Methods and apparatus for module arbitration

• US 10,210,003 B2
• Filed: 09/30/2014
• Issued: 02/19/2019
• Est. Priority Date: 09/30/2014
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Ducatelle, Frederick, Grober, Marcus, Martens, Gaetan
• Primary Examiner(s): Abebe, Daniel
• Application Number:
  US15/509,884
• Publication Number:
  US 20170308389A1
• Time in Patent Office: 1,603 Days
• Field of Search: 704275
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• CPC Class Codes: G06F 9/451 Execution arrangements for user interfaces
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G10L 2015/228 of application context

Abstract

Methods and apparatus to process a user input on independent applications that provide classifier outputs to an arbitration module, which selects one of the application to respond to the user input. The classifier outputs include a probability that the user input is in domain for the application functionality.

Claims

1. A method, comprising

receiving, in a first device in response to a user input, a first input from a first classifier of a first application executing in the first device, the first classifier having been trained using an application-independent vocabulary, the first input including a probability of at least a portion of the user input being in domain for the first application, wherein in domain for the first application refers to functionality provided by the first application using a first subset of the application-independent vocabulary;

receiving, in the first device in response to the user input, a second input from a second classifier of a second application executing in a second device, the second classifier having been trained using the application-independent vocabulary, the second input including a probability of at least a portion of the user input being in domain for the second application, wherein in domain for the second application refers to functionality provided by the second application using a second subset of the application-independent vocabulary; and

performing, in the first device, arbitration on the user input using the first and second inputs to prioritize a first one of the first and second applications for responding to the user input.

2. The method according to claim 1, further including performing automated speech recognition (ASR) on the user input by an ASR module in the first application specific to the first application and generating a first feature list for the user input.

3. The method according to claim 2, further including processing the first feature list by the first classifier in the first application.

4. The method according to claim 1, wherein the first device comprises a head unit of a vehicle.

5. The method according to claim 4, wherein the second device comprises a mobile phone wirelessly connected to the head unit.

6. The method according to claim 1, wherein the first classifier includes an in domain/out of domain classifier trained using input data relevant to the first application for the in domain data and input data that is not relevant to the first application for out of domain data.

7. The method according to claim 1, wherein the first and second applications were independently developed.

8. The method according to claim 1, wherein the first classifier receives a feature set as input derived from the user input.

9. An article, comprising:

a non-transitory computer readable storage medium having stored instructions that enable a first device to;

receive, in response to a user input, a first input from a first classifier of a first application executing in the first device, the first classifier having been trained using an application-independent vocabulary, the first input including a probability of at least a portion of the user input being in domain for the first application, wherein in domain for the first application refers to functionality provided by the first application using a first subset of the application-independent vocabulary;

receive, in response to the user input, a second input from a second classifier of a second application executing in a second device, the second classifier having been trained using the application-independent vocabulary, the second input including a probability of at least a portion of the user input being in domain for the second application, wherein in domain for the second application refers to functionality provided by the second application using a second subset of the application-independent vocabulary; and

perform arbitration on the user input using the first and second inputs to prioritize a first one of the first and second applications for responding to the user input.

10. The article according to claim 9, wherein the first device comprises a head unit of a vehicle.

11. The article according to claim 10, wherein the second device comprises a mobile phone wirelessly connected to the head unit.

12. The article according to claim 9, wherein the first classifier includes an in domain/out of domain classifier trained using input data relevant to the first application for the in domain data and input data that is not relevant to the first application for out of domain data.

13. The article according to claim 9, wherein the first and second applications were independently developed.

14. The article according to claim 9, wherein the first classifier receives a feature set as input derived from the user input.

15. A system, comprising:

an interface to communicate with first and second applications; and

an arbitration means coupled to the interface, the arbitration means to arbitrate on a user input from a user based on a first input from a first classifier of a first application executing in a first device and a second input from a second classifier of a second application executing in a second, the arbitration means to prioritize a first one of the first and second applications for responding to the user input, the first and second classifiers having been trained using an application-independent vocabulary,

the first input including a probability of at least a portion of the user input being in domain for the first application, wherein in domain for the first application corresponds to functionality provided by the first application using a first subset of the application-independent vocabulary, and

the second input including a probability of at least a portion of the user input being in domain for the second application, wherein in domain for the second application corresponds to functionality provided by the second application using a second subset of the application-independent vocabulary.

16. The system according to claim 15, wherein the first device comprises a vehicle head unit and the second device comprises a device wirelessly connected to the head unit.

17. The system according to claim 15, wherein the first classifier includes an in domain/out of domain classifier trained using input data relevant to the first application for the in domain data and input data that is not relevant to the first application for out of domain data.

18. The system according to claim 15, wherein the first classifier receives a feature set from an automated speech recognition module as input derived from the user input.