Method, apparatus, and computer program product for one-step correction of voice interaction

• US 7,720,684 B2
• Filed: 04/29/2005
• Issued: 05/18/2010
• 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: Pieraccini, Roberto, Huerta, Juan Manuel
• Primary Examiner(s): Hudspeth; David R
• Assistant Examiner(s): GUERRA-ERAZO, EDGAR X
• Application Number:
  US11/117,951
• Publication Number:
  US 20060247913A1
• Time in Patent Office: 1,845 Days
• Field of Search: 704/201, 704/275, 704/270.1, 704/E15.022, 704/E15.04, 704/E15.044, 704/257, 704/231, 704/270
• US Class Current: 704/257
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A one-step correction mechanism for voice interaction is provided. Correction of a previous state is enabled simultaneously with recognition in a current or subsequent state. An application is decomposed into a set of tasks. Each task is associated with the collection of one piece of information. Each task may be in a different state. At any point during the interaction, while a task/state pair is active, the dialog manager may enable multiple other task/state pairs to be active in latent fashion. The application developer may then use those facilities or resources to the active task/state and the latent task/state pairs depending on contextual condition of the interaction state of the application.

Claims

1. A method, in an interactive voice response system, for processing corrections in a voice interaction, the method comprising:

receiving first information from a first input spoken by a user for a first interaction state;

prompting the user for second information for a second interaction state;

in response to receiving a second input spoken by the user that complies with a latent state grammar for the first interaction state,

correcting the first information in accordance with the second input, and

providing a first correction confirmation and a re-prompting of the user for the second information within a single interaction prompt; and

in response to receiving a third input that is spoken by the user after the second input and that complies with the latent state grammar for the first interaction state,

correcting the first information in accordance with the third input,

providing a second correction confirmation, and

after receiving a user acceptance of the second correction confirmation, re-prompting the user for the second information.

2. The method of claim 1, wherein providing the first correction confirmation comprises echoing the corrected first information.

3. The method of claim 1, wherein the third input is a next input spoken by the user after the second input.

4. The method of claim 1, further comprising invoking a failure handler in response to receiving a fourth input that is spoken by the user after the third input and that complies with the latent state grammar for the first interaction state.

5. The method of claim 1, further comprising:

if the second input complies with an active state grammar for the second interaction state, prompting the user for third information for a third interaction state.

6. The method of claim 5, wherein prompting the user for the third information comprises echoing the second information.

7. An apparatus for processing corrections in a voice interaction with an interactive voice response system, the apparatus comprising:

a speech recognizer;

a prompt player; and

a dialog manager configured to;

receive from the speech recognizer first information from a first input spoken by a user for a first interaction state;

prompt the user via the prompt player for second information for a second interaction state; and

in response to receiving a second input spoken by the user that complies with a latent state grammar for the first interaction state,

correct the first information in accordance with the second input, and

provide a first correction confirmation and a re-prompting of the user for the second information within a single interaction prompt via the prompt player; and

in response to receiving a third input that is spoken by the user after the second input and that complies with the latent state grammar for the first interaction state,

correct the first information in accordance with the third input,

provide a second correction confirmation, and

after receiving a user acceptance of the second correction confirmation, re-prompt the user via the prompt player for the second information.

8. The apparatus of claim 7, wherein the dialog manager is configured to provide the first correction confirmation by echoing the corrected first information.

9. The apparatus of claim 7, wherein the dialog manager is configured to receive the third input as a next input spoken by the user after the second input.

10. The apparatus of claim 7, wherein the dialog manager is further configured to invoke a failure handler in response to receiving a fourth input that is spoken by the user after the third input and that complies with the latent state grammar for the first interaction state.

11. The apparatus of claim 7, wherein the dialog manager is further configured to:

if the second input complies with an active state grammar for the second interaction state, prompt the user via the prompt player for third information for a third interaction state.

12. The apparatus of claim 11, wherein the dialog manager is further configured to:

if the second input complies with the active state grammar for the second interaction state, echo the second information.

13. At least one computer-readable storage medium encoded with a plurality of computer-executable instructions that, when executed, perform a method comprising:

receiving first information from a first input spoken by a user for a first interaction state;

prompting the user for second information for a second interaction state;

in response to receiving a second input spoken by the user that complies with a latent state grammar for the first interaction state,

correcting the first information in accordance with the second input, and

providing a first correction confirmation and a re-prompting of the user for the second information within a single interaction prompt; and

in response to receiving a third input that is spoken by the user after the second input and that complies with the latent state grammar for the first interaction state,

correcting the first information in accordance with the third input,

providing a second correction confirmation, and

after receiving a user acceptance of the second correction confirmation, re-prompting the user for the second information.

14. The at least one computer-readable storage medium of claim 13, wherein providing the first correction confirmation comprises echoing the corrected first information.

15. The at least one computer-readable storage medium of claim 13, wherein the third input is a next input spoken by the user after the second input.

16. The at least one computer-readable storage medium of claim 13, wherein the method further comprises invoking a failure handler in response to receiving a fourth input that is spoken by the user after the third input and that complies with the latent state grammar for the first interaction state.

17. The at least one computer-readable storage medium of claim 13, wherein the method further comprises:

if the second input complies with an active state grammar for the second interaction state, prompting the user for third information for a third interaction state.

18. The at least one computer-readable storage medium of claim 17, wherein prompting the user for the third information comprises echoing the second information.