Wikidata Knowledge Graph to Enable Equitable and Validated Ge...
summary
Wikidata provides an open CC0 knowledge graph of over 100M crowd-sourced items across ~300 languages and is being vectorised so LLMs can use it for retrieval and validation. A prototype GraphRAG pipeline embeds ~20 billion statement chunks and a few hundred million paragraph chunks, stores them in an Astra vector DB, and uses Hugging Face/Jina embeddings plus a REST API to fetch full JSON context. The demo successfully answers queries and the code and Docker pipeline are on GitHub for reuse.
AI-written summary · about
chapters (7)
- 0:00Introduction to Wikidata
- 0:32Wikidata as a Knowledge Graph
- 7:13Integrating Wikidata with AI
- 10:17Building Graph RAG systems
- 16:32Live demonstration
- 18:30Future goals and collaboration
- 21:06Q&A session
related talks
31mEmpowering Open-Source Generative AI by Integrating the Wikidat...
53mPractical GraphRAG - Making LLMs smarter with Knowledge Graphs
20mPractical GraphRAG: Making LLMs smarter with Knowledge Graphs
42mMaking OSS AI Real with Knowledge Graphs
19mWisdom-Driven Knowledge Augmented Generation at Scale
30mMaking Your Knowledge Graph LLM-Ready: Quality Assessment and Design Strategies for GraphRAG
26mPractical GraphRAG Making LLMs smarter with Knowledge Graphs
24mEmpowering Lifelong Learning: Transforming Data into Insights With GraphRAG