Providing contextual information associated with a source document using information from external reference documents

• US 9,965,459 B2
• Filed: 09/18/2014
• Issued: 05/08/2018
• Est. Priority Date: 08/07/2014
• Status: Active Grant

• Current Assignee: Accenture Global Services Limited (Accenture PLC)
• Original Assignee: Accenture Global Services Limited (Accenture PLC)
• Inventors: Sengupta, Shubhashis, Karukapadath Mohamedrasheed, Annervaz, Pathak, Neetu
• Primary Examiner(s): Corrielus, Jean M
• Application Number:
  US14/489,693
• Publication Number:
  US 20160042061A1
• Time in Patent Office: 1,328 Days
• Field of Search: 707738, 707739, 707741
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A device may receive a source document to be processed for contextual information associated with named entities included in the source document. The device may identify a named entity included in the source document, and may identify a context of the source document. The device may identify a reference document associated with the named entity. The reference document may be different from the source document. The device may perform a semantic similarity analysis based on the context of the source document and further based on the reference document. The device may identify contextual information, included in the reference document, based on performing the semantic similarity analysis. The contextual information may relate to the context of the source document. The device may provide the contextual information.

Claims

1. A device, comprising:

one or more processors to;

receive a source document to be processed for contextual information associated with one or more named entities included in the source document;

provide, for display on a representation of the source document on a user interface, a first input mechanism for a user;

identify, based on a first user interaction with the first input mechanism, a first named entity, of the one or more named entities, included in the source document;

identify, based on a second user interaction with the first input mechanism, a context of the source document by using context terms, of the source document, that are different than the first named entity;

provide the first named entity as a search query;

identify a first reference document based on providing the first named entity as the search query,

the first reference document being associated with a result of the search query, and

the first reference document being different from the source document;

identify a second named entity, of the one or more named entities, included in the source document;

identify a second reference document associated with the second named entity;

analyze the first reference document and the second reference document;

classify the first named entity as a primary entity based on analyzing the first reference document and the second reference document;

classify the second named entity as a secondary entity based on analyzing the first reference document and the second reference document;

perform a semantic similarity analysis based on the context of the source document and based on classifying the first named entity as the primary entity and the second named entity as the secondary entity;

provide, for display on the user interface, a second input mechanism for the user to cause contextual information to be provided; and

identify contextual information, associated with the source document, based on performing the semantic similarity analysis and based on a third user interaction with the second input mechanism,

the contextual information including one or more reference text sections having a threshold semantic similarity score with respect to the secondary entity and the context of the source document.

2. The device of claim 1, where the one or more processors, when performing the semantic similarity analysis, are to:

generate a semantic similarity score for a relationship between the context of the source document and a text section included in the first reference document; and

where the one or more processors, when identifying the contextual information, are to;

identify the text section as contextual information based on the semantic similarity score.

3. The device of claim 2, where the semantic similarity score indicates a degree of semantic relatedness between the context of the source document and the text section.

4. The device of claim 1, where the one or more processors are further to:

provide, for display, the contextual information in association with the source document.

5. The device of claim 1, where the one or more processors are further to:

provide, for display, an indication of a relationship between the contextual information and the first named entity.

6. The device of claim 1, where the one or more processors are further to:

calculate one or more relevance scores for the first named entity; and

where, when classifying the first named entity as the primary entity, the one or more processors are to;

classify the first named entity as the primary entity based on calculating the one or more relevance scores for the first named entity.

7. The device of claim 1, where the one or more processors are further to:

calculate one or more relevance scores for the second named entity; and

where, when classifying the second named entity as the secondary entity, the one or more processors are to;

classify the second named entity as the secondary entity based on calculating the one or more relevance scores for the second named entity.

8. A computer-readable medium storing instructions, the instructions comprising:

one or more instructions that, when executed by one or more processors, cause the one or more processors to;

receive a source document to be processed for contextual information relating to the source document;

provide, for display on a representation of the source document on a user interface, a first input mechanism for a user;

identify, based on a first user interaction with the first input mechanism, a first named entity included in the source document;

identify, based on a second user interaction with the first input mechanism, a context associated with the source document by using context terms, of the source document, that are different than the first named entity;

provide the first named entity as a search query;

identify a first reference document based on providing the first named entity as the search query,

the first reference document being associated with a result of the search query, and

the first reference document being different from the source document;

identify a second named entity included in the source document;

identify a second reference document associated with the second named entity;

analyze the first reference document and the second reference document;

classify the first named entity as a primary entity based on analyzing the first reference document and the second reference document;

classify the second named entity as a secondary entity based on analyzing the first reference document and the second reference document;

perform a semantic similarity analysis using the context associated with the source document and based on classifying the first named entity as the primary entity and the second named entity as the secondary entity;

provide, for display on the user interface, a second input mechanism for the user to cause contextual information to be provided; and

identify contextual information, based on performing the semantic similarity analysis and based on a third user interaction with the second input mechanism,

the contextual information including one or more reference text sections having a threshold semantic similarity score with respect to the secondary entity and the context associated with the source document, and not being included in the source document.

