Method and system for providing a voice interface

• US 9,978,365 B2
• Filed: 10/31/2008
• Issued: 05/22/2018
• Est. Priority Date: 10/31/2008
• Status: Active Grant

• Current Assignee: Nokia Technologies Oy (Nokia Corporation)
• Original Assignee: Nokia Technologies Oy (Nokia Corporation)
• Inventors: Adler, Mark R., Kiss, Imre, Polifroni, Joseph H., Wu, Tao
• Primary Examiner(s): Thomas, Ashish K
• Assistant Examiner(s): Ohba, Mellissa M
• Application Number:
  US12/263,012
• Publication Number:
  US 20100114944A1
• Time in Patent Office: 3,490 Days
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• CPC Class Codes: G10L 13/027 Concept to speech synthesisers; Generation of natural phrases from machine-based concepts generation of parameters for speech synthesis out of text G10L13/08
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A classifier voice interface of a user terminal may receive a query, may parse the query to identify an attribute, and may process the query to select a first domain-specific voice interface of a plurality of domain-specific voice interfaces based on the attribute, wherein each of the domain-specific voice interfaces comprises specialized information to process queries of different types. The classifier voice interface may further instruct the first domain-specific voice interface to process the query.

Claims

1. A method comprising:

determining weights for attributes by ranking the attributes based on user interactions with a user terminal;

storing the weights for the attributes in a memory of the user terminal;

processing, on the user terminal after the storing, a voice input in a first analysis, wherein the first analysis includes identifying one of the attributes;

identifying, as a result of the first analysis, one domain of a plurality of domains, wherein the identifying the one domain includes retrieving the stored weight for the identified attribute and identifying the one domain based on the identified attribute and the retrieved weight;

processing, on the user terminal, the voice input in a second analysis using specialized information of the one domain, wherein each of the plurality of domains comprises different respective specialized information; and

outputting as synthesized speech a response resulting from the second analysis.

2. The method of claim 1, further comprising

receiving in the voice input a predefined name;

and

identifying speech in the voice input that immediately follows the predefined name as a query, wherein the first analysis identifies the one of the attributes in the query.

3. The method of claim 2, wherein the predefined name corresponds to an identity assigned to the user terminal.

4. The method of claim 1, wherein the one domain is selected by identifying keywords in the voice input.

5. The method of claim 1, wherein the first analysis includes natural language processing of the voice input.

6. The method of claim 5, wherein the natural language processing includes using knowledge of user data stored on the user terminal that receives the voice input.

7. The method of claim 5, wherein the natural language processing includes identifying, in the voice input, each of one or more words as one of a verb and a noun.

8. The method of claim 5, wherein the natural language processing includes resolving a meaning of ambiguous words in the voice input by determining which of a plurality of meanings is more likely based on usage statistics associated with a user of the user terminal that receives the voice input.

9. The method of claim 5, wherein the natural language processing includes using preferences set by a user of the user terminal that receives the voice input.

10. The method of claim 1, further comprising:

determining, based on performance of the second analysis, that the second analysis was not appropriate for interpreting the voice input; and

performing a third analysis with another domain of the plurality of domains in response to the determining that the second analysis was not appropriate for interpreting the voice input, wherein the synthesized speech is specific to the another domain when the third analysis is performed.

11. The method of claim 1, further comprising identifying context information other than the voice input;

wherein the identifying the one domain is based on the context information.

12. The method of claim 11, wherein the context information comprises one of a date, a time, and a location.

13. The method of claim 1, further comprising identifying a language of the voice input.

14. The method of claim 1, wherein the identified attribute, the specialized information, or both the identified attribute and the specialized information are stored on the user terminal prior to the first analysis.

15. The method of claim 1, wherein the weight of the identified attribute is a predefined user preference.

16. The method of claim 1, wherein the retrieved weight of the identified attribute is updated based upon an outcome of the second analysis.

17. One or more non-transitory computer readable media storing computer-executable instructions which, when executed by a processor, cause the processor to:

determine weights for attributes by ranking the attributes based on user interactions with a user terminal;

store the weights for the attributes in a memory;

process, after the weights are stored in the memory, a voice input in a first analysis, wherein the first analysis includes identifying one of the attributes;

identify, as a result of the first analysis, one domain of a plurality of domains wherein the identifying the one domain includes retrieving the stored weight for the identified attribute and identifying the one domain based on the identified attribute and the retrieved weight;

process the voice input in a second analysis using specialized information of the one domain, wherein each of the plurality of domains comprises different respective specialized information; and

cause a response resulting from the second analysis to be output as synthesized speech.

18. The one or more non-transitory computer readable media of claim 17, storing further computer-executable instructions which, when executed by the processor, cause the processor to:

receive in the voice input a predefined name; and

identify speech in the voice input that immediately follows the predefined name as a query, wherein the first analysis identifies the one of the attributes in the query.

19. The one or more non-transitory computer readable media of claim 17, storing further computer-executable instructions which, when executed by the processor, cause the processor to:

identify context information other than the voice input;

wherein the one domain is identified based on the context information.

20. The one or more non-transitory computer readable media of claim 17, storing further computer-executable instructions which, when executed by the processor, cause the processor to:

identify a language of the voice input.

21. An apparatus comprising:

a processor; and

memory including computer readable instructions, wherein the memory including the computer readable instructions are configured to, with the processor, cause the apparatus to;

determine weights for attributes by ranking the attributes based on user interactions with the apparatus;

store the weights for the attributes in the memory;

process, after the weights are stored in the memory, a voice input in a first analysis, wherein the first analysis includes identifying one of the attributes;

identify, as a result of the first analysis, one domain of a plurality of domains wherein the identifying the one domain includes retrieving the stored weight for the identified attribute and identifying the one domain based on the identified attribute and the retrieved weight;

process the voice input in a second analysis using specialized information of the one domain, wherein each of the plurality of domains comprises different respective specialized information; and

cause a response resulting from the second analysis to be output as synthesized speech.

22. The apparatus of claim 21, wherein the memory and the computer readable instructions are configured to, with the processor, cause the apparatus to:

identify a language of the voice input.

23. An apparatus comprising:

means for determining weights for attributes by ranking the attributes based on user interactions with a user terminal;

means for storing the weights for the attributes in a memory;

means for processing, after the weights are stored in the memory, a voice input in a first analysis, wherein the first analysis includes identifying one of the attribute;

means for identifying, as a result of the first analysis, one domain of a plurality of domains wherein the identifying the one domain includes retrieving the stored weight for the identified attribute and identifying the one domain based on the identified attribute and the retrieved weight;

means for processing the voice input in a second analysis using specialized information of the one domain, wherein each of the plurality of domains comprises different respective specialized information; and

means for outputting as synthesized speech a response resulting from the second analysis.

24. A method comprising:

determining weights for attributes by ranking the attributes based on user interactions with a user terminal;

storing the weights for the attributes in a memory of the user terminal;

performing, on the user terminal after the storing, a first analysis of a voice input, wherein the first analysis includes identifying one of the attributes;

identifying, as a result of the first analysis, one domain of a plurality of domains, wherein the identifying the one domain includes retrieving the stored weight for the identified attribute and identifying the one domain based on the identified attribute and the retrieved weight;

performing, based on the one domain, a second analysis on the user terminal of the voice input using specialized information of the one domain, wherein each of the plurality of domains comprises different respective specialized information; and

outputting as synthesized speech a response resulting from the second analysis, wherein the identified attribute, the specialized information, or both the identified attribute and the specialized information are based on user data stored on the user terminal prior to the first analysis.