Agnostic creation, version control, and contextual query of knowledge graph

Patent metadata

Field	Value
Publication number	US-2021304021-A1
Application number	US-202016831391-A
Country	US
Kind code	A1
Filing date	Mar 26, 2020
Priority date	Mar 26, 2020
Publication date	Sep 30, 2021
Grant date	—

Abstract

A lifecycle platform for creation, ingestion, version control, and contextual query of knowledge graph is disclosed. Such a platform may be used to create and deploy a knowledge graph by reusing and merging knowledge defined in existing and validated data models. The platform tracks changes made to the knowledge graph after being deployed and provides version tracking of the knowledge graph and its underlying namespaces. The platform further provides a subscribable service for contextual viewing and query of portions and/or subset versions of the knowledge graph. Such a platform may be provided as an agnostic plugin to a specific vendor knowledge graph solution space.

First claim

What is claimed is:

1 . A method including:
automatically extracting a first namespace from a first data model and a second namespace from a second data model independent of the first data model; recording the first namespace and the second namespace in a namespace repository; generating a composite data model dependent on the first namespace and the second namespace; automatically generating an agnostic schema from the composite data model and the first namespace and the second namespace; automatically launching and deploying a composite knowledge graph based on the agnostic schema; automatically back-extracting different versions of the composite knowledge graph at different times as the composite knowledge graph is modified while deployed; automatically determining changes made in the different versions of the composite knowledge graph at the different times; automatically extracting and generating different versions of the first namespace and the second namespace at the different times corresponding to the different versions of the composite knowledge graph based on the changes; and tracking the different versions of the first namespace and the second namespace in the namespace repository.

2 . The method of claim 1 , wherein the first data model comprises a first schema associated with a first knowledge graph and a second schema associated with a second knowledge graph independent of the first knowledge graph.

3 . The method of claim 2 , wherein the first schema associated with the first knowledge graph and the second schema associated with the second knowledge graph are expressed using a first graphic data description framework and a second graphic description framework distinct from the first graphic data description framework.

4 . The method of claim 1 , wherein the first data model and the second data model belong to a model library containing a plurality of pre-validated domain data models.

5 . The method of claim 1 , wherein the composite data model inherits at least a portion of the first namespace and the second namespace.

6 . The method of claim 5 , wherein the composite data model comprises definitions of graph nodes, graph edges, or relationships between the graph nodes.

7 . The method of claim 5 , wherein:
the composite data model comprises a definition of a graph node based on the first namespace; and the definition of the graph node comprises a property definition inherited from the second namespace.

8 . The method of claim 1 , wherein automatically generating the agnostic schema based on the composite data model and the first namespace and the second namespace comprises:
determining a plurality of graph nodes or graph edges based on the composite data model and the first namespace and the second namespace; identifying a namespace collision between a first graph node and a second graph node among the plurality of graph nodes; resolving the namespace collision by automatically renaming the second graph node according to a predefined naming rule; and including definitions of the plurality of graph nodes in the agnostic schema with the second graph node renamed.

9 . The method of claim 8 , wherein generating the agnostic schema based on the composite data model and first namespace and the second namespace further comprises:
determining a plurality of graph relationships between the plurality of graph nodes; modifying the plurality of graph relationships according to the renamed second graph node; and including definitions of the modified plurality of the graph relationships in the agnostic schema.

10 . The method of claim 1 , further comprising providing a contextual service for viewing and querying the composite knowledge graph to a user.

11 . The method of claim 10 , wherein the contextual service for viewing and querying the composite knowledge graph is based on a user service subscription.

12 . The method of claim 11 , wherein the user service subscription comprises a subscription to selected portions and/or selected versions of the composite knowledge graph.

13 . The method of claim 12 , wherein the selected portions of the composite knowledge graph corresponds to one of the first data model and the second data model.

14 . The method of claim 12 , further comprising receiving a query from the user;
searching a subset of the composite knowledge graph determine by the user service subscription to obtain a query result; and curating a view of the query result to the user via a user interface on a user terminal device.

15 . The method of claim 1 , wherein automatically launching and deploying the composite knowledge graph based on the agnostic schema comprises:
automatically converting the agnostic schema into a converted schema of a predefined graphic data description framework; and automatically launching the composite knowledge graph according to the converted schema.

16 . The method of claim 15 , wherein:
the automatically back-extracting the different versions of the composite knowledge graph, automatically determining changes made in the different versions of the composite knowledge graph, and automatically extracting and generating different versions of the first namespace and the second namespace are performed in a manner agnostic to the predefined graphic data description framework.

17 . A computer system comprising processors and memories for storing computer instructions, where the processors are configured to execute the computer instructions to:
automatically extract a first namespace from a first data model and a second namespace from a second data model independent of the first data model; record the first namespace and the second namespace in a namespace repository; generate a composite data model dependent of the first namespace and the second namespace; automatically generate an agnostic schema from the composite data model and the first namespace and the second namespace; automatically launch and deploy a composite knowledge graph based on the agnostic schema; automatically back-extract different versions of the composite knowledge graph at different times as the composite knowledge graph is modified while deployed; automatically determine changes made in the different versions of the composite knowledge graph at the different times; automatically extract and generate different versions of the first namespace and the second namespace at the different times corresponding to the different versions of the composite knowledge graph based on the changes; and track the different versions of the first namespace and the second namespace in the namespace repository.

18 . The computer system of claim 17 , wherein to automatically generate the agnostic schema based on the composite data model and the first namespace and the second namespace, the processors are configured to execute the computer instructions to:
determine a plurality of graph nodes or graph edges based on the composite data model and the first namespace and the second namespace; identify a namespace collision between a first graph node and a second graph node among the plurality of graph nodes; resolve the namespace collision by automatically renaming the second graph node according to a predefined naming rule; and include definitions of the plurality of graph nodes in the agnostic schema with the second graph node renamed.

19 . The computer system of claim 17 , wherein to automatically launch the deploy the composite knowledge graph based on the agnostic sch

Assignees

• Accenture Global Solutions Ltd

Inventors

• Puri Colin Anil
• Vasquez Reymonrod Geli
• Kujawinski Matthew
• Tung Teresa Sheausan

Classifications

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06F16/9024 Primary

  Graphs; Linked lists (G06F16/9027 takes precedence) · CPC title

• G06F16/213

  with details for schema evolution support · CPC title

• G06N5/022 Primary

  Knowledge engineering; Knowledge acquisition · CPC title

• G06F16/2458

  Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries · CPC title