9. The computer-readable medium of claim 8, where the one or more instructions, that cause the one or more processors to perform the semantic similarity analysis, cause the one or more processors to:

generate a semantic similarity score for a relationship between the context associated with the source document and reference information included in the first reference document or the second reference document;

where the one or more instructions, when executed by the one or more processors, further cause the one or more processors to;

determine that the semantic similarity score satisfies the threshold semantic similarity score; and

where the one or more instructions, that cause the one or more processors to identify the contextual information, cause the one or more processors to;

identify the reference information as contextual information based on determining that the semantic similarity score satisfies the threshold.

10. The computer-readable medium of claim 8,

where the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:

classify the first named entity as the primary entity based on a quantity of times that the first named entity is included in the second reference document.

11. The computer-readable medium of claim 10, where the one or more instructions, that cause the one or more processors to classify the first named entity, cause the one or more processors to:

classify the first named entity as the primary entity further based on a quantity of times that the second named entity is included in the first reference document.

12. The computer-readable medium of claim 8,

where the one or more instructions, that cause the one or more processors to identify the contextual information, cause the one or more processors to:

identify the contextual information based on a plurality of relationships between the first named entity and the second named entity.

13. The computer-readable medium of claim 8, where the one or more instructions, that cause the one or more processors to perform the semantic similarity analysis, cause the one or more processors to:

perform at least one of;

an Adapted Lesk algorithm, or

a Jiang-Conrath algorithm.

14. The computer-readable medium of claim 8, where the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:

calculate a relevance score for the first named entity and a relevance score for the second named entity; and

where the one or more instructions, that cause the one or more processors to classify the first named entity as the primary entity, cause the one or more processors to;

classify the first named entity as the primary entity based on the first named entity having a higher relevance score than the second named entity.

15. A method, comprising:

receiving, by a device, a source document to be processed for contextual information relating to the source document;

providing, by the device and for display on a representation of the source document on a user interface, a first input mechanism for a user;

identifying, by the device and based on a first user interaction with the first input mechanism, a first named entity included in the source document;

determining, by the device and based on a second user interaction with the first input mechanism, a context associated with the source document by using context terms, of the source document, that are different than the first named entity;

providing, by the device, the first named entity as a search query;

receiving, by the device, a first reference document based on providing the first named entity as the search query,

the first reference document being associated with a result of the search query, and

the first reference document being different from the source document;

identifying, by the device, a second named entity included in the source document;

identifying, by the device, a second reference document associated with the second named entity;

analyzing, by the device, the first reference document and the second reference document;

classifying, by the device, the first named entity as a primary entity based on analyzing the first reference document and the second reference document;

classifying, by the device, the second named entity as a secondary entity based on analyzing the first reference document and the second reference document;

performing, by the device, a semantic similarity analysis based on the context associated with the source document and based on classifying the first named entity as the primary entity and the second named entity as the secondary entity; and

identifying, by the device, contextual information, associated with the source document, based on performing the semantic similarity analysis and based on a third user interaction with a second input mechanism,

the contextual information including one or more reference text sections having a threshold semantic similarity score with respect to the secondary entity and the context associated with the source document.

16. The method of claim 15, further comprising:

identifying a third named entity included in the source document,

the third named entity being different from the first named entity and the second named entity; and

identifying a text section, included in the source document, that includes the third named entity; and

where performing the semantic similarity analysis comprises;

performing the semantic similarity analysis using the text section that includes the third named entity.

17. The method of claim 15, further comprising:

identifying a third named entity included in the source document,

the third named entity being different from the first named entity and the second named entity; and

receiving a third reference document associated with the third named entity,

the third reference document being different from the first reference document, the second reference document, and the source document; and

where performing the semantic similarity analysis comprises;

performing the semantic similarity analysis using the third reference document.

18. The method of claim 15, where performing the semantic similarity analysis comprises:

determining a source text section, included in the source document, that is associated with the context associated with the source document;

determining a reference text section, included in the first reference document and the second reference document, that is associated with the context associated with the source document; and

determining a score associated with the source text section and the reference text section.

19. The method of claim 15,

where the method further comprises:

providing contextual information associated with a relationship between the first named entity and the second named entity.

20. The method of claim 15, where performing the semantic similarity analysis comprises:

generating a semantic similarity score associated with the first named entity and the second named entity; and

where the method further comprises;

providing a visual indication of the semantic similarity score.