Automated centralized updating of speech recognition systems

• US 6,456,975 B1
• Filed: 01/13/2000
• Issued: 09/24/2002
• Est. Priority Date: 01/13/2000
• Status: Expired due to Term

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Chang, Eric I-Chao
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US09/482,738
• Time in Patent Office: 985 Days
• Field of Search: 704/270, 704/275, 704/243-245, 704/231, 704/251, 704/255-257, 704/270.1
• 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
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting

Abstract

In one embodiment, a speech recognition program at a client receives data that is unrecognized, such as an unrecognized word, an unrecognized pronunciation of a known word, an unrecognized dialect of a known, and/or a substantially new word frequency usage. The client transmits the data to a provider, which processes the data into known data, and transmits the known data back to a number of clients, including the client that initially sent the unrecognized data. In one embodiment, the unrecognized data is sent from the client to the provider via a third party, to anonymize the data.

Claims

1. A computer-implemented method comprising:

receiving unrecognized data by a speech recognition program at a particular client;

transmitting the unrecognized data from the particular client to a provider;

processing the unrecognized data at the provider into a known data; and

transmitting the known data to at least one client other than the particular client.

2. The method of claim 1, wherein transmitting the unrecognized data from the particular client to a provider comprises transmitting the unrecognized data from the particular client to the provider by way of a third party to at least anonymize the unrecognized data.

3. The method of claim 1, wherein transmitting the unrecognized data from the particular client to a provider comprises transmitting the unrecognized data from the particular client to the provider by way of a third party to at least group the particular client with other clients within a predetermined linguistic class.

4. The method of claim 1, wherein receiving unrecognized data comprises receiving unrecognized data comprising at least one of:

data representing an unrecognized word;

data representing an unrecognized pronunciation of a known word;

data representing an unrecognized dialect of a known word; and

, data representing a substantially new word frequency usage.

5. The method of claim 1, wherein receiving unrecognized data comprises receiving a speech waveform.

6. The method of claim 1, wherein receiving unrecognized data comprises receiving the unrecognized data in an implicit manner.

7. The method of claim 1, wherein receiving unrecognized data comprises receiving the unrecognized data in an explicit manner.

8. A computer-implemented method comprising:

receiving unrecognized data by a speech recognition program at a particular client;

transmitting the unrecognized data from the particular client to a third party to group the particular client with other clients within a predetermined linguistic class;

transmitting the unrecognized data from the third party to a provider;

processing the unrecognized data at the provider into known data; and

,

transmitting the known data to at least one client including the particular client.

9. The method of claim 8, wherein receiving unrecognized data comprises receiving unrecognized data comprising at least one of:

data representing an unrecognized word;

data representing an unrecognized pronunciation of a known word;

data representing an unrecognized dialect of a known word; and

, data representing a substantially new word frequency usage.

10. The method of claim 8, wherein receiving unrecognized data comprises receiving a speech waveform.

11. The method of claim 8, wherein receiving unrecognized data comprises receiving the unrecognized data in one of an explicit manner and an implicit manner.

12. A machine-readable medium having instructions stored thereon for execution by a processor of a speech recognition provider to perform a method comprising:

receiving unrecognized data generated by a speech recognition program of a particular client, and initially sent by the particular client;

processing the unrecognized data into known data, wherein the known data is associated with one or more linguistic groups; and

,

transmitting the known data to a plurality of clients associated with said one or more different linguistic groups.

13. The medium of claim 12, wherein receiving unrecognized data comprises receiving the unrecognized data directly from the particular client.

14. The medium of claim 12, wherein receiving unrecognized data comprises receiving the unrecognized data from a third party to which the particular client directly sent the unrecognized data, such that an identity of the particular client remains unknown to the speech recognition provider.

15. The medium of claim 12, wherein receiving unrecognized data comprises receiving unrecognized data comprising at least one of:

data representing an unrecognized word;

data representing an unrecognized pronunciation of a known word;

data representing an unrecognized dialect of a known word; and

, data representing a substantially new word frequency usage.

16. The medium of claim 12, wherein receiving unrecognized data comprises receiving a speech waveform.

17. A computerized system comprising:

a particular client capable of having a speech recognition program running thereon designed to receive recognized data and unrecognized data; and

,

a provider designed to process the unrecognized data into known data and to transmit the known data to at least one client including the particular client, the particular client being further designed to transmit unrecognized data to the provider such that unrecognized data received by the provider is substantially independent of recognized data.

18. The system of claim 17, further comprising a third party designed at least to receive the unrecognized data from the particular client and to transmit the unrecognized data to the provider so that the particular client remains unknown to the provider.

19. The system of claim 18, wherein the third party is further designed to group the particular client with other clients within a predetermined linguistic class.

20. The system of claim 17, wherein the unrecognized data comprising at least one of:

data representing an unrecognized word;

data representing an unrecognized pronunciation of a known word;

data representing an unrecognized dialect of a known word; and

, data representing a substantially new word frequency usage.

21. The system of claim 17, wherein the unrecognized data comprises a speech waveform.

22. A client computer system configured to perform speech recognition, comprising:

a microphone capable of detecting sound data;

means for

receiving unrecognized sound data from the microphone;

transmitting the unrecognized sound data to a provider; and

,

receiving known data from the provider previously processed by the provider from unrecognized sound data.

23. The client computer system of claim 22, wherein the means is to transmit the unrecognized sound data to the provider by way of a third party to at least anonymize the unrecognized sound data.

24. The client computer system of claim 22, wherein the means is to transmit the unrecognized sound data to the provider by way of a third party to at least group the client computer system with other client computer systems within a predetermined linguistic class.

25. The client computer system of claim 22, wherein the unrecognized sound data comprises at least one of:

data representing an unrecognized word;

data representing an unrecognized pronunciation of a known word;

data representing an unrecognized dialect of a known word; and

, data representing a substantially new word frequency usage.

26. The client computer system of claim 22, wherein the unrecognized sound data comprises a speech waveform.