Method of and system for inferring user intent in search input in a conversational interaction system

• US 9,183,183 B2
• Filed: 05/01/2013
• Issued: 11/10/2015
• Est. Priority Date: 07/20/2012
• Status: Active Grant

• Current Assignee: Veveo, Inc. (Adeia Inc.)
• Original Assignee: Veveo, Inc. (Adeia Inc.)
• Inventors: Barve, Rakesh, Aravamudan, Murali, Venkataraman, Sashikumar, Welling, Girish
• Primary Examiner(s): Uddin, Mohammed R
• Application Number:
  US13/874,523
• Publication Number:
  US 20140025706A1
• Time in Patent Office: 923 Days
• Field of Search: 707/771, 707/706, 707/714, 707/999.003, 707/999.005
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2457 with adaptation to user needs
• CPC Class Codes: G06F 16/24573 using data annotations, e.g. user-defined metadata
• CPC Class Codes: G06F 16/84 Mapping; Conversion
• CPC Class Codes: G06F 16/9558 Details of hyperlinks; Management of linked annotations
• CPC Class Codes: G06F 40/10 Text processing natural language analysis G06F40/20; semantic analysis G06F40/30; processing or translation of natural language G06F40/40
• CPC Class Codes: G06F 40/237 Lexical tools
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30

Abstract

A method of inferring user intent in search input in a conversational interaction system is disclosed. A method of inferring user intent in a search input includes providing a user preference signature that describes preferences of the user, receiving search input from the user intended by the user to identify at least one desired item, and determining that a portion of the search input contains an ambiguous identifier. The ambiguous identifier is intended by the user to identify, at least in part, a desired item. The method further includes inferring a meaning for the ambiguous identifier based on matching portions of the search input to the preferences of the user described by the user preference signature and selecting items from a set of content items based on comparing the search input and the inferred meaning of the ambiguous identifier with metadata associated with the content items.

Claims

1. A method of inferring user intent in a search input based on resolving ambiguous portions of the search input, the method comprising:

providing access to a set of content items, each of the content items being associated with metadata that describes the corresponding content item;

providing a user preference signature, the user preference signature describing preferences of the user for at least one of (i) particular content items and (ii) metadata associated with the content items;

receiving search input from the user, the search input being intended by the user to identify at least one desired content item;

determining a state space associated with the set of content items and with at least a portion of the search input, wherein the state space includes entities and relationships between the entities;

determining that a portion of the search input contains an ambiguous identifier, the ambiguous identifier intended by the user to identify, at least in part, the at least one desired content item;

inferring a meaning for the ambiguous identifier by performing the following;

determining whether an entity and a relationship in the state space disambiguates the ambiguous identifier,

upon a condition that at least one of the entity and the relationship in the state space disambiguates the ambiguous identifier, the inferring the meaning for the ambiguous identifier being based at least in part on the at least one of the entity and the relationship in the state space that disambiguates the ambiguous identifier,

upon a condition that the at least one of the entity and the relationship in the state space does not disambiguate the ambiguous identifier, the inferring the meaning for the ambiguous identifier including formulating and presenting a follow-up conversational question formed in part based on selecting at least one response template, wherein the at least one response template is selected based at least in part on matching characteristics of at least one of entities and relationships along a path that includes an entity and a relationship that matches portions of the search input to the preferences of the user described by the user preference signature;

determining, based on the state space, whether the search input is related to a portion of a previously received search input; and

in response to determining that the search input is not related to the portion of the previously-received search input, selecting content items from the set of content items based on comparing the search input and the inferred meaning of the ambiguous identifier with metadata associated with the content items and further based on disregarding entities related to the portion of the previously-received search input.

2. The method of claim 1, the ambiguous identifier being a pronoun or syntactic expletive.

3. The method of claim 1, the ambiguous identifier being an identifier of an entertainment genre.

4. The method of claim 1, the ambiguous identifier being at least a portion of a name.

5. The method of claim 1, the metadata associated with the content items including a mapping of relationships between entities associated with the content items.

6. The method of claim 1, the user preference signature being based on explicit preferences provided by the user.

7. The method of claim 1, the user preference signature being based on analyzing content item selections by the user conducted over a period of time.

8. The method of claim 1, the user preference signature describing preferences of the user for metadata associated with the content items, the metadata including entities preferred by the user.

9. The method of claim 1, the content items including media entertainment items and the metadata including at least one of crew, characters, actors, teams, leagues, tournaments, athletes, composers, music artists, performers, albums, songs, news personalities, and content distributors.

10. The method of claim 1,

the content items including at least one of electronic mail items and address book items, and

the metadata including at least one of electronic mail threads, contacts, senders, recipients, company names, business departments, business units, electronic mail folders, office location information, contact names, electronic mail address, telephone number, address, and employer.

