Juan Sequeda and Dean Allemang on boosting LLMs with Knowledge Graphs for Enterprise Q&A.
Subscribe: Apple • Spotify • Overcast • Google • AntennaPod • Podcast Addict • Amazon • RSS.
Juan Sequeda (Principal Scientist & Head of AI Lab) and Dean Allemang (Principal Solutions Architect) are knowledge graph experts at data.world, a startup that offers a data catalog powered by a knowledge graph to help organizations better understand and gain value from their data. In this episode we discuss their recent paper, “A Benchmark to Understand the Role of Knowledge Graphs on Large Language Model’s Accuracy for Question Answering on Enterprise SQL Databases”.
The primary motivation of using KGs is to provide a business context layer to LLMs, addressing the gaps in understanding and contextualizing enterprise data. This enhancement aims to reduce inaccuracies and hallucinations, thereby increasing the trustworthiness and effectiveness of LLMs in enterprise settings.
Interview highlights – key sections from the video version:
- Motivation behing incorporating knowledge graphs (KGs) into Q&A systems against enterprise databases
- Setup used for their experiments
- Tools for knowledge graph creation
- Creating a quadrant for questions
- Evaluating the results
- System architecture
- Cypher Query Language
- Investing in KGs will pay off
- How to convince your organization to invest in KGs
- RDF
- Next steps and testing generalizability of the results
- RAG for hybrid (structured and unstructured) data sources
Related content:
- A video version of this conversation is available on our YouTube channel.
- Boosting LLMs with External Knowledge: The Case for Knowledge Graphs
- Philipp Moritz and Goku Mohandas: Navigating the Nuances of Retrieval Augmented Generation
- Emil Eifrem: The Future of Graph Databases
- Sudhir Hasbe: Trends in Data Management: From Source to BI and Generative AI
- Semih Salihoglu: Kùzu – A simple, extremely fast, and embeddable graph database
- Waleed Kadous: Best Practices for Building LLM-Backed Applications
- Michele Catasta: Software Development with AI and LLMs
If you enjoyed this episode, please support our work by encouraging your friends and colleagues to subscribe to our newsletter: