Knowledge Graph
Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary
A network of entities and their relationships, stored as nodes and edges, that lets machines answer questions by following connections between facts.
What is a Knowledge Graph?
A knowledge graph is a network of entities and their relationships, stored as nodes and edges, that lets a machine answer questions by following connections between facts.
It differs from an ordinary database in what it makes central. A table stores rows about one kind of thing; a knowledge graph stores the links between things, typed and queryable, so “find Sci-Fi films directed by Christopher Nolan” is a walk along edges rather than a join across tables.
How a Knowledge Graph Works
The basic unit is a triple: subject, relationship, object. “Inception – directedBy – Christopher Nolan” is one fact; millions of such triples interlink into a graph where every entity is reachable from its neighbors.
Entities and relationships carry attributes (a film’s release year, a relationship’s date range). An ontology sits on top, defining which entity types and relationship types exist, so the graph stays consistent as data flows in from different sources. This structure makes knowledge graphs the most visible modern descendant of knowledge representation.
Applications of Knowledge Graphs
- Search: Graphs let a search engine answer entity questions (“Who directed Inception?”) directly instead of returning pages that mention the words.
- Recommendations: Users, products, and their interactions form a natural graph; paths through it surface items connected to what someone already likes.
- Grounding language models: A knowledge graph supplies verified facts for a model to draw on at answer time, a graph-shaped variant of RAG that reduces invented answers.
Example of a Knowledge Graph
Google introduced its Knowledge Graph in 2012 with the slogan “things, not strings.” Before it, a search for “Marie Curie” matched pages containing those characters; after it, the query resolved to an entity with typed connections to her discoveries, her Nobel Prizes, and her family.
Ask “How tall is the Eiffel Tower?” and the answer comes straight off the entity’s attributes, no web page required. The information panels beside Google results are the graph made visible.
The design shift is the point of the example: once facts are stored as connected entities rather than indexed text, the system answers questions instead of retrieving documents.
Related AI terms: Knowledge Representation · RAG · Graph Neural Network · Natural Language Processing
Did you like the Knowledge Graph 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