INCREMENTAL SPEECH RECOGNITION FOR DIALOG SYSTEMS

• US 20140156268A1
• Filed: 11/30/2012
• Published: 06/05/2014
• Est. Priority Date: 11/30/2012
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd. (Samsung Group)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: ARIZMENDI, Iker, WILLIAMS, Jason, SELFRIDGE, Ethan
• Granted Patent: US 9,015,048 B2
• Time in Patent Office: Days
• US Class Current: 704/231
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding

Abstract

A system and method for integrating incremental speech recognition in dialog systems. An example system configured to practice the method receives incremental speech recognition results of user speech as part of a dialog with a user, and copies a dialog manager operating on the user speech to generate temporary instances of the dialog manager. Then the system evaluates actions the temporary instances of the dialog manager would take based on the incremental speech recognition results, and identifies an action that would advance the dialog and a corresponding temporary instance of the dialog manager. The system can then execute the action in the dialog and optionally replace the dialog manager with the corresponding temporary instance of the dialog manager. The action can include making a turn-taking decision in the dialog, such as whether, what, and when to speak or whether to be silent.

Claims

1. A method comprising:

receiving incremental speech recognition results of user speech as part of a dialog with a user;

copying a dialog manager operating on the user speech to generate a temporary instance of the dialog manager;

evaluating, via a processing device, an action the temporary instance of the dialog manager would take based on the incremental speech recognition results;

determining that the action that would advance the dialog; and

executing the action in the dialog.

2. The method of claim 1, further comprising:

replacing the dialog manager with the temporary instance of the dialog manager.

3. The method of claim 1, wherein executing the action in the dialog comprises making a turn-taking decision in the dialog.

4. The method of claim 3, wherein the turn-making decision comprises deciding when to speak or when to be silent.

5. The method of claim 1, wherein executing the action further comprises providing a spoken response in the dialog.

6. The method of claim 1, further comprising:

discarding the temporary instance of the dialog manager when the action would not sufficiently advance the dialog.

7. The method of claim 1, further comprising:

deleting the dialog manager; and

substituting the dialog manager with the temporary instance of the dialog manager.

8. The method of claim 1, further comprising:

discarding incremental speech recognition results that do not correspond to a complete user utterance.

9. The method of claim 1, further comprising:

receiving a stability score for each incremental speech recognition result; and

discarding incremental speech recognition results having a corresponding stability score below a threshold.

10. The method of claim 1, wherein an incremental interaction manager receives the incremental speech recognition results.

11. A system comprising:

a processor; and

a computer-readable storage medium storing instructions which, when executed by the processor, cause the processor to perform steps comprising;

receiving, at a dialog manager an N-best list of incremental speech recognition results of user speech as part of a dialog with a user, wherein the incremental speech recognition results incorporate contextual information from a beginning of the dialog until the incremental speech recognition results were received;

receiving, at the dialog manager, respective confidence scores for each incremental speech recognition result in the N-best list; and

updating a dialog state associated with the dialog based on the respective confidence scores and the incremental speech results, and further based on an assumption that the incremental speech results are complete.

12. The system of claim 11, wherein the instructions, when executed by the processor, cause the processor further to:

advance the dialog based on the dialog state.

13. The system of claim 12, wherein advancing the dialog further comprises:

executing, in the dialog, an action associated with the dialog state.

14. The system of claim 11, wherein updating the dialog state further comprises:

copying the dialog manager operating to generate temporary instances of the dialog manager;

evaluating actions the temporary instances of the dialog manager would take based on the incremental speech recognition results; and

updating the dialog state based on the actions evaluated.

15. A tangible computer-readable storage medium having stored therein instructions which, when executed by a computing device, cause the computing device to perform a method comprising:

receiving an incremental speech recognition result of user speech as part of a dialog with a user;

copying a dialog manager operating on the user speech to generate a temporary dialog manager;

evaluating an action of the temporary dialog manager based on the incremental speech recognition result; and

upon determining that the action would advance the dialog and that the action is different than an existing action of the dialog manager;

executing the action in the dialog; and

replacing the dialog manager with the temporary dialog manager.

16. The tangible computer-readable storage medium of claim 15, wherein executing the action in the dialog comprises making a turn-taking decision in the dialog.

17. The tangible computer-readable storage medium of claim 16, wherein the turn-making decision comprises deciding when to speak or when to be silent.

18. The tangible computer-readable storage medium of claim 15, wherein executing the action further comprises providing a spoken response in the dialog.

19. The tangible computer-readable storage medium of claim 15, wherein the instructions, when executed by the computing device, cause the computing device to perform a method further comprising:

discarding temporary instances of the dialog manager which do not sufficiently advance the dialog.

20. The tangible computer-readable storage medium of claim 15, wherein the instructions, when executed by the computing device, cause the computing device to perform a method further comprising:

discarding incremental speech recognition results that do not correspond to a complete user utterance.