Confidence measure generation for speech related searching

• US 8,165,877 B2
• Filed: 08/03/2007
• Issued: 04/24/2012
• Est. Priority Date: 08/03/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Wang, Ye-Yi, Ju, Yun-Cheng, Yu, Dong
• Primary Examiner(s): Saint Cyr, Leonard
• Application Number:
  US11/888,928
• Publication Number:
  US 20090037175A1
• Time in Patent Office: 1,726 Days
• Field of Search: None
• US Class Current: 704/240
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding

Abstract

A voice search system has a speech recognizer, a search component, and a dialog manager. A confidence measure generator receives speech recognition features from the speech recognizer, search features from the search component, and dialog features from the dialog manager, and calculates an overall confidence measure for voice search results based upon the features received. The invention can be extended to include the generation of additional features, based on those received from the individual components of the voice search system.

Claims

1. A method of conducting a dialog with a user of a voice search system that includes a speech recognizer, a search component and a dialog manager, the method comprising:

generating a dialog feature, with the dialog manager, indicative of how well the voice search system is responding to the user;

recognizing a user speech input with the speech recognizer to generate a query and a speech recognition feature indicative of a measure of speech recognition accuracy;

searching a data store, with the search component, for search results based on the query, and generating a search feature indicative of a measure of search accuracy;

generating a voice search confidence measure, indicative of a confidence of the search results for the query, based on the dialog feature indicative of how well the voice search system is responding to the user, the speech recognition feature, and the search feature; and

conducting the dialog by conducting dialog turns with the user, each dialog turn corresponding to an interaction with the user, and conducting the dialog according to a dialog strategy chosen based, at least in part, on the voice search confidence measure, the dialog strategy determining what type of dialog turn to next conduct with the user, the dialog feature being derived based on one or more of the dialog turns conducted with the user.

2. The method of claim 1 and further comprising:

generating additional features based on at least some of the dialog feature, the speech recognition feature and the search feature.

3. The method of claim 1 wherein generating a voice search confidence measure comprises:

applying a maximum entropy model to the dialog feature, the speech recognition feature and the search feature to generate the voice search confidence measure.

4. The method of claim 3 wherein applying a maximum entropy model comprises:

classifying the search results as correct or incorrect, and calculating a probability associated with the classification.

5. The method of claim 1 wherein generating a dialog feature comprises:

generating a plurality of dialog features.

6. The method of claim 1 wherein recognizing a user speech input to generate a speech recognition feature comprises:

generating a plurality of speech recognition features.

7. The method of claim 1 wherein generating a search feature comprises:

generating a plurality of search features.

8. The method of claim 1, wherein the dialog feature is indicative of how well the voice search system is responding to the user.

9. The method of claim 1, wherein the dialog feature is indicative of how many dialog turns have been conducted with the user.

10. The method of claim 1, wherein the dialog feature represents the dialog turn at which the spoken input was received.

11. The method of claim 1, wherein the dialog feature comprises a previous turn occurrence feature representing whether the search result was presented to the user in a dialog turn prior to the dialog turn at which the spoken input was received.

12. A voice search system, comprising:

a processor;

a speech recognition system receiving a spoken input from a user and generating a query from the spoken input;

a search system receiving the query and searching a data store for a search result based on the query;

a dialog component conducting a user dialog by conducting dialog turns with the user, each dialog turn corresponding to an interaction with the user, the dialog component generating at least one dialog feature based at least in part on a plurality of dialog turns conducted with the user; and

a confidence measure generator receiving features, including at least one dialog feature received from the dialog component and at least one feature received from one or more of the speech recognizer and the search component, the confidence measure generator generating a voice search confidence measure corresponding to a confidence of the search result based on the received features using the processor, the dialog component conducting the user dialog according to a strategy based on the voice search confidence measure.

13. The voice search system of claim 12 wherein the speech recognition system is configured to generate at least one of the features as a speech recognition feature indicative of an accuracy of speech recognition of the spoken input.

14. The voice search system of claim 13 wherein the search system is configured to generate at least one of the features as a search feature indicative of an accuracy of the search result, given the query.

15. The voice search system of claim 14 wherein the dialog component is configured to generate at least one of the features as a dialog feature indicative of a performance of the voice search system in responding to the spoken input.

16. The voice search system of claim 15 wherein the search system is configured to receive a keyboard text input as a second query and search the data store for a second search result based on the second query and wherein the confidence measure generator is further configured to generate a text search confidence measure corresponding to the second search result.

17. The voice search system of claim 12, wherein the at least one dialog feature comprises a feature indicative of how well the voice search system is responding to the user.

18. The voice search system of claim 12, wherein the at least one dialog feature comprises a feature indicative of how many dialog turns have been conducted with the user.

19. The voice search system of claim 12, wherein the at least one dialog feature comprises a feature representing the dialog turn at which the spoken input was received.

20. The voice search system of claim 12, wherein the at least one dialog feature comprises a previous turn occurrence feature representing whether the search result was presented to the user in a dialog turn prior to the dialog turn at which the spoken input was received.