A knowledge graph stores information as nodes representing entities such as people, organizations, events, and documents, linked by typed relationships describing how they relate.
Unlike text retrieval, which returns passages, a graph supports traversal, so a system can follow relationships across many documents to assemble an answer.
Alternative Names:
Semantic Graph, Entity Graph
Why it Matters?
Certain litigation questions are structural rather than textual. Identifying every document a particular custodian touched regarding a specific issue, or tracing how an assertion propagated across witnesses, requires following relationships rather than matching text. Graph structure is what makes those questions answerable without reading everything.
Frequently Confused with
Related terms
Frequently asked questions
How is a knowledge graph different from a database?
Why combine a graph with retrieval?
