Knowledge Bases are large repositories of structured or unstructured data for use within an information system. The term was originally coined to differentiate between databases, which at the time held flat, tabular data, and new knowledge bases with relational and hierarchical data. Today databases can be as multifaceted as knowledge bases of old. With the connotation that today’s knowledge bases are that much closer to semantic and organizationally multifaceted data.
While some researchers use the terms knowledge base and knowledge graph interchangeably, the use of the term graph is intentional and there are distinctions. All knowledge graphs are knowledge bases, while not every knowledge base qualifies as a knowledge graph. The key differentiator between knowledge graphs and bases is that graphs are centered around the relationships between entities. Graphs are mutable as well as semantic, meaning the meaning of data is encoded alongside the data. This provides huge possibilities for machine learning in automated knowledge bases or knowledge graphs such as Diffbot’s Knowledge Graph™.