Entity Resolution refers to a portion of a knowledge graph build process in which information from separate records is reconciled to create a single set of facts that comprise an entity. Additionally, successful entity resolution tackles many time consuming challenges within the data cleaning process such as disambiguating entities, ensuring a lack of overmerging, and verifying a coherency to facts about an entity.
A more concrete example of entity resolution could involve the discernment of which John Smith is being referenced within an article based on the surrounding context. Additionally, if more than one source of information is contradictory, a truthiness score will be assessed to each in an attempt to pull together one authoritative entity.