KGcreator is written in Common Lisp (a research version written in Haskell is available by special request) and is available here for live demos and also in three editions of a commercial product:
KGcreator was written by Mark Watson and uses both classical symbolic natural language processing (NLP) techniques and deep learning models for coreference (anaphora resolution) and link analysis.
Data created by KGcreator is stored internally in an generalized data structure and rendered for use in two formats:
Screen shot of part of an automatically generated graph in Neo4j:
Knowledge Graphs are implemented as directed graphs where nodes can represent people, locations, documents, organizations, business processes, documents on the web, etc. Graph edges represent property relations between nodes. Both nodes and edges can contain named properties. KGcreator generates graph data in two formats for both Neo4J and RDF/OWL linked data sources.
I worked as a contractor at Google on their Knowledge Graph. Facebook also has a Knowledge Graph to store information on users and connections.
Here is some reading material:
Privacy statement: this site KGcreator.com collects no personal data.