Speech recognition adaptation systems based on adaptation data

• US 9,495,966 B2
• Filed: 06/29/2012
• Issued: 11/15/2016
• Est. Priority Date: 05/31/2012
• Status: Active Grant

• Current Assignee: Elwha LLC (Intellectual Ventures LLC)
• Original Assignee: Elwha LLC (Intellectual Ventures LLC)
• Inventors: Levien, Royce A., Lord, Richard T., Lord, Robert W., Malamud, Mark A.
• Primary Examiner(s): Neway, Samuel G
• Application Number:
  US13/538,855
• Publication Number:
  US 20130325446A1
• Time in Patent Office: 1,600 Days
• Field of Search: 704200-504
• US Class Current: 1/1
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

The instant application includes computationally-implemented systems and methods that include acquiring indication of a speech-facilitated transaction between a particular party and a target device, receiving adaptation data correlated to the particular party, the receiving facilitated by a particular device associated with the particular party, processing audio data from the particular party at least partly using the received adaptation data correlated to the particular party, and updating the adaptation data based at least in part on a result of the processed audio data, such that the updated adaptation data is configured to be transmitted to the particular device. In addition to the foregoing, other aspects are described in the claims, drawings, and text.

Claims

1. A computationally-implemented method, comprising:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving, from a particular device associated with the particular party, an indication of a location on at least one server of adaptation data correlated to the particular party;

receiving, at the target device, the adaptation data correlated to the particular party from the location on the at least one server indicated by the particular device associated with the particular party;

processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device; and

updating the adaptation data on the at least one server based at least in part on a result of the processing speech from the particular party received via the target device.

2. The computationally-implemented method of claim 1, wherein said acquiring an indication of a speech interaction between a particular party and a target device includes:

acquiring an indication of an interaction in which the particular party interacts with the target device at least partly using speech and partly interacting with one or more portions of the target device.

3. The computationally-implemented method of claim 1, wherein said acquiring an indication of a speech interaction between a particular party and a target device includes:

acquiring an indication that the particular party is within a particular proximity of the target device.

4. The computationally-implemented method of claim 1, wherein said acquiring an indication of a speech interaction between a particular party and a target device includes:

acquiring an indication that the particular party is speaking to the target device.

5. The computationally-implemented method of claim 4, wherein said acquiring an indication that the particular party is speaking to the target device includes:

acquiring an indication that the particular party is speaking to the target device based on detection of one or more words spoken by the particular party.

6. The computationally-implemented method of claim 5, wherein said acquiring an indication that the particular party is speaking to the target device based on detection of one or more words spoken by the particular party includes:

acquiring an indication that the particular party is speaking to the target device based on detection of the particular party speaking a target sentence.

10. The computationally-implemented method of claim 1, further comprising:

receiving adaptation data correlated to the particular party from the particular device.

11. The computationally-implemented method of claim 1, wherein said receiving, at the target device, the adaptation data correlated to the particular party from the location on the at least one server indicated by the particular device associated with particular party includes:

receiving adaptation data that includes at least one speech characteristic of the particular party, said receiving facilitated by a particular device associated with the particular party.

12. The computationally-implemented method of claim 1, wherein said receiving, at the target device, the adaptation data correlated to the particular party from the location on the at least one server indicated by the particular device associated with particular party includes:

receiving adaptation data that includes instructions for adapting one or more speech recognition components, said adaptation data received from a particular device associated with the particular party.

13. The computationally-implemented method of claim 1, wherein said receiving, at the target device, the adaptation data correlated to the particular party from the location on the at least one server indicated by the particular device associated with particular party includes:

receiving adaptation data that includes data linking pronunciation of one or more phonemes by the particular party to one or more concepts, said receiving facilitated by a particular device associated with the particular party.

14. The computationally-implemented method of claim 1, wherein said processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device includes:

applying the received adaptation data correlated to the particular party to the speech from the particular party.

15. The computationally-implemented method of claim 1, wherein said processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device includes:

applying the received adaptation data correlated to the particular party to a speech recognition component of the target device, wherein the received adaptation data includes a pronunciation dictionary.

16. The computationally-implemented method of claim 1, wherein said processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device includes:

processing received speech from the particular party at least partly using the received adaptation data correlated to the particular party.

17. The computationally-implemented method of claim 16, wherein said processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device includes:

processing received speech from the particular party at least partly using a received pronunciation dictionary correlated to the particular party.

