Systems and methods for dynamic re-configurable speech recognition

• US 9,536,524 B2
• Filed: 05/01/2014
• Issued: 01/03/2017
• Est. Priority Date: 10/13/2000
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Rose, Richard C., Gajic, Bojana
• Primary Examiner(s): Shah, Paras D
• Application Number:
  US14/267,133
• Publication Number:
  US 20140244251A1
• Time in Patent Office: 978 Days
• Field of Search: 704243-244, 704/246, 704/251, 704/255
• US Class Current: 1/1
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates
• CPC Class Codes: G10L 15/142 Hidden Markov Models [HMMs]
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence

Abstract

Speech recognition models are dynamically re-configurable based on user information, background information such as background noise and transducer information such as transducer response characteristics to provide users with alternate input modes to keyboard text entry. The techniques of dynamic re-configurable speech recognition provide for deployment of speech recognition on small devices such as mobile phones and personal digital assistants as well environments such as office, home or vehicle while maintaining the accuracy of the speech recognition.

Claims

1. A method comprising:

generating a user identifier using a voice request, the voice request received from a device;

estimating, via successive comparisons, a transducer noise parameter of the device, wherein a delay between the successive comparisons is increased when successive changes do not exceed a threshold value;

comparing stored user identities to the user identifier, to yield a comparison;

when, based on the comparison, a user is associated with the device;

retrieving a parameterizable speech recognition model associated with the user identifier; and

adapting the parameterizable speech recognition model based on the transducer noise parameter to yield an adapted parameterizable speech recognition model; and

performing speech recognition on the voice request using the adapted parameterizable speech recognition model.

2. The method of claim 1, wherein the parameterizable speech recognition model is speaker independent.

3. The method of claim 2, wherein generating of the user identifier occurs after receiving a unique user code in the voice request.

4. The method of claim 1, wherein the parameterizable speech recognition model is generated based on a background model, a transducer model, and the user identifier.

5. The method of claim 1, further comprising estimating, via successive comparisons, a background noise parameter.

6. The method of claim 5, wherein the adapting of the parameterizable speech recognition model is further based on the background noise parameter.

7. A system comprising:

a processor; and

a computer-readable storage medium having instructions stored which, when executed by the processor, cause the processor to perform operations comprising;

generating a user identifier using a voice request, the voice request received from a device;

estimating, via successive comparisons, a transducer noise parameter of the device, wherein a delay between the successive comparisons is increased when successive changes do not exceed a threshold value;

comparing stored user identities to the user identifier, to yield a comparison;

when, based on the comparison, a user is associated with the device;

retrieving a parameterizable speech recognition model associated with the user identifier; and

adapting the parameterizable speech recognition model based on the transducer noise parameter to yield an adapted parameterizable speech recognition model; and

performing speech recognition on the voice request using the adapted parameterizable speech recognition model.

8. The system of claim 7, wherein the parameterizable speech recognition model is speaker independent.

9. The system of claim 8, wherein generating of the user identifier occurs after receiving a unique user code in the voice request.

10. The system of claim 7, wherein the parameterizable speech recognition model is generated based on a background model, a transducer model, and the user identifier.

11. The system of claim 7, the computer-readable storage medium having additional instructions which result in operations comprising estimating, via successive comparisons, a background noise parameter.

12. The system of claim 11, wherein the adapting of the parameterizable speech recognition model is further based on the background noise parameter.

18. The system of claim 11, wherein the adapting of the parameterizable speech recognition model is further based on the background noise parameter.

13. A computer-readable storage device having instructions stored which, when executed by a computing device, cause the computing device to perform operations comprising:

generating a user identifier using a voice request, the voice request received from a device;

estimating, via successive comparisons, a transducer noise parameter of the device, wherein a delay between the successive comparisons is increased when successive changes do not exceed a threshold value;

comparing stored user identities to the user identifier, to yield a comparison;

when, based on the comparison, a user is associated with the device;

retrieving a parameterizable speech recognition model associated with the user identifier; and

adapting the parameterizable speech recognition model based on the transducer noise parameter to yield an adapted parameterizable speech recognition model; and

performing speech recognition on the voice request using the adapted parameterizable speech recognition model.

14. The computer-readable storage device of claim 13, wherein the parameterizable speech recognition model is speaker independent.

15. The computer-readable storage device of claim 14, wherein generating of the user identifier occurs after receiving a unique user code in the voice request.

16. The computer-readable storage device of claim 13, wherein the parameterizable speech recognition model is generated based on a background model, a transducer model, and the user identifier.

17. The computer-readable storage device of claim 13, having additional instructions which result in operations comprising estimating, via successive comparisons, a background noise parameter.