but the system remains work in progress.
AI-written summary · about
chapters (8)
- 0:00Introduction and team
- 2:54Why use knowledge graphs
- 7:10Building the knowledge graph
- 10:44Making the graph dynamic
- 15:49Turning data into insights
- 19:22Demo of web portal
- 21:50Exploring GraphRAG
- 23:57Q&A
related talks
53mPractical GraphRAG - Making LLMs smarter with Knowledge Graphs
26mPractical GraphRAG Making LLMs smarter with Knowledge Graphs
27mAdaptive Knowledge Graphs for Dynamic RAG Systems
32mGraphRAG: Powering Up LLMs with Knowledge Graphs
42mMaking OSS AI Real with Knowledge Graphs
36mAgentic GraphRAG: AI’s Logical Edge
01h 19mIntro to GraphRAG
20mPractical GraphRAG: Making LLMs smarter with Knowledge Graphs