Catalog & Cocktails: The Honest, No-BS Data Podcast Ontologies Aren’t New with Oscar Corcho
8 snips
Jan 22, 2026 In this engaging discussion, Professor Oscar Corcho, an expert in ontology engineering at the Technical University of Madrid, unveils the rich history and evolution of ontologies. He traces their roots in AI from the 1990s, emphasizes the transition from prototypes to practical applications, and showcases their role in semantic interoperability, like harmonizing railway data across Europe. Oscar also shares insights on starting ontology projects and how using LLMs can enhance the ontology development process, urging teams to treat ontologies as vital software products.
AI Snips
Chapters
Books
Transcript
Episode notes
Origins In The Knowledge Acquisition Problem
- Ontologies originated in the 1990s to solve the knowledge acquisition bottleneck for early AI systems.
- The goal was reusable, modular knowledge representations to avoid re-encoding domain expertise repeatedly.
Maturation Through Standards And Tools
- Ontology work matured through standards (OWL), methodologies, and better tools, mirroring how data management evolved around SQL.
- This maturation enabled ontologies to move from toy prototypes to robust artifacts used in production systems.
Data-Center Root-Cause Use Case
- Oscar described building a network of ontologies for a data-center company to map assets, virtual services, and logs.
- They generated knowledge graphs from logs to run queries that pinpoint root causes of failures across complex infrastructure.
