Method and apparatus for multiple value confirmation and correction in spoken dialog system

• US 8,433,572 B2
• Filed: 04/02/2008
• Issued: 04/30/2013
• Est. Priority Date: 04/29/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Caskey, Sasha Porto, Huerta, Juan Manuel, Pieraccini, Roberto
• Primary Examiner(s): GUERRA-ERAZO, EDGAR X
• Application Number:
  US12/061,291
• Publication Number:
  US 20080183470A1
• Time in Patent Office: 1,854 Days
• Field of Search: 704/255, 704/257, 704/270, 704/270.1, 704/275, 704/258, 704/260, 704/261, 704/233, 704/235, 704/243, 704/9, 704/10, 704/251
• US Class Current: 704/255
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A method for multiple value confirmation and correction in spoken dialog systems. A user is allowed to correct errors in values captured by the spoken dialog system, such that the interaction necessary for error correction between the system and the user is reduced. When the spoken dialog system collects a set of values from a user, the system provides a spoken confirmation of the set of values to the user. The spoken confirmation comprises the set of values and possibly pause associated with each value. Upon hearing an incorrect value, the user may react and barge-in the spoken confirmation and provide a corrected value. Responsive to detecting the user interruption during the pause or after the system speaking of a value, the system halts the spoken confirmation and collects the corrected value. The system then provides a new spoken confirmation to the user, wherein the new spoken confirmation includes the corrected value.

Claims

1. A method for multiple value confirmation and correction in a spoken dialog system, the method comprising:

responsive to collecting a set of values from a user, inviting the user to interrupt a spoken confirmation of the set of values, previously collected from the user, upon hearing an incorrect value, to correct the incorrect value;

presenting the spoken confirmation of the set of previously collected values to the user, wherein the spoken confirmation includes a sequence of prompts spoken by the spoken dialog system, the sequence of prompts comprising spoken confirmation values of the set of previously collected values as understood by the spoken dialog system and a pause after each of the spoken confirmation values, each pause except the last immediately followed by a next spoken confirmation value in the sequence of prompts;

detecting a user interruption of a spoken confirmation value or a pause in the sequence of prompts, wherein the user interruption comprises an indication that the spoken confirmation value is incorrect;

responsive to detecting the user interruption, halting the sequence of prompts being spoken by the spoken dialog system and collecting a corrected value from the user; and

providing a new spoken confirmation to the user, wherein the new spoken confirmation includes the corrected value.

2. The method of claim 1, wherein detecting a user interruption comprises using one or more correction grammars to allow for stating a value correction in multiple ways.

3. The method of claim 1, wherein a time stamp is used to determine the value after which the user interruption occurred.

4. The method of claim 1, wherein the spoken confirmation and the new spoken confirmation comprise at least one carrier phrase.

5. The method of claim 1, further comprising:

repeating the detecting, halting, and providing steps until the set of values in the new spoken confirmation are correct.

6. The method of claim 1, wherein the new spoken confirmation is dynamically tuned based on the spoken dialog system'"'"'s confidence in accuracy of the set of values.

7. The method of claim 1, further comprising:

modifying the duration of the pause based on the spoken dialog system'"'"'s confidence in an accuracy of the value.

8. The method of claim 1, further comprising:

modifying the intonation of the new spoken confirmation based on the spoken dialog system'"'"'s confidence in an accuracy of the value.

9. The method of claim 1, wherein the user interruption for the value is detected within a predetermined time interval after the value is spoken.

