Knowledge graph

A structured representation of entities and the relationships among them. It encodes information as connected data, often using labeled nodes and edges.

Entity: A distinct, identifiable object, person, place, event, or idea represented in a system. Entities are the things represented as nodes in a knowledge graph.
Graph database: A database designed to store and query data as nodes and relationships. Graph databases store connected entities in a form suited to traversing their relationships.
Google Knowledge Graph: Google's knowledge base of entities and relationships, used to enhance search results. It illustrates how a large graph can supply structured facts to search.
Relational database: A database that organizes data into tables of rows and columns linked through keys. Relational tables encode connections differently from a graph's explicit edges.
Relation: A connection or association between two or more entities. Relations form the links that give a graph its structure and meaning.
Knowledge graph embedding: A method that maps graph entities and relations into vectors for machine-learning tasks. Embeddings turn graph structure into numerical representations that models can use.
Wikidata: A collaboratively edited knowledge base that stores structured data in a multilingual format. Wikidata is a prominent open graph of entities and their statements.
Taxonomy: A hierarchical classification system that organizes items into named categories. A taxonomy mainly represents category hierarchy, while knowledge graphs can encode many relation types.
Knowledge representation: The methods used to encode information so that computers can store, interpret, and reason with it. Knowledge graphs are one way to represent structured information for computation.
Entity linking: The task of matching a mention in text to the entity it denotes in a knowledge base. Entity linking connects textual mentions to the graph's canonical entities.

Linked from 11 pages