Controlling a listening horizon of a speech recognition system for use in handsfree conversational dialog

• US 6,782,364 B2
• Filed: 07/08/2002
• Issued: 08/24/2004
• Est. Priority Date: 05/17/1999
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Horvitz, Eric
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): Storm, Donald L.
• Application Number:
  US10/190,978
• Publication Number:
  US 20020184021A1
• Time in Patent Office: 778 Days
• Field of Search: 704/200, 704/251, 704/270, 704/275
• US Class Current: 704/275
• CPC Class Codes: G06F 3/16 Sound input; Sound output speech processing G10L
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

Conversational dialog with a computer or other processor-based device without requiring push-to-talk functionality, wherein computer-implemented methods probabilistically determines a user desire to engage in a dialog, and/or the context of the desire. Based thereon the method either turns on a speech recognition functionality for a period of time referred to as a listening horizon, or facilitates the user'"'"'s desire. Upon the listening horizon expiring, the method turns off the speech recognition functionality.

Claims

1. A computer-implemented method comprising:

turning on a command recognition functionality for a listening horizon based at least in part on a probability a user would desire to engage in dialog given a particular computational state, wherein turning on a command recognition functionality for a listening horizon comprises determining a length of the listening horizon; and

,

turning off the command recognition functionality after the listening horizon has expired.

2. The method of claim 1, command recognition functionality further comprises speech recognition.

3. The method of claim 1, further comprising providing user feedback during the length of the listening horizon.

4. The method of claim 3, providing user feedback further comprises at least one of displaying a visual indicator to the user and providing an audio indication to the user.

5. The method of claim 3, providing user feedback further comprises providing one or more command options to the user.

6. The method of claim 1, determining a length of the listening horizon further comprises setting a predetermined length of time.

7. The method of claim 1, wherein determining a length of the listening horizon further comprises setting a length of time via a probabilistic analysis.

8. The method of claim 1, wherein the computational state corresponds to a computer query.

9. The method of claim 1, wherein the computational state corresponds to receipt or retrieval of electronic mail.

10. A system for controlling portions of an application, comprising:

a listening component to receive user commands; and

an application component that invokes the listening component in order to process the user commands, the application component having one or more listening segments contained therein, the listening segments associated with a dynamically determined listening horizon having a beginning time and an ending time, wherein the listening segments invoke the listening component automatically based on a determined probability that a user will issue a command.

11. The system of claim 10, the application component including at least one of a scheduling application, a text application, a help application, an e-mail application, and a calendar application.

12. The system of claim 10, the listening segments further comprise at least one question directed to the user in order to generate a response from the user.

13. The system of claim 10, the listening horizon is a function of at least one of a predetermined length, a probabilistic analysis, an inference, a cost-benefit analysis, and an expected utility analysis.

14. The system of claim 10, the listening component is a speech recognizer.

15. The system of claim 10, further comprising a display object that is turned on during the listening horizon.

16. The system of claim 10, the listening horizon is a function of a computational state of the application component.

17. A computer-readable medium having computer-executable instructions stored thereon to execute the components of claim 10.

18. A system for receiving responses to a query, comprising:

means for determining a listening segment having a dynamically determined delay time;

means for automatically turning on the listening segment during at least one portion of a computer process in order to receive the user input based on a determined probability that a user will issue a command; and

means for turning off the listening segment after the delay time.