System and method for automatically detecting and interactively displaying information about entities, activities, and events from multiple-modality natural language sources

Patent metadata

Field	Value
Publication number	US-10698964-B2
Application number	US-201715419615-A
Country	US
Kind code	B2
Filing date	Jan 30, 2017
Priority date	Jun 11, 2012
Publication date	Jun 30, 2020
Grant date	Jun 30, 2020

Abstract

A method for automatically extracting and organizing information by a processing device from a plurality of data sources is provided. A natural language processing information extraction pipeline that includes an automatic detection of entities is applied to the data sources. Information about detected entities is identified by analyzing products of the natural language processing pipeline. Identified information is grouped into equivalence classes containing equivalent information. At least one displayable representation of the equivalence classes is created. An order in which the at least one displayable representation is displayed is computed. A combined representation of the equivalence classes that respects the order in which the displayable representation is displayed is produced.

First claim

What is claimed is:

1. A method for processing information by a processing device, the method comprising:
receiving a user query;
inferring a user query intention from the user query to develop an inferred user intention;
detecting a query type of the inferred user intention selected from a set of query types comprising a person type, an organization type, and an event type;
identifying a plurality of information categories based on the query type;
searching for information within each of the plurality of information categories based on identifying the plurality of information categories, wherein the searching is further based on natural language processing of a corpus of documents having multiple modalities comprising at least one of text, audio and video;
identifying a plurality of items in the searched information for each of the plurality of information categories, wherein each of the plurality of items comprises a span of text from the corpus of documents including two or more entities and a relation mention explicitly describing a relation between the two or more entities;
organizing the items into a plurality of equivalence classes, wherein each item having a same equivalence class comprises an equivalent relation mention;
selecting a representative item from the plurality of items for each of the equivalence classes; and
automatically generating a page in response to the user query by adaptively building a template with a plurality of page elements that correspond to the plurality of information categories based on the inferred user intention, wherein each of the plurality of page elements displays the searched information for a single information category from the plurality of information categories based on the selected representative items.

2. The method of claim 1 , further comprising: when the user query selects a person who has a political status, detecting the political status, and identifying the information categories comprising at least one of an election campaign, public appearances, statements, and public service history based on detecting the political status.

3. The method of claim 1 , further comprising: when the user query selects a company identifying the information categories comprising at least one of recent news about the company, information on the company's top officials, and press releases for the company based on the user query selecting the company.

4. The method of claim 1 , further comprising: when the user query selects an event identifying the information categories comprising at least one of news items about the event and reactions to the event based on the user query selecting the event.

5. The method of claim 4 , wherein entities in the event are identified and relevant information about the entities is searched.

6. A non-transitory computer program storage device embodying instructions executable by a processor, the non-transitory computer program storage device comprising storage memory configured to store:
program code that receives a user query;
program code that infers a user query intention from the user query to develop an inferred user intention;
program code that detects a query type of the inferred user intention selected from a set of query types comprising a person type, an organization type, and an event type;
program code that identifies a plurality of information categories based on the query type;
program code that searches for information within each of the plurality of information categories based on identifying the plurality of information categories, wherein the searching is further based on natural language processing of a corpus of documents having multiple modalities comprising at least one of text, audio and video;
program code that identifies a plurality of items in the searched information for each of the plurality of information categories, wherein each of the plurality of items comprises a span of text from the corpus of documents including two or more entities and a relation mention explicitly describing a relation between the two or more entities;
program code that organizes the items into a plurality of equivalence classes, wherein each item having a same equivalence class comprises an equivalent relation mention;
program code that selects a representative item from the plurality of items for each of the equivalence classes; and
program code that automatically generates a page in response to the user query by adaptively building a template with a plurality of page elements that correspond to the plurality of information categories based on the inferred user intention, wherein each of the plurality of page elements displays the searched information for a single information category from the plurality of information categories based on the selected representative items.

7. The non-transitory computer program storage device of claim 6 , further comprising: program code that, when the user query selects a person who has a political status, detects the political status, identifies the information categories comprising at least one of an election campaign, public appearances, statements, and public service history based on detecting the political status.

8. The non-transitory computer program storage device of claim 6 , further comprising program code that, when the user query selects a company, identifies the information categories comprising at least one of recent news about the company, information on the company's top officials, and press releases for the company based on the user query selecting the company.

9. The non-transitory computer program storage device of claim 6 , further comprising program code that, when the user query selects an event, identifies the information categories comprising at least one of news items about the event and reactions to the event based on the user query selecting the event.

10. The non-transitory computer program storage device of claim 9 , wherein entities in the event are identified and relevant information about the entities is searched.

11. A method for processing information by a processing device, the method comprising:
receiving a user query;
inferring a user query intention from the user query to develop an inferred user intention;
detecting a query type based on the inferred user intention;
searching for a plurality of information categories in a corpus of documents based on the query type;
identifying a plurality of items for each of the plurality of information categories based on the search, wherein each of the plurality of items comprises a span of text from the corpus of documents including two or more entities and a relation mention explicitly describing a relation between the two or more entities;
organizing the items into a plurality of equivalence classes, wherein each item having a same equivalence class comprises an equivalent relation mention;
selecting a representative item from the plurality of items for each of the plurality of equivalence classes; and
generating a page based on the query type, the page comprising a plurality of tabs, wherein each tab contains the representative items corresponding to the plurality of equivalence classes of an information category from the plurality of information categories.

12. The method of claim 11 , further comprising:
scoring each of the equivalence classes; and
prioritizing the equivalence classes based on the scoring.

Assignees

• IBM

Inventors

• Castelli Vittorio
• Florian Radu
• Luo Xiaoqiang
• Raghavan Hema

Classifications

• G06F40/295 Primary

  Named entity recognition · CPC title

• G06F16/35

  Clustering; Classification · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/3344

  using natural language analysis · CPC title

• G06F16/36

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