METHOD AND SYSTEM FOR CONSIDERING INFORMATION ABOUT AN EXPECTED RESPONSE WHEN PERFORMING SPEECH RECOGNITION

• US 20140288933A1
• Filed: 06/10/2014
• Published: 09/25/2014
• Est. Priority Date: 02/04/2005
• Status: Active Grant

• Current Assignee: Vocollect, Inc. (Honeywell International Inc.)
• Original Assignee: Vocollect, Inc. (Honeywell International Inc.)
• Inventors: Braho, Keith, El-Jaroudi, Amro, Pike, Jeffrey
• Granted Patent: US 10,068,566 B2
• Time in Patent Office: Days
• US Class Current: 704/251
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning

Abstract

A speech recognition system receives and analyzes speech input from a user in order to recognize and accept a response from the user. Under certain conditions, information about the response expected from the user may be available. In these situations, the available information about the expected response is used to modify the behavior of the speech recognition system by taking this information into account. The modified behavior of the speech recognition system comprises adjusting the rejection threshold when speech input matches the predetermined expected response.

Claims

1. A method for recognizing speech comprising the steps of:

in a speech dialog, determining an expected response to be received as input speech from a user;

based upon the determined expected response, modifying a speech recognition match/search algorithm to tailor it to favor the expected response;

processing input speech from a user with the modified match/search algorithm to generate a hypothesis and confidence factor for the input speech.

2. The method of claim 1, comprising comparing the hypothesis confidence factor to an acceptance threshold for accepting the hypothesis.

3. The method of claim 1, wherein:

the match/search algorithm processes the input speech through a plurality of states; and

the method comprises modifying transition probabilities associated with transitions between multiple states in the match/search algorithm.

4. The method of claim 1, wherein:

the match/search algorithm processes the input speech through a plurality of states; and

the method comprises modifying initial state probabilities associated with a path through multiple states in the match/search algorithm.

5. The method of claim 1, wherein:

the match/search algorithm utilizes multiple paths through acoustic models for the input speech; and

the method comprises modifying the acoustic models based on the expected response.

6. The method of claim 1, wherein:

the match/search algorithm utilizes multiple paths through acoustic models for the input speech; and

the method comprises modifying the insertion penalty associated with a given acoustic model.

7. The method of claim 1, wherein:

processing of the input speech results in frames of the features of the input speech; and

the method comprises;

utilizing multiple paths through acoustic models for the input speech feature frames to generate scores; and

applying at least one of a bias or penalty per frame to the input speech to affect the scores of paths through the models.

8. An apparatus for recognizing speech, comprising:

processing circuitry including a match/search algorithm for performing speech recognition;

the match search algorithm being dynamically modifiable based on an expected response to be received as input speech to tailor it to favor the expected response from a user, and further being operable for processing input speech from a user as modified to generate a hypothesis and confidence factor for the input speech.

9. The apparatus of claim 8, wherein the processing circuitry is operable for comparing the hypothesis confidence factor to an acceptance threshold for accepting the hypothesis.

10. The apparatus of claim 8, wherein the match/search algorithm processes the input speech through a plurality of states and is modifiable by modifying transition probabilities associated with transitions between multiple states in the match/search algorithm.

11. The apparatus of claim 8, wherein the match/search algorithm processes the input speech through a plurality of states and is modifiable by modifying initial state probabilities associated with a path through multiple states in the match/search algorithm.

12. The apparatus of claim 8, wherein the match/search algorithm utilizes multiple paths through acoustic models for the input speech and is modifiable by modifying the acoustic models based on the expected response.

13. The apparatus of claim 8, wherein the match/search algorithm utilizes multiple paths through acoustic models for the input speech and is modifiable by modifying the insertion penalty associated with a given acoustic model.

14. The apparatus of claim 8, wherein:

the processing circuitry creates frames of the features of the input speech; and

the match/search algorithm utilizes multiple paths through acoustic models for the input speech feature frames to generate scores, the algorithm applying at least one of a bias or penalty per frame to the input speech to affect the scores of paths through the model.

15. A method for recognizing speech, comprising:

in a speech dialog, determining, with a processor, an expected response to be received as input speech from a user;

based upon the determined expected response, modifying, with the processor, a speech recognition match/search algorithm to tailor it to favor the expected response;

processing, with the processor, input speech from a user with the modified match/search algorithm to generate a hypothesis, confidence factor, and frames of features for the input speech.

16. The method of claim 15, comprising comparing the hypothesis confidence factor to an acceptance threshold for accepting the hypothesis.

17. The method of claim 15, wherein:

the match/search algorithm processes the input speech through a plurality of states; and

the method comprises modifying transition probabilities associated with transitions between multiple states in the match/search algorithm.

18. The method of claim 15, wherein:

the match/search algorithm processes the input speech through a plurality of states; and

the method comprises modifying initial state probabilities associated with a path through multiple states in the match/search algorithm.

19. The method of claim 15, wherein:

the match/search algorithm utilizes multiple paths through acoustic models for the input speech; and

the method comprises modifying the acoustic models based on the expected response.

20. The method of claim 15, wherein:

the match/search algorithm utilizes multiple paths through acoustic models for the input speech; and

the method comprises modifying the insertion penalty associated with a given acoustic model.