Knowledge Representation
Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary
The branch of AI concerned with encoding facts about the world in structured forms a machine can reason over, from if-then rules to ontologies.
What is Knowledge Representation?
Knowledge representation (KR) is the branch of AI concerned with encoding facts about the world in structured forms a machine can store, retrieve, and reason over.
KR dominated AI’s first decades, when intelligence was pursued by writing knowledge down rather than learning it from data. Machine learning has displaced it as the default approach, but explicit knowledge structures persist wherever answers must be verifiable, most visibly in the knowledge graph.
Types of Knowledge Representation
- Logical representation states facts and relationships in formal logic, which permits rigorous, provable inference at the cost of expressive flexibility.
- Semantic networks encode concepts as nodes and relationships as edges. They are the direct ancestors of today’s knowledge graphs.
- Frames bundle everything known about a type of thing into a structure of slots and values: a “patient” frame holds age, symptoms, history.
- Rules express knowledge as if-then statements, the format behind the classic expert system.
- Ontologies define what kinds of entities exist in a domain and how they relate, so separate systems interpret the same terms the same way.
Each format trades expressiveness against tractability. Richer representations describe more of the world; leaner ones let the machine reason faster and more reliably.
Example of Knowledge Representation
MYCIN, built at Stanford in the 1970s, encoded infectious-disease expertise as several hundred if-then rules, such as: if the organism is gram-positive and grows in chains, then there is suggestive evidence it is streptococcus. Each rule carried a certainty factor, so evidence accumulated rather than being simply true or false.
Given a patient’s lab results, MYCIN chained through its rules to identify the likely organism and recommend an antibiotic dose. In evaluations its recommendations compared well with those of infectious-disease specialists.
It never entered routine clinical use, but its architecture became the template for a generation of expert systems. The rules were the knowledge representation: expertise made explicit, inspectable, and executable, which is exactly what learned models today do not give you for free.
Related AI terms: Knowledge Graph · Expert System · Fuzzy Logic · Natural Language Processing
Did you like the Knowledge Representation gist?
Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.
Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About