Voice recognition using barge-in time

• US 20030093274A1
• Filed: 11/09/2001
• Published: 05/15/2003
• Est. Priority Date: 11/09/2001
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Netbytel Incorporated (West Corporation)
• Inventors: Thompson, Scott A.
• Granted Patent: US 7,069,213 B2
• Time in Patent Office: Days
• US Class Current: 704/246
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A system, method and computer program product for automated voice recognition using the barge-in time of a user'"'"'s utterance is provided. A list of items is audibly provided to a user. The user selects one item from the list by providing an utterance representing the user'"'"'s selection. The utterance time (i.e., the barge-in time) is calculated from the utterance of the user. Then, the utterance is compared to each item in the list so as to generate a confidence score for each item in the list, with the confidence score being based on the similarity between the item and the utterance. The confidence scores are also based on the barge-in time. One of the items from the list is selected based on the confidence scores of the items.

Claims

1. A method for automated voice recognition, said method comprising the steps of:

providing an audible list containing a plurality of items;

receiving an utterance indicating a selected one of the items from the list; and

matching the utterance with a matched one of the items from the list based at least partially on a time that the utterance was received.

2. The method of claim 1, wherein the matching step includes the sub-steps of:

comparing the utterance to each of the items in the list so as to generate a confidence score for each of the items in the list; and

selecting the matched one of the items based on the confidence scores of the items in the list.

3. The method of claim 2, wherein the comparing sub-step includes:

calculating an initial confidence score for each of the items in the list based on a comparison of the utterance to the item; and

selectively modifying the initial confidence scores based on the time that the utterance was received relative to the items in the list.

4. The method of claim 3, wherein the comparison of the utterance to the item is performed using natural language processing.

5. The method of claim 3, wherein in the sub-step of selectively modifying the initial confidence scores, the initial confidence score of the item that was provided immediately before the time that the utterance was received is increased by a first predetermined amount.

6. The method of claim 5, wherein in the sub-step of selectively modifying the initial confidence scores, the initial confidence score of the item that precedes the item that was provided immediately before the time that the utterance was received is increased by a second predetermined amount, which is less than the first predetermined amount.

7. The method of claim 3, wherein in the sub-step of selectively modifying the initial confidence scores, the initial confidence scores of the items that were provided after the time that the utterance was received are decreased.

8. The method of claim 3, wherein in the sub-step of selectively modifying the initial confidence scores, none of the initial confidence scores are modified if at least one of the initial confidence scores exceeds a predetermined threshold.

9. The method of claim 1, wherein in the matching step, the utterance is matched with the matched one of the items based on the degree of proximity between the time of the utterance and a time when each of the items was provided.

10. The method of claim 1, wherein the matching step includes the sub-step of determining the time that the utterance was received based on an echo cancellation time for the utterance.

11. A machine-readable medium encoded with a program for automated voice recognition, said program containing instructions for performing the steps of:

providing an audible list containing a plurality of items;

receiving an utterance indicating a selected one of the items from the list; and

matching the utterance with a matched one of the items from the list based at least partially on a time that the utterance was received.

12. The machine-readable medium of claim 11, wherein the matching step includes the sub-steps of:

comparing the utterance to each of the items in the list so as to generate a confidence score for each of the items in the list; and

selecting the matched one of the items based on the confidence scores of the items in the list.

13. The machine-readable medium of claim 12, wherein the comparing sub-step includes:

calculating an initial confidence score for each of the items in the list based on a comparison of the utterance to the item; and

selectively modifying the initial confidence scores based on the time that the utterance was received relative to the items in the list.

14. The machine-readable medium of claim 13, wherein in the sub-step of selectively modifying the initial confidence scores, the initial confidence score of the item that was provided immediately before the time that the utterance was received is increased by a first predetermined amount.

15. The machine-readable medium of claim 13, wherein in the sub-step of selectively modifying the initial confidence scores, the initial confidence scores of the items that were provided after the time that the utterance was received are decreased.

16. The machine-readable medium of claim 13, wherein in the sub-step of selectively modifying the initial confidence scores, none of the initial confidence scores are modified if at least one of the initial confidence scores exceeds a predetermined threshold.

17. The machine-readable medium of claim 11, wherein in the matching step, the utterance is matched with the matched one of the items based on the degree of proximity between the time of the utterance and a time when each of the items was provided.

18. The machine-readable medium of claim 11, wherein the matching step includes the sub-step of determining the time that the utterance was received based on an echo cancellation time for the utterance.

19. An automated voice recognition system comprising:

a first input for receiving a list containing a plurality of items;

a second input for receiving an utterance indicating a selected one of the items from the list;

a third input for receiving a time that the utterance was received; and

a comparator for outputting a matched one of the items from the list based on the first, second, and third inputs.

20. The automated voice recognition system of claim 19, wherein the comparator compares the utterance to each of the items in the list so as to generate a confidence score for each of the items in the list, and selects the matched one of the items based on the confidence scores of the items in the list.

21. The automated voice recognition system of claim 20, wherein the comparator generates the confidence scores by:

calculating an initial confidence score for each of the items in the list based on a comparison of the utterance to the item; and

selectively modifying the initial confidence scores based on the time that the utterance was received relative to the items in the list.

22. The automated voice recognition system of claim 21, wherein the initial confidence score of the item that was provided immediately before the time that the utterance was received is increased by a first predetermined amount.

23. The automated voice recognition system of claim 21, wherein none of the initial confidence scores are modified if at least one of the initial confidence scores exceeds a predetermined threshold.

24. The automated voice recognition system of claim 19, wherein the comparator matches the utterance with the matched one of the items based on the degree of proximity between the time of the utterance and a time when each of the items was provided.