18. The computationally implemented method of claim 17, wherein said processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device includes:

replacing a pronunciation definition of at least one word stored in a pronunciation dictionary of the target device with a corresponding pronunciation definition of at least one word stored in the received pronunciation dictionary; and

processing the received speech using the pronunciation dictionary of the target device including the replaced pronunciation definition of the at least one word.

19. The computationally-implemented method of claim 1, wherein said updating the adaptation data on the at least one server based at least in part on a result of the processing speech from the particular party received via the target device includes:

updating the adaptation data based at least in part on a received result of the processed speech, such that the updated adaptation data is configured to be transmitted to the particular device.

20. The computationally-implemented method of claim 19, wherein said updating the adaptation data on the at least one server based at least in part on a result of the processing speech from the particular party received via the target device includes:

updating the adaptation data based at least in part on a received result indicating a subjective opinion of the particular party regarding a success of the speech-facilitated transaction.

23. The computationally-implemented method of claim 1, wherein said updating the adaptation data on the at least one server based at least in part on a result of the processing speech from the particular party received via the target device includes:

updating the adaptation data based at least in part on a received result indicating a subjective state of the particular party after completing the speech interaction.

25. The computationally-implemented method of claim 23, wherein said updating the adaptation data based at least in part on a received result indicating a subjective state of the particular party after completing the speech interaction includes:

updating the adaptation data based at least in part on a received result from the particular device indicating the particular party'"'"'s subjective state after completing the speech interaction.

26. The computationally-implemented method of claim 25, wherein said updating the adaptation data based at least in part on a received result from the particular device indicating the particular party'"'"'s subjective state after completing the speech interaction includes:

updating the adaptation data based at least in part on a received result from the particular device indicating the particular party'"'"'s subjective state inputted into the particular device by the particular party after completing the speech interaction.

35. The computationally-implemented method of claim 1, wherein said updating the adaptation data on the at least one server based at least in part on a result of the processing speech from the particular party received via the target device includes:

determining that the adaptation data should not be modified based at least in part on a result of the processed speech.

37. The computationally-implemented method of claim 1, wherein said updating the adaptation data on the at least one server based at least in part on a result of the processing speech from the particular party received via the target device includes:

updating the adaptation data based at least in part on a determined result of the processed speech.

38. The computationally-implemented method of claim 37, wherein said updating the adaptation data based at least in part on a determined result of the processed speech includes:

updating the adaptation data based at least in part on a result calculated by an inferred success of the speech interaction.

39. The computationally-implemented method of claim 38, wherein said updating the adaptation data based at least in part on a result calculated by an inferred success of the speech interaction includes:

updating the adaptation data based at least in part on a result calculated by an inferred success of the speech interaction, inferred from at least one characteristic of the received speech from the particular party.

40. The computationally-implemented method of claim 39, wherein said updating the adaptation data based at least in part on a result calculated by an inferred success of the speech interaction, inferred from at least one characteristic of the received speech from the particular party includes:

updating the adaptation data based at least in part on a result calculated by an inferred success of the speech interaction, inferred from a type of word used in the received speech from the particular party.

43. The computationally-implemented method of claim 1, wherein said updating the adaptation data on the at least one server based at least in part on a result of the processing speech from the particular party received via the target device includes:

updating the adaptation data based at least in part on a confidence rate of the processed speech.

44. The computationally-implemented method of claim 1, wherein said receiving, from a particular device associated with the particular party, an indication of a location on at least one server of adaptation data correlated to the particular party includes:

receiving, from the particular device associated with the particular party, at least one of;

an address, one or more credentials, one or more instructions, or an authorization.

7. A computationally-implemented method comprising:

acquiring an indication that the particular party is speaking to a target device based on a position of a body part of the particular party;

receiving, from a particular device associated with the particular party, an indication of a location on at least one server of adaptation data correlated to the particular party;

receiving, at the target device, the adaptation data correlated to the particular party from the location on the at least one server indicated by the particular device associated with the particular party;

processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device; and

updating the adaptation data on the at least one server based at least in part on a result of the processed speech.

8. The computationally-implemented method of claim 7, wherein said acquiring an indication that the particular party is speaking to a target device based on a position of a body part of the particular party includes:

acquiring an indication that the particular party is speaking to the target device based on a position of a head of the particular party.

9. A computationally-implemented method comprising:

acquiring an indication that the particular party is speaking to the target device based on a position of an arm of the particular party;

receiving adaptation data correlated to the particular party from a location on at least one server specified by a particular device associated with the particular party;

processing speech from the particular party at least partly using the adaptation data received from the at least one server; and

updating the adaptation data on the at least one server based at least in part on a result of the processed speech.

