Triple

A triple is the atomic unit of knowledge representation in knowledge graphs, consisting of three components: a subject (the entity being described), a predicate (the relationship or property), and an object (the related entity or value). For example, (Alice Smith, WORKS_AT, Acme Corp) states that Alice Smith works at Acme Corp.

Triples are the foundation of the Resource Description Framework (RDF), the W3C standard for representing information on the semantic web. In property graph databases like Neo4j, the same concept is expressed using node-relationship-node patterns. Triple extraction from unstructured text is a key step in knowledge graph construction, typically performed using NER followed by relation extraction.

Triples can represent both relationships between entities (Alice, MANAGES, Bob) and properties of entities (Alice, AGE, 30). Collections of triples form knowledge graphs, where subjects and objects become nodes and predicates become edges. The simplicity and composability of triples makes them a powerful primitive for building arbitrarily complex knowledge structures.