Method and apparatus for improving voice recognition performance in a voice application distribution system

• US 7,643,998 B2
• Filed: 07/02/2002
• Issued: 01/05/2010
• Est. Priority Date: 07/03/2001
• Status: Active Grant

• Current Assignee: HTC Corporation
• Original Assignee: Apptera Incorporated
• Inventors: Chiu, Leo, Yuen, Michael S.
• Primary Examiner(s): Dorvil; Richemond
• Assistant Examiner(s): YEN, ERIC L
• Application Number:
  US10/190,080
• Publication Number:
  US 20030009339A1
• Time in Patent Office: 2,744 Days
• Field of Search: 704/260, 704270-275
• US Class Current: 704/275
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: H04M 3/4938 comprising a voice browser which renders and interprets, e.g. VoiceXML

Abstract

A vocabulary management system for constraining speech recognition processing associated with text-to-speech and speech-to-text rendering associated with use of a voice application in progress between a user accessing a data source through a voice portal has a vocabulary management server connected to a voice application server and to a telephony server, and an instance of vocabulary management software running on the management server for enabling vocabulary establishment and management for speech recognition software. The system is characterized in that an administrator accessing the vocabulary management server uses the software to create unique vocabulary sets that are specific to selected portions of vocabulary associated with target data sources the vocabulary sets differing in content according to administrator direction.

Claims

1. A vocabulary management system for constraining speech recognition processing associated with text-to-speech and speech-to-text rendering with use of a voice application in progress between a user accessing a data source through a voice portal comprising:

a vocabulary management server connected to a voice application server and to a telephony server; and

an instance of vocabulary management software running on the vocabulary management server to enable vocabulary establishment and management for speech recognition software, the instance of vocabulary management software further to render user-specific data and activity-specific data from the data source to the user and further to provide user-specific voice recognition options to each user, wherein the user-specific voice recognition options restrict voice recognition options to options associated with both the user and a specific activity being performed by the vocabulary management system;

characterized in that an administrator accessing the vocabulary management server uses the software to create unique vocabulary sets that are specific to and associated with target data.

2. The system of claim 1 wherein the voice portal is an interactive voice response unit operating in a telephony environment.

3. The system of claim 1 wherein the data source is hosted on the Internet network and accessed from a telephony network.

4. The system of claim 1 wherein the voice application server renders the data as VXML for text to speech synthesis.

5. The system of claim 1 wherein the unique vocabulary sets are associated with unique rules sets for voice presentation to accessing users.

6. The system of claim 1 wherein the unique vocabulary sets are associated with unique vocabulary sub-sets that are specific to particular accessing users.

7. The system of claim 6 wherein the unique vocabulary sub-sets are associated with unique rules sub-sets.

8. The system of claim 1 wherein the vocabulary sets contain both words found at the data source and words introduced by an administrator.

9. A vocabulary management server for processing text-to-speech and speech-to-text rendering associated with use of a voice application in progress between a user accessing a data source through a voice portal, the vocabulary management server comprising:

a vocabulary set created specific to and associated with the data source, wherein the vocabulary set limits data rendered from the data source by the vocabulary management server, wherein the vocabulary management server further provides user-specific voice recognition options to each user, and wherein the user-specific voice recognition options limit voice recognition options to those options associated with both a particular user and a specific activity being requested by the particular user; and

an interface for communicating with a text-to-speech or speech-to-text engine;

wherein the server uses the vocabulary set specific to and associated with speech recognition generating dialog for the voice application in progress.

10. The vocabulary management server of claim 9 wherein the voice portal is an interactive voice response unit operating in a telephony environment.

11. The vocabulary management server of claim 9 wherein the data source is hosted on the Internet network and accessed from a telephony network.

12. The vocabulary management server of claim 9 wherein the voice application server renders the data as VXML for text to speech synthesis.

13. The vocabulary management server of claim 9 wherein the vocabulary set is associated with a unique rules set for voice presentation to accessing users.

14. The vocabulary management server of claim 13 wherein the vocabulary set is associated with a unique vocabulary sub-set that is specific to a particular accessing user.

15. The vocabulary management server of claim 14 wherein the unique vocabulary sub-set is associated with a unique rules sub-set.

16. The vocabulary management server of claim 9 wherein the vocabulary set contains both words found at the data source and words introduced by an administrator.

17. The vocabulary management server of claim 9 wherein voice is rendered to the user over a telephone receiver.

18. A method for performing speech recognition, the method comprising:

(a) providing a vocabulary management software instance for creating, editing, and organizing vocabulary sets for voice recognition, wherein the vocabulary management software instance limits data rendered from a data source to a user;

(b) creating a vocabulary set using the software, the vocabulary set containing a portion of available vocabulary words specific to and associated with a data source to be used in dialog creation by the voice application;

(c) configuring the voice application to use the created vocabulary set when a speech recognition portion of the application is triggered;

(d) configuring the voice application to provide user-specific voice recognition options for each user, wherein the user-specific voice recognition options restrict voice recognition options to options associated with both the user and a specific activity being performed by the user; and

(e) deploying the voice application to execution.

19. The method of claim 18 wherein in step (a) the vocabulary management software has access to application logic and a database resource adapter of the voice application system.

20. The method of claim 18 wherein the vocabulary sets are also specific to client activity in association with the data source.

21. The method of claim 18 wherein in step (a) the vocabulary sets are associated with one or more vocabulary subsets.

22. The method of claim 18 wherein in step (b) the software is manipulated through administrator input from a connected workstation hosting client voice application software.

23. The method of claim 18 wherein in step (b) the vocabulary words in the set also contain words introduced by an administrator.

24. The method of claim 18 wherein in step (c) configuration includes accessing applicable rules associated with the vocabulary set.

25. The method of claim 24 wherein in step (c) configuration further includes access of vocabulary subset organized under the vocabulary set and a rules subset organized under the rules set wherein sub components are specific to a single client.

26. The method of claim 18 wherein in step (e) the voice application is stored for execution in the voice application server.

27. The method of claim 26 wherein the execution includes distribution resulting from trigger by an accessing user, or presenting dialog to the user by way of outbound calling.