Taxonomies, Vocabularies, and Ontologies
Choosing the right level of structure for a domain, from a flat controlled vocabulary to a formal ontology, and knowing when to stop.
Not every system needs an ontology. Many need a good controlled vocabulary and a clear taxonomy. This guide lays out the spectrum of knowledge structures, what each one can and cannot express, and how to pick the lightest structure that supports the decisions your AI system has to make.
What you will be able to do
- Explain the difference between vocabularies, taxonomies, thesauri, schemas, and ontologies.
- Match a knowledge structure to the questions a system must answer.
- Grow a structure incrementally without rebuilding it.
Inside the guide
- 01
The spectrum of structure
From word lists to formal logic.
- 02
Controlled vocabularies
Terms, synonyms, and the value of saying one thing one way.
- 03
Taxonomies and thesauri
Hierarchy, broader and narrower terms, and faceting.
- 04
Schemas and ontologies
Classes, properties, constraints, and inference.
- 05
Choosing and growing
Starting light and adding formality when a question demands it.
From the guide
Structure has a cost that is paid every day by whoever maintains it. The right amount is the least that still answers the questions your system is asked.
Who it is for
- Information architects and librarians entering AI projects.
- Engineers who have been told to 'build an ontology' and are not sure they should.
More on Ontologies
Topic overview- Ontologies
Ontologies for AI Agents
How to define the entities, relationships, and rules an agent reasons over, and how to hand that definition to the agent at runtime.