Retrieval and management of spoken language understanding personalization data

• US 9,361,289 B1
• Filed: 08/30/2013
• Issued: 06/07/2016
• Est. Priority Date: 08/30/2013
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Jampani, Madan Mohan Rao, Rajasekaram, Arushan, Strom, Nikko, Watanabe, Yuzo, Salvador, Stan Weidner
• Primary Examiner(s): Spooner, Lamont
• Application Number:
  US14/015,697
• Time in Patent Office: 1,012 Days
• Field of Search: 704/1, 704/9, 704/10, 704/251, 704/257, 707706-708
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/18 using natural language modelling

Abstract

Features are disclosed for maintaining data that can be used to personalize spoken language processing, such as automatic speech recognition (“ASR”), natural language understanding (“NLU”), natural language processing (“NLP”), etc. The data may be obtained from various data sources, such as applications or services used by the user. User-specific data maintained by the data sources can be retrieved and stored for use in generating personal models. Updates to data at the data sources may be reflected by separate data sets in the personalization data, such that other processes can obtain the update data sets separate from other data.

Claims

1. A computer system comprising:

a data store storing spoken language understanding personalization data for a plurality of users, wherein spoken language understanding personalization data for a first user of the plurality of users comprises;

one or more first data sets comprising first personalization data received from a