Method and apparatus for providing a dynamic speech-driven control and remote service access system

• US 7,013,275 B2
• Filed: 12/28/2001
• Issued: 03/14/2006
• Est. Priority Date: 12/28/2001
• Status: Active Grant

• Current Assignee: Voice Inventions LLC
• Original Assignee: SRI International, Inc.
• Inventors: Arnold, James F., Franco, Horacio E., Israel, David J., Culy, Christopher
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): Lerner, Martin
• Application Number:
  US10/033,772
• Publication Number:
  US 20030125955A1
• Time in Patent Office: 1,537 Days
• Field of Search: 704/231, 704/233, 704/243, 704/244, 704/246, 704/250, 704/251, 704/255, 704/270, 704/270.1, 704/275
• US Class Current: 704/244
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

An apparatus and a concomitant method for speech recognition. In one embodiment, a distributed speech recognition system provides speech-driven control and remote service access. The distributed speech recognition system comprises a client device and a central server.

Claims

1. Method for performing speech recognition, said method comprising the steps of:

(a) receiving a speech signal locally from a user via a client device;

(b) performing speech recognition on said speech signal in accordance with an embedded speech recognizer of said client device to produce a recognizable text signal, wherein said embedded speech recognizer employs a language model and a natural language understanding module;

(c) forwarding said recognizable text signal to a remote server; and

(d) updating said language model by dynamically receiving an update from said remote server in accordance with said recognizable text signal.

2. The method of claim 1, wherein said remote server generates said update in accordance with a task model.

3. The method of claim 2, wherein said remote server monitors one or more states of said task model to determine progress toward satisfying a goal of said user.

4. The method of claim 1, further comprising the step of:

(e) storing at least a portion of said language model in a cache of said client device.

5. The method of claim 1, further comprising the step of:

(b1) adapting said performance of speech recognition based on at least one local parameter of said speech signal.

6. The method of claim 5, wherein said at least one local parameter is representative of an environmental noise.

7. The method of claim 5, wherein said at least one local parameter is representative of an acoustic environment.

8. The method of claim 5, wherein said at least one local parameter is representative of a pronunciation of said user.

9. Method for performing speech recognition, said method comprising the steps of:

(a) receiving a recognizable text signal representative of a user speech signal from a client device, wherein said recognizable text is generated using a speech recognizer having a language model and a natural language understanding module on said client device;

(b) processing said recognizable text signal in accordance with a task model; and

(c) forwarding a language model update to said client device in accordance with said recognizable text signal.

10. The method of claim 9, wherein a remote server monitors one or more states of said task model to determine progress toward satisfying a goal of said user.

11. The method of claim 9, wherein said recognizable text is further generated in accordance with adapting said performance of speech recognition based on at least one local parameter of said speech signal.

12. The method of claim 11, wherein said at least one local parameter is representative of an environmental noise.

13. The method of claim 11, wherein said at least one local parameter is representative of an acoustic environment.

14. The method of claim 11, wherein said at least one local parameter is representative of a pronunciation of said user.

15. A distributed system for performing speech recognition, said system comprising:

a client device for receiving a speech signal locally from a user, said client device having an embedded speech recognizer with a language model and a natural language understanding module for performing speech recognition on said speech signal to produce a recognizable text signal, and wherein said embedded speech recognizer further adapts said performance of speech recognition based on at least one local parameter of said speech signal; and

a remote server for receiving said recognizable text signal and forwarding a language model update to said client device in accordance with said recognizable text signal.

16. A client device for performing speech recognition, said client device comprising:

means for receiving a speech signal locally from a user;

means for performing speech recognition on said speech signal to produce a recognizable text signal, wherein said speech recognition means employs a language model and a natural language understanding module;

means for forwarding said recognizable text signal to a remote server; and

means for updating said language model by dynamically receiving an update from said remote server in accordance with said recognizable text signal.

17. The client device of claim 16, further comprising:

a cache for storing at least a portion of said language model.

18. The client device of claim 16, further comprising:

means for adapting said performance of speech recognition based on at least one local parameter of said speech signal.

19. The client device of claim 18, wherein said at least one local parameter is representative of an environmental noise.

20. The client device of claim 18, wherein said at least one local parameter is representative of an acoustic environment.

21. The client device of claim 18, wherein said at least one local parameter is representative of a pronunciation of said user.

22. A server for performing speech recognition, said server comprising:

means for receiving a recognizable text signal representative of a user speech signal from a client device, wherein said recognizable text is generated using a speech recognizer having a language model and a natural language understanding module on said client device;

means for processing said recognizable text signal in accordance with a task model; and

means for forwarding a language model update to said client device in accordance with said recognizable text signal.

23. The server of claim 22, wherein said forwarding means is a grammar manager.

24. The server of claim 22, further comprising:

means for monitoring one or more states of said task model to determine progress toward satisfying a goal of said user.

25. The server of claim 24, wherein said monitoring means is a dialog manager.

26. The server of claim 22, wherein said recognizable text is further generated in accordance with adapting said performance of speech recognition based on at least one local parameter of said speech signal.

27. The server of claim 26, wherein said at least one local parameter is representative of an environmental noise.

28. The server of claim 26, wherein said at least one local parameter is representative of an acoustic environment.

29. The server of claim 26, wherein said at least one local parameter is representative of a pronunciation of said user.

30. A computer-readable medium having stored thereon a plurality of instructions, the plurality of instructions including instructions which, when executed by a processor, cause the processor to perform the steps comprising of:

(a) receiving a speech signal locally from a user via a client device;

(b) performing speech recognition on said speech signal in accordance with an embedded speech recognizer of said client device to produce a recognizable text signal, wherein said embedded speech recognizer employs a language model and a natural language understanding module;

(c) adapting said performance of speech recognition based on at least one local parameter of said speech signal;

(d) forwarding said recognizable text signal to a remote server; and

(e) updating said language model by dynamically receiving an update from said remote server in accordance with said recognizable text signal.

31. A computer-readable medium having stored thereon a plurality of instructions, the plurality of instructions including instructions which, when executed by a processor, cause the processor to perform the steps comprising of:

(a) receiving a recognizable text signal representative of a user speech signal from a client device, wherein said recognizable text is generated using a speech recognizer having a language model and a natural language understanding module on said client device, and wherein said recognizable text is generated in accordance with adapting said performance of speech recognition based on at least one local parameter of said speech signal;

(b) processing said recognizable text signal in accordance with a task model; and

(c) forwarding a language model update to said client device in accordance with said recognizable text signal.