21. A computationally-implemented method, comprising:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving adaptation data correlated to the particular party, said receiving facilitated by a particular device associated with the particular party;

processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device; and

updating the adaptation data based at least in part on a received result of the processed speech indicating a subjective opinion of the particular party as a numerical representation of a success of the speech interaction, such that the updated adaptation data is configured to be transmitted to the particular device.

22. The computationally-implemented method of claim 21, wherein said updating the adaptation data based at least in part on a received result of the processed speech indicating a subjective opinion of the particular party as a numerical representation of a success of the speech interaction includes:

updating the adaptation data based at least in part on a received result indicating the particular party'"'"'s numerical ranking of success of a speech interaction.

24. A computationally-implemented method comprising speech:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving adaptation data correlated to the particular party from a location on at least one server specified by a particular device associated with the particular party;

processing speech from the particular party at least partly using the adaptation data received from the at least one server; and

updating the adaptation data based at least in part on a received result from a social network website indicating the particular party'"'"'s subjective state after completing the speech interaction.

27. A computationally-implemented method comprising speech:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving adaptation data correlated to the particular party from a location on at least one server specified by a particular device associated with the particular party;

processing speech from the particular party at least partly using the adaptation data received from the at least one server; and

updating the adaptation data based at least in part on a received result from the particular device indicating the particular party'"'"'s subjective state inputted into the particular device by the particular party after completing the speech interaction.

28. A computationally-implemented method, comprising:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving, from a particular device associated with the particular party, an indication of a location on at least one server of adaptation data correlated to the particular party;

receiving, at the target device, the adaptation data correlated to the particular party, from the location on the at least one server indicated by the at least one other device associated with the particular party;

processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device;

requesting feedback from the particular party regarding a success of the speech interaction;

receiving feedback from the particular party regarding the success of the speech interaction; and

updating the adaptation data based at least in part on the feedback from the particular party.

29. The computationally-implemented method of claim 28, wherein said requesting feedback from the particular party regarding a success of the speech interaction includes:

transmitting a location at which feedback is requested from the particular party regarding a success of the speech interaction.

31. The computationally-implemented method of claim 28, wherein said requesting feedback from the particular party regarding a success of the speech interaction includes:

sending a message to the particular device requesting feedback from the particular party regarding a success of the speech interaction.

32. The computationally-implemented method of claim 28, wherein said requesting feedback from the particular party regarding a success of the speech interaction includes:

transmitting a message configured to be presented on the particular device requesting feedback from the particular party regarding a success of the speech interaction.

33. The computationally-implemented method of claim 28, wherein said requesting feedback from the particular party regarding a success of the speech interaction includes:

sending a request to the particular device requesting presentation of a message requesting feedback from the particular party regarding a success of the speech interaction.

34. The computationally-implemented method of claim 28, wherein said receiving feedback from the particular party regarding the success of the speech interaction includes:

receiving feedback via the particular device regarding the success of the speech interaction.

30. A computationally-implemented method, comprising:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving adaptation data correlated to the particular party, said receiving facilitated by a particular device associated with the particular party;

processing speech from the particular party at least partly using the received adaptation data correlated to the particular party;

transmitting a world wide web address at which feedback is requested from the particular party regarding a success of the speech interaction;

receiving feedback from the particular party regarding the success of the speech interaction; and

updating the adaptation data based at least in part on a result of the processed speech.

36. A computationally-implemented method comprising:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving, from a particular device associated with the particular party, an indication of a location on at least one server of adaptation data correlated to the particular party;

receiving, at the target device, the adaptation data correlated to the particular party from the location on the at least one server indicated by the particular device associated with particular party;

processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server device;

determining that the adaptation data should not be modified based at least in part on a result of the processed speech; and

transmitting, to the at least one server, an indication that a speech interaction has taken place and a determination that the adaptation data should not be modified has been made.

41. A computationally-implemented method comprising:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving adaptation data correlated to the particular party from a location on at least one server specified by a particular device associated with the particular party;

processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server; and

updating the adaptation data based at least in part on a result calculated by an inferred success of the speech interaction, inferred from a tone of voice used in the received speech from the particular party.

42. A computationally-implemented method comprising:

acquiring an indication of a speech interaction between a particular party and a target device;

receiving adaptation data correlated to the particular party from a location on at least one server specified by a particular device associated with the particular party;

processing speech from the particular party, received via the target device, at least partly using the adaptation data received from the at least one server; and

updating the adaptation data based at least in part on a result calculated by an inferred success of the speech interaction, inferred from a number of times a portion of the speech interaction was repeated.