Knowledge Fusion is a crucial and differentiating step within Diffbot’s Knowledge Graph™ build pipeline. Occuring after the linking of records, knowledge fusion merges all origins of data for a given entity together. Subsequently, knowledge fusion is the stage at which the Knowledge Graph™ chooses which data (and from which source) gets pulled into a finalized entity.
Examples of knowledge fusion — sometimes called entity resolution — would include reconciling where a person entity is currently employed if there are multiple competing sources of this fact online. Discerning the most up-to-date web page of an organization, employee count, market caps of publicly-traded companies, and any fact that may change over time are other examples of what may be reconciled within knowledge fusion.