11. The method of claim 1, the content items including travel-related items and the metadata including at least one of cities, hotels, hotel brands, individual points of interest, categories of points of interest, retail chains, car rental websites, and car rental company names.

12. The method of claim 1, the content items including electronic commerce items and the metadata including at least one of product items, product categories, product subcategories, product brand, and retail stores.

13. The method of claim 1, the content items including network-based documents and the metadata including at least one of domain names, internet media types, filenames, directories, and filename extensions.

14. The method of claim 1, wherein the determining the state space is further based on at least a portion of search input received from the user before the immediately received search input.

15. The method of claim 1, wherein the selection of the at least one response template further comprises:

determining a number of entities recognized in the search input; and

selecting the at least one response template based in part on the number of entities recognized in the search input.

16. A method of inferring user intent in a search input based on resolving ambiguous portions of the search input, the method comprising:

providing access to a set of content items, each of the content items being associated with metadata that describes the corresponding content item;

receiving search input from the user, the search input being intended by the user to identify at least one desired content item;

determining a state space associated with the set of content items and with at least a portion of the search input, wherein the state space includes entities and relationships between the entities;

determining whether or not a portion of the search input contains an ambiguous identifier, the ambiguous identifier intended by the user to identify, at least in part, the at least one desired content item;

upon a condition in which a portion of the search input contains an ambiguous identifier, performing the following;

determining whether at least one of an entity and a relationship in the state space disambiguates the ambiguous identifier,

upon the condition that at least one of the entity and the relationship in the state space disambiguates the ambiguous identifier,

inferring a meaning for the ambiguous identifier based at least in part on the at least one of the entity and the relationship in the state space that disambiguates the ambiguous identifier, and

selecting content items from the set of content items based on comparing the search input and the inferred meaning of the ambiguous identifier with metadata associated with the content items; and

upon a condition in which the at least one of the entity and the relationship in the state space does not disambiguate the ambiguous identifier,

inferring the meaning for the ambiguous identifier including formulating and presenting a follow-up conversational question formed in part based on selecting at least one response template, wherein the at least one response template is selected based at least in part on matching characteristics of at least one of entities and relationships along a path that includes an entity and a relationship that matches portions of the search input to the preferences of the user described by the user preference signature,

determining, based on the state space, whether the search input is related to a portion of a previously received search input; and

in response to determining that the search input is not related to the portion of the previously-received search input, selecting content items from the set of content items based on comparing the search input and the inferred meaning of the ambiguous identifier with metadata associated with the content items and further based on matching portions of the search input to preferences of the user described by a user preference signature and further based on disregarding entities related to the portion of the previously-received search input.

17. The method of claim 16, the ambiguous identifier being a pronoun or syntactic expletive.

18. The method of claim 16, the ambiguous identifier being an identifier of an entertainment genre.

19. The method of claim 16, the ambiguous identifier being at least a portion of a name.

20. The method of claim 16, the metadata associated with the content items including a mapping of relationships between entities associated with the content items.

21. The method of claim 16, the user preference signature being based on explicit preferences provided by the user.

22. The method of claim 16, the user preference signature being based on analyzing content item selections by the user conducted over a period of time.

23. The method of claim 16, the user preference signature describing preferences of the user for metadata associated with the content items, the metadata including entities preferred by the user.

24. The method of claim 16, the user preference signature describing preferences of the user for particular content items.

25. The method of claim 16, the content items including media entertainment items and the metadata including at least one of crew, characters, actors, teams, leagues, tournaments, athletes, composers, music artists, performers, albums, songs, news personalities, and content distributors.

26. The method of claim 16,

the content items including at least one of electronic mail items and address book items, and

the metadata including at least one of electronic mail threads, contacts, senders, recipients, company names, business departments, business units, electronic mail folders, office location information, contact names, electronic mail address, telephone number, address, and employer.

27. The method of claim 16, the content items including travel-related items and the metadata including at least one of cities, hotels, hotel brands, individual points of interest, categories of points of interest, retail chains, car rental websites, and car rental company names.

28. The method of claim 16, the content items including electronic commerce items and the metadata including at least one of product items, product categories, product subcategories, product brand, and retail stores.

29. The method of claim 16, the content items including network-based documents and the metadata including at least one of domain names, internet media types, filenames, directories, and filename extensions.

30. The method of claim 16, wherein the determining the state space is further based on at least a portion of search input received from the user before the immediately received search input.

31. The method of claim 16, wherein the selection of the at least one response template further comprises:

determining a number of entities recognized in the search input; and

selecting the at least one response template based in part on the number of entities recognized in the search input.