Unified client-server distributed architectures for spoken dialogue systems

• US 6,738,743 B2
• Filed: 03/28/2001
• Issued: 05/18/2004
• Est. Priority Date: 03/28/2001
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Larson, Jim A., Sharma, Sangita R.
• Primary Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US09/819,832
• Publication Number:
  US 20020143551A1
• Time in Patent Office: 1,147 Days
• Field of Search: 704/270, 704/270.1, 709/218, 709/224, 709/226, 709/249, 709/229
• US Class Current: 704/270.1
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering

Abstract

The invention relates to unified client-server distributed architectures for spoken dialogue systems. Particularly, the invention provides a method, apparatus, and system for dynamically switching between various architectural configurations based upon user functionality and network conditions to implement speech recognition functions for a spoken dialogue system. In one embodiment, a client device is coupled to a server through a network. A configuration selection switch located within the client device selects a configuration based upon user functionality and network conditions to implement speech recognition functions for the spoken dialogue system. For example, the configuration selection switch may select one of a client-based configuration, a server-based configuration, a client-server distributed configuration, or a client-gateway-server distributed configuration based upon user functionality and network conditions. Each of these configurations is selected to provide the user with the most efficient speech recognition for the function being utilized by the user based upon network conditions.

Claims

1. An apparatus for connection to a server over a network, the apparatus comprising:

a client device to be coupled to the server through the network; and

a configuration selection switch to select a configuration based upon user functionality and network conditions to implement speech recognition functions for a spoken dialogue system.

2. The apparatus of claim 1, wherein if user functionality includes utilizing a small vocabulary, the configuration selection switch to select a client-based configuration.

3. The apparatus of claim 1, wherein if network conditions are such that a network connection is not available or not reliable, the configuration selection switch to select a client-based configuration.

4. The apparatus of claim 3, wherein in the client-based configuration, the client device performs speech recognition functions.

5. The apparatus of claim 1, wherein if user functionality includes utilizing a large vocabulary and a network connection is available with sufficient bandwidth for transmitting audio data, the configuration selection switch to select a server-based configuration.

6. The apparatus of claim 5, wherein in the server-based configuration, the server performs speech recognition functions.

7. The apparatus of claim 1, wherein if user functionality includes utilizing a large vocabulary and a network connection is available having reduced bandwidth for transmitting audio data, the configuration selection switch to select a client-server distributed configuration.

8. The apparatus of claim 7, wherein in the client-server distributed configuration, the client device performs the speech recognition functions of analog to digital conversion and feature extraction and the server performs the speech recognition functions of decoding and language processing.

9. The apparatus of claim 1, wherein if the client device does not have computing capabilities and reduced bandwidth to the server for transmitting audio data is desired, the configuration selection switch to select a client-gateway-server distributed configuration.

10. The apparatus of claim 9, wherein in the client-gateway-server distributed configuration, a gateway performs the speech recognition functions of analog to digital conversion and feature extraction and the server performs the speech recognition functions of decoding and language processing.

11. A method comprising:

determining user functionality and network conditions in a spoken dialogue system; and

selecting a configuration for the spoken dialogue system based upon user functionality and network conditions to implement speech recognition functions.

12. The method of claim 11, wherein if user functionality is determined to include utilizing a small vocabulary, a client-based configuration is selected.

13. The method of claim 11, wherein if network conditions are determined to indicate that a network connection is not available or not reliable, a client-based configuration is selected.

14. The method of claim 13, wherein in the client-based configuration, a client device performs speech recognition functions.

15. The method of claim 11, wherein if user functionality is determined to include utilizing a large vocabulary and network conditions are determined to include a network connection that is available with sufficient bandwidth for transmitting audio data, a server-based configuration is selected.

16. The method of claim 15, wherein in the server-based configuration, a server performs speech recognition functions.

17. The method of claim 11, wherein if user functionality is determined to include utilizing a large vocabulary and network conditions are determined to include a network connection that is available having reduced bandwidth for transmitting audio data, a client-server distributed configuration is selected.

18. The method of claim 17, wherein in the client-server distributed configuration, a client device performs the speech recognition functions of analog to digital conversion and feature extraction and a server performs the speech recognition functions of decoding and language processing.

19. The method of claim 11, wherein if a client device does not have computing capabilities and reduced bandwidth to a server for transmitting audio data is desired, a client-gateway-server distributed configuration is selected.

20. The method of claim 19, wherein in the client-gateway-server distributed configuration, a gateway performs the speech recognition functions of analog to digital conversion and feature extraction and the server performs the speech recognition functions of decoding and language processing.

21. A machine-readable medium having stored thereon instructions, which when executed by a machine, cause the machine to perform the following:

determining user functionality and network conditions in a spoken dialogue system; and

selecting a configuration for the spoken dialogue system based upon user functionality and network conditions to implement speech recognition functions.

22. The machine-readable medium of claim 21, wherein if user functionality is determined to include utilizing a small vocabulary, a client-based configuration is selected.

23. The machine-readable medium of claim 21, wherein if network conditions are determined to indicate that a network connection is not available or not reliable, a client-based configuration is selected.

24. The machine-readable medium of claim 23, wherein in the client-based configuration, a client device performs speech recognition functions.

25. The machine-readable medium of claim 21, wherein if user functionality is determined to include utilizing a large vocabulary and network conditions are determined to include a network connection that is available with sufficient bandwidth for transmitting audio data, a server-based configuration is selected.

26. The machine-readable medium of claim 25, wherein in the server-based configuration, a server performs speech recognition functions.

27. The machine-readable medium of claim 21, wherein if user functionality is determined to include utilizing a large vocabulary and network conditions are determined to include a network connection that is available having reduced bandwidth for transmitting audio data, a client-server distributed configuration is selected.

28. The machine-readable medium of claim 21, wherein in the client-server distributed configuration, a client device performs the speech recognition functions of analog to digital conversion and feature extraction and a server performs the speech recognition functions of decoding and language processing.

29. An apparatus comprising:

means for determining user functionality and network conditions in a spoken dialogue system; and

means for selecting a configuration for the spoken dialogue system based upon user functionality and network conditions to implement speech recognition functions.

30. The apparatus of claim 29, wherein if user functionality is determined to include utilizing a small vocabulary, a client-based configuration is selected.

31. The apparatus of claim 29, wherein if user functionality is determined to include utilizing a large vocabulary and network conditions are determined to include a network connection that is available having reduced bandwidth for transmitting audio data, a client-server distributed configuration is selected.