10. A spoken dialog system comprising a processor and a memory encoded with computer-executable instructions that, when executed by the processor, perform a method for multiple value confirmation and correction, comprising:

responsive to collecting a set of values from a user, inviting the user to interrupt a spoken confirmation of the set of values, previously collected from the user, upon hearing an incorrect value, to correct the incorrect value;

presenting the spoken confirmation of the set of previously collected values to the user wherein the spoken confirmation includes a sequence of prompts spoken by the spoken dialog system, the sequence of prompts comprising spoken confirmation values of the set of previously collected values as understood by the spoken dialog system and a pause after each of the spoken confirmation values, each pause except the last immediately followed by a next spoken confirmation value in the sequence of prompts;

detecting a user interruption of a spoken confirmation value or a pause in the sequence of prompts, wherein the user interruption comprises an indication that the spoken confirmation value is incorrect;

halting the sequence of prompts being spoken by the spoken dialog system in response to detecting the user interruption;

collecting a corrected value from the user in response to detecting the user interruption; and

providing a new spoken confirmation to the user, wherein the new spoken confirmation includes the corrected value.

11. The spoken dialog system of claim 10, wherein detecting a user interruption comprises using one or more correction grammars to allow for stating a value correction in multiple ways.

12. The spoken dialog system of claim 10, wherein a time stamp is used to determine the value after which the user interruption occurred.

13. The spoken dialog system of claim 10, wherein the spoken confirmation and the new spoken confirmation comprise at least one carrier phrase.

14. The spoken dialog system of claim 10, wherein the method further comprises:

repeating the detecting, halting, and providing steps until the set of values in the new spoken confirmation are correct.

15. The spoken dialog system of claim 10, wherein the new spoken confirmation is dynamically tuned based on the spoken dialog system'"'"'s confidence in accuracy of the set of values.

16. The spoken dialog system of claim 10, wherein the method further comprises:

modifying the duration of the pause based on the spoken dialog system'"'"'s confidence in an accuracy of the value.

17. The spoken dialog system of claim 10, wherein the method further comprises:

modifying an intonation of the new spoken confirmation based on the spoken dialog system'"'"'s confidence in an accuracy of the value.

18. A computer-readable storage device encoded with computer-executable instructions that, when executed by a computing device, perform a method for multiple value confirmation and correction in a spoken dialog system, the method comprising:

responsive to collecting a set of values from a user, inviting the user to interrupt a spoken confirmation of the set of values, previously collected from the user, upon hearing an incorrect value, to correct the incorrect value;

presenting the spoken confirmation of the set of previously collected values to the user, wherein the spoken confirmation includes a sequence of prompts spoken by the spoken dialog system, the sequence of prompts comprising spoken confirmation values of the set of previously collected values as understood by the spoken dialog system and a pause after each of the spoken confirmation values, each pause except the last immediately followed by a next spoken confirmation value in the sequence of prompts;

detecting a user interruption of a spoken confirmation value or a pause in the sequence of prompts, wherein the user interruption comprises an indication that the spoken confirmation value is incorrect;

halting the sequence of prompts being spoken by the spoken dialog system and collecting a corrected value from the user in response to detecting the user interruption; and

providing a new spoken confirmation to the user, wherein the new spoken confirmation includes the corrected value.

19. A method for a multiple value confirmation and correction in a spoken dialog system, the method comprising:

responsive to collecting a set of values from a user, inviting the user to interrupt a spoken confirmation of the set of values, previously collected from the user, upon hearing an incorrect value, to correct the incorrect value;

presenting the spoken confirmation of the set of previously collected values to the user, wherein the spoken confirmation includes a sequence of prompts spoken by the spoken dialog system, the sequence of prompts comprising spoken confirmation values of the set of previously collected values as understood by the spoken dialog system and a pause after each of the spoken confirmation values, each pause except the last immediately followed by a next spoken confirmation value in the sequence of prompts;

detecting a user interruption of at least one of a spoken confirmation value or a pause during the spoken confirmation, wherein the user interruption comprises an indication that the spoken confirmation value is incorrect;

responsive to detecting the user interruption, halting the sequence of prompts being spoken by the spoken dialog system and collecting from the user a corrected value corresponding to the spoken confirmation value;

providing a new spoken confirmation to the user, wherein the new spoken confirmation includes the corrected value;

repeating the acts of detecting, halting, collecting and providing until values in the new spoken confirmation are correct; and

responsive to lack of detecting a user interruption during the sequence of prompts of he spoken confirmation, confirming the set of values.