System and method of disambiguating natural language processing requests

• US 10,331,784 B2
• Filed: 07/31/2017
• Issued: 06/25/2019
• Est. Priority Date: 07/29/2016
• Status: Active Grant

• Current Assignee: Cerence Inc.
• Original Assignee: VoiceBox Technologies Corporation (Microsoft Corporation)
• Inventors: Swart, Erik, Drouin, Emilie
• Primary Examiner(s): Azad, Abul K
• Application Number:
  US15/664,249
• Publication Number:
  US 20180032503A1
• Time in Patent Office: 694 Days
• CPC Class Codes: G06F 16/219 Managing data history or versioning querying versioned data G06F16/2474; querying temporal data G06F16/2477
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/24575 using context
• CPC Class Codes: G06F 16/90332 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A system and method is provided of disambiguating natural language processing requests based on smart matching, request confirmations that are used until ambiguities are resolved, and machine learning. Smart matching may match entities (e.g., contact names, place names, etc.) based on user information such as call logs, user preferences, etc. If multiple matches are found and disambiguation has not yet been learned by the system, the system may request that the user identify the intended entity. On the other hand, if disambiguation has been learned by the system, the system may execute the request without confirmations. The system may use a record of confirmations and/or other information to continuously learn a user'"'"'s inputs in order to reduce ambiguities and no longer prompt for confirmations.

Claims

1. A computer implemented method of natural language disambiguation, the method being implemented in a computer system having one or more physical processors programmed with computer program instructions that, when executed by the one or more physical processors, cause the computer system to perform the method, the method comprising:

processing, by the computer system, using an intent recognition engine, a natural language input;

determining, by the computer system, based on the processing, at least a type of request and an input entity name;

identifying, by the computer system, at least a first match between the input entity name and a first identified entity name and at least a second match between the input entity name and a second identified entity name;

selectively performing, by the computer system, one of a passive confirmation or an active confirmation of whether the first identified entity name or the second identified entity name was intended by the natural language input, depending on a threshold comparison applied to previous confirmation history of a number of previous confirmations of whether the first identified entity name or the second entity name was intended by the natural language input,

wherein the active confirmation comprises;

requesting, by the computer system, a confirmation of whether the first identified entity name or the second identified entity name was intended by the natural language input; and

receiving, by the computer system, a response to the requested confirmation;

wherein the passive confirmation comprises;

selecting, as a response, the first identified entity name or the second identified entity name as the intended entity name;

generating, by the computer system, a request based on the type of request and the response; and

updating the number of previous responses regarding whether the first identified entity name or the second identified entity name was intended by the natural language input.

2. The computer implemented method of claim 1, wherein the input entity name includes at least one of a contact name, an address, or a business name.

3. The computer implemented method of claim 1, wherein the type of request includes a phone dialing request and the first identified entity name or the second identified entity name includes a user contact.

4. The computer implemented method of claim 1, wherein the type of request includes a search query and the first identified entity name or the second identified entity name includes a search object.

5. The computer implemented method of claim 1, wherein the type of request includes a navigation query and the first identified entity name or the second identified entity name includes a street address.

6. The computer implemented method of claim 1, wherein identifying at least a first match between the input entity name and the first identified entity name and at least a second match between the input entity name and the second identified entity name is at least partially based on at least one of a user location, a time of day, a user request history, a user request frequency, and a user preference.

7. The computer implemented method of claim 1, wherein identifying at least a first match between the input entity name and the first identified entity name and at least a second match between the input entity name and the second identified entity name is at least partially based on a user preference for contacting at least one of family, friends, and business associates.

8. The computer implemented method of claim 1, wherein:

the active confirmation is performed if a number of previous confirmations of whether the first identified entity name or the second identified entity name was intended by the natural language input is below a predetermined threshold; and

wherein the passive confirmation is performed if a number of previous confirmations of whether the first identified entity name or the second identified entity name was intended by the natural language input is above a predetermined threshold.

9. The computer implemented method of claim 1,

wherein the method further comprises:

alerting the user that execution of the request will be performed;

waiting, by the computer system, a predetermined amount of time for user cancellation of the announced execution prior to executing the request; and

updating the number of previous responses regarding whether the first identified entity name or the second identified entity name was intended by the natural language input.

10. The computer implemented method of claim 9, further comprising:

receiving, by the computer system, a cancellation command from the user;

requesting that the user select an alternate request different from the generated request; and

storing the alternate request selected by the user for use in future natural language input processing.

11. The computer implemented method of claim 1,

wherein the method further comprises:

processing, by the computer system, using the intent recognition engine, a second natural language input;

determining, by the computer system, a second input entity name based on the processing;

identifying, by the computer system, at least a third match between the second input entity name and the first identified entity name and at least a fourth match between the second input entity name and the second identified entity name;

determining, by the computer system, that the first number of times that the first identified entity name was intended exceeds a predetermined threshold; and

determining, by the computer system, that the first identified entity name is intended by the second natural language input based on the determination that the first number of times that the first identified entity name was intended exceeds the predetermined threshold.

12. The computer implemented method of claim 1, wherein requesting a confirmation comprises:

requesting that the user provide an alternate entity name, different from the first identified entity name and the second identified entity name;

providing information about the alternate entity name as the response to the requested confirmation,

and wherein the method further comprises;

storing, by the computer system, information indicating the response for use in future natural language input processing.

