Constructing enterprise-specific knowledge graphs

Patent metadata

Field	Value
Publication number	US-10915577-B2
Application number	US-201815928288-A
Country	US
Kind code	B2
Filing date	Mar 22, 2018
Priority date	Mar 22, 2018
Publication date	Feb 9, 2021
Grant date	Feb 9, 2021

Abstract

A framework is provided for constructing enterprise-specific knowledge bases from enterprise-specific data that includes structured and unstructured data. Relationships between entities that match known relationships are identified for each of a plurality of tuples included in the structured data. Where possible, relationships between entities that match known relationships also are identified for tuples included in the unstructured data. If matching relationships between entities that cannot be identified for tuples in the unstructured data, extracted relationships are sequentially clustered to similar relationships and a relationship is assigned to the clustered tuples. An enterprise-specific knowledge graph is constructed from the structured-data-tuples and their identified relationships, the unstructured-data-tuples where the relationships could be mapped to a known relationship and their identified relationships, and the unstructured-data-tuples that could not be mapped to a known relationship and their assigned relationships. The knowledge graph is enriched with any information determined to be missing therefrom.

First claim

What is claimed is:

1. A computer-implemented method for constructing a knowledge graph, the method comprising:
receiving data, a first portion of the data being structured data and a second portion of the data being unstructured data, the structured data having a plurality of text fields, each of the plurality of text fields having a corresponding value such that the structured data includes a plurality of text-field/value pairs;
generating a first plurality of tuples based on the plurality of text-field/value pairs from the structured data, each tuple from the first plurality of tuples including a relationship corresponding to each text-field/value pair identified as matching one of a plurality of predefined relationships included in a predefined relationship taxonomy;
generating a second plurality of tuples based on a first plurality of relationships extracted from the unstructured data and identified as matching one of the plurality of predefined relationships included in the predefined relationship taxonomy;
generating a third plurality of tuples based on a second plurality of relationships extracted from the unstructured data and identified by:
clustering relationships from the second plurality of relationships such that similar relationships are grouped together; and
attributing an assigned relationship to at least part of the tuples from the third plurality of tuples based upon the clustering; and

constructing the knowledge graph from, at least in part, the first plurality of tuples, the second plurality of tuples, and the third plurality of tuples.

2. The computer-implemented method of claim 1 , further comprising enriching the knowledge graph by adding a fourth plurality of tuples that is determined to be missing based upon existing patterns.

3. The computer-implemented method of claim 1 , wherein each tuple of the first plurality of tuples has a <subject> element, a <predicate> element, and an <object> element arranged in a <subject><predicate><object> format.

4. The computer-implemented method of claim 3 , wherein within a given tuple, the <subject> element and the <object> element refer to entities and the <predicate> element refers to a correlation between the entities referred to by the <subject> and <object> elements.

5. The computer-implemented method of claim 4 , wherein the relationship corresponding to each text-field/value pair of the first plurality of text-field/value pairs of the structured data is identified based on a particular predefined relationship of the plurality of predefined relationships included in the predefined relationship taxonomy that matches the <predicate> element of each tuple from the first plurality of tuples.

6. The computer-implemented method of claim 1 , wherein each tuple of the second plurality of tuples and the third plurality of tuples has a <subject> element, a <predicate> element, and an <object> element arranged in a <subject><predicate><object> format.

7. The computer-implemented method of claim 6 , wherein within a tuple, the <subject> element and the <object> element refer to entities and the <predicate> element refers to a correlation between the entities referred to by the <subject> and <object> elements.

8. The computer-implemented method of claim 7 , wherein the relationship corresponding to each tuple from the second plurality of tuples extracted from the unstructured data is identified based on a particular predefined relationship of the plurality of predefined relationships included in the predefined relationship taxonomy that matches the <predicate> element of each tuple from the second plurality of tuples.

9. The computer-implemented method of claim 1 , wherein the one of the plurality of predefined relationships included in the predefined relationship taxonomy that matches each s relationship from the first plurality of relationships extracted from the unstructured data is identified utilizing one or more of semantic mapping, syntactic mapping, and pattern-based mapping.

10. One or more computer storage media storing computer-useable instructions that, when executed by one or more processors, cause the one or more processors to perform a method for constructing a knowledge graph, the method comprising:
receiving data from a repository, a first portion of the data being structured data and a second portion of the data being unstructured data;
generating a plurality of structured-data tuples from the structured data, each structured-data tuple including a structured-data-tuple <subject> element, a structured-data-tuple <predicate> element, and a structured-data-tuple <object> element arranged in a <subject><predicate><object> format;
identifying a relationship corresponding to each of the plurality of structured-data tuples that matches one of a plurality of predefined relationships included in a predefined relationship taxonomy;
extracting a plurality of unstructured-data tuples from the unstructured data, each unstructured-data tuple including an unstructured-data-tuple <subject> element, an unstructured-data-tuple <predicate> element, and an unstructured-data-tuple <object> element arranged in the <subject><predicate><object> format;
determining that each unstructured-data tuple of a first portion of the plurality of unstructured-data tuples refers to a relationship that matches one of the plurality of predefined relationships included in the predefined relationship taxonomy and each unstructured-data tuple of a second portion of the plurality of unstructured-data tuples refers to a relationship that does not match one of the plurality of predefined relationships included in the predefined relationship taxonomy;
identifying one of the plurality of predefined relationships included in the predefined relationship taxonomy that matches each unstructured-data tuple of the first portion of the plurality of unstructured-data tuples;
clustering the relationships referenced by the second portion of the plurality of unstructured-data tuples such that similar relationships are grouped together;
attributing an assigned relationship to at least part of the unstructured-data tuples of the second portion of the plurality of unstructured-data tuples based upon the clustering;
constructing the knowledge graph from, at least in part, the structured-data tuples of the plurality of structured-data tuples and their respective identified relationships, the unstructured-data tuples of the first portion of the plurality of unstructured-data tuples and their respective identified relationships, and at least part of the unstructured-data tuples of the second portion of the plurality of unstructured-data tuples and their respective assigned relationships; and
enriching the knowledge graph by adding a plurality of tuples that is determined to be missing from the knowledge graph based upon existing patterns.

11. The one or more computer storage media of claim 10 , wherein within a given structured-data tuple, the structured-data-tuple <subject> element and the structured-data-tuple <object> element refer to entities and the structured-data-tuple <predicate> element refers to a correlation between the entities referred to by the structured-data-tuple <subject> and <object> elements.

12. The one or more computer storage media of claim 11 , wherein identifying the relationship corresponding to each of the plurality of structured-data tuples that matches one of the plurality of predefined relationships included in the predefined relationship taxonomy comprises identifying a particular predefined relationship of the plurality of predefined relationships included in the predefined relationship taxonomy that matches the structured-data-tu

Assignees

• Adobe Inc

Inventors

• Srinivasan Balaji Vasan
• Chaturvedi Rajat
• Goyal Tanya
• Maheshwari Paridhi
• Monsy Anish Valliyath
• Sancheti Abhilasha

Classifications

• G06F16/9024 Primary

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

• G06F16/36 Primary

  Creation of semantic tools, e.g. ontology or thesauri · CPC title

• G06F16/81

  Indexing, e.g. XML tags; Data structures therefor; Storage structures · CPC title

• G06F16/367 Primary

  Ontology · CPC title

• G06F16/907

  Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually · CPC title