CONSTRUCTING A KNOWLEDGE GRAPH EMPLOYING MULTIPLE SUBGRAPHS AND A LINKING LAYER INCLUDING MULTIPLE LINKING NODES

摘要

A graph-based data storage and retrieval system in which multiple subgraphs representing respective datasets in different namespaces are interconnected via a linking or "canonical" layer. Respective datasets represented by subgraphs in different namespaces may pertain to a particular information domain, and may include heterogeneous datasets. The canonical layer provides for a substantial reduction of graph complexity required to interconnect respective corresponding nodes in different subgraphs, which in turn offers distinct advantages as the number of subgraphs increases for the particular domain(s) of interest. Examples of such advantages include reductions in data storage and retrieval times, enhanced query/search efficacy and discovery of relationships in different parts of the system, enhanced ability to infer relationships in different parts of the system, and enhanced ability to train data models for natural language processing and other purposes based on information extracted from the system.