RAG 5 Jan 2026

Knowledge Graphs - Ontology

Ontology or knowledge graphs is basically a bunch of data interconnected with some meaning. It defines relationship between objects in that data.

When do you need Ontology

  1. Understanding data
  2. Reusability
  3. Explainability

It started with keyword search/inverted indexes in the 90s, until it didn't work. Then we got Google's Page Rank alogithm. Google then went on to use Knowledge graphs in 2012. Google in 2024 is now using LLM + GraphRAG