13. The computer implemented method of claim 1, further comprising:

updating the intent recognition engine with information indicating a match between the input entity name and the first identified entity name or second identified entity name based on one of the passive confirmation or the active confirmation,

wherein the computer system recognizes the first identified entity name and the second identified entity name as being an exact match.

14. The computer implemented method of claim 1, further comprising:

updating the intent recognition engine with information indicating a match between the input entity name and the first identified entity name or second identified entity name based on one of the passive confirmation or the active confirmation,

wherein the computer system recognizes the first identified entity name and the second identified entity name as being a phonetic match.

15. The computer implemented method of claim 1, further comprising:

updating the intent recognition engine with information indicating a match between the input entity name and the first identified entity name or second identified entity name based on one of the passive confirmation or the active confirmation,

wherein the computer system recognizes the first identified entity name and the second identified entity name as being a partial match.

16. A system of natural language disambiguation, the system comprising:

a computer system comprising one or more physical processors programmed by computer program instructions that, when executed, cause the computer system to;

process, using an intent recognition engine, a natural language input;

determine, based on the processed natural language input, at least a type of request and an input entity name;

identify at least a first match between the input entity name and a first identified entity name and at least a second match between the input entity name and a second identified entity name;

selectively perform one of a passive confirmation or an active confirmation of whether the first identified entity name or the second identified entity name was intended by the natural language input, depending on a threshold comparison applied to previous confirmation history of a number of previous confirmations of whether the first identified entity name or the second entity name was intended by the natural language input,

wherein the active confirmation comprises;

requesting a confirmation of whether the first identified entity name or the second identified entity name was intended by the natural language input; and

receiving a response to the requested confirmation;

wherein the passive confirmation comprises;

selecting, as a response, the first identified entity name or the second identified entity name as the intended entity name;

generate a request based on the type of request and the response; and

update the number of previous responses regarding whether the first identified entity name or the second identified entity name was intended by the natural language input.

17. The system of claim 16, wherein the input entity name includes at least one of a contact name, an address, and a business name.

18. The system of claim 16, wherein the request includes a phone dialing request and the first identified entity name or the second identified entity name includes a user contact.

19. The system of claim 16, wherein the type of request includes a search query and the first identified entity name or the second identified entity name includes a search object.

20. The system of claim 16, wherein the type of request includes a navigation query and the first identified entity name or the second identified entity name includes a street address.

21. The system of claim 16, wherein the computer program instructions further cause the computer system to:

identify at least the first match between the input entity name and the first identified entity name and at least the second match between the input entity name and the second identified entity name at least partially based on at least one of a user location, a time of day, a user request history, a user request frequency, and a user preference.

22. The system of claim 16, wherein the computer program instructions further cause the computer system to:

identify at least the first match between the input entity name and the first identified entity name and at least the second match between the input entity name and the second identified entity name at least partially based on user preferences including a preference for contacting at least one of family, friends, and business associates.

23. The system of claim 16, wherein the computer program instructions further cause the computer system to:

perform the active confirmation if a number of previous confirmations of whether the first identified entity name or the second identified entity name was intended by the natural language input is below a predetermined threshold; and

perform the passive confirmation if a number of previous confirmations of whether the first identified entity name or the second identified entity name was intended by the natural language input is above a predetermined threshold.

24. The system of claim 16, wherein

the computer program instructions further cause the computer system to:

alert the user that execution of the intended request will be performed;

wait a predetermined amount of time for user cancellation of the announced execution prior to executing the intended request; and

update the number of previous responses regarding whether the first identified entity name or the second identified entity name was intended by the natural language input.

25. The system of claim 24, further comprising instructions to cause the computer system to:

receive a cancellation command from the user;

request that the user select an alternate request different from the generated request; and

store the alternate request selected by the user for use in future natural language input processing.

26. The system of claim 16,

wherein the computer program instructions further cause the computer system to:

process, using the intent recognition engine, a second natural language input;

identify a second input entity name based on the processing;

identify at least a third match between the second input entity name and the first identified entity name and at least a fourth match between the second input entity name and the second identified entity name;

determine that the first number of times that the first identified entity name was intended exceeds a predetermined threshold; and

determine that the first identified entity name is intended by the second natural language input based on the determination that the first number of times that the first identified entity name was intended exceeds the predetermined threshold.

27. The system of claim 16, wherein the computer program instructions to request a confirmation further cause the computer system to:

request that the user provide an alternate entity name, different from the first identified entity name and the second identified entity name;

provide information about the alternate entity name as the response to the requested confirmation, and

wherein the computer system is further caused to store information indicating the response for use in future natural language input processing.

28. The system of claim 16, wherein the computer program instructions further cause the computer system to:

update the intent recognition engine with information indicating a match between the input entity name and the first identified entity name or second identified entity name based on one of the passive confirmation or the active confirmation,

wherein the computer system recognizes the first identified entity name and the second identified entity name as being an exact match.

29. The system of claim 16, wherein the computer program instructions further cause the computer system to:

update the intent recognition engine with information indicating a match between the input entity name and the first identified entity name or second identified entity name based on one of the passive confirmation or the active confirmation,

wherein the computer system recognizes the first identified entity name and the second identified entity name as being a phonetic match.

30. The system of claim 16, wherein the computer program instructions further cause the computer system to:

update the intent recognition engine with information indicating a match between the input entity name and the first identified entity name or second identified entity name based on one of the passive confirmation or the active confirmation,

wherein the computer system recognizes the first identified entity name and the second identified entity name as being a partial match.