How AI Is Built  cover image

How AI Is Built

#043 Knowledge Graphs Won't Fix Bad Data

Feb 20, 2025
Juan Sequeda, a Principal Scientist at data.world and an authority on knowledge graphs, shares his insights on improving data quality. He discusses the importance of integrating technical and business metadata to create a 'brain' for AI applications. Sequeda explains how traditional silos hinder effective data management and emphasizes the need for collaboration in startups. He also addresses the balance between automation and human oversight in knowledge graphs and outlines strategies for defining robust entities and relationships, ensuring accurate data connections.
01:10:59

Episode guests

Podcast summary created with Snipd AI

Quick takeaways

  • Metadata is essential for constructing enterprise knowledge graphs, providing the necessary context and structure for effective data management and AI applications.
  • The rise of knowledge graphs addresses the challenge of data silos by enabling better integration and usability of discrete data sources across organizations.

Deep dives

Understanding Knowledge Graphs and Metadata

Knowledge graphs have gained significant attention, yet they are built upon foundational concepts such as ontologies and taxonomies, which have been researched for decades. Metadata serves as the backbone of these structures, providing context by defining data meanings, origins, and interrelations. The resurgence in interest is largely attributed to the increasing need for artificial intelligence (AI) systems to have contextual understanding, as traditional large language models (LLMs) struggle without clear definitions and connections. These elements underscore the relevance of knowledge graphs in enhancing data usability across AI and business intelligence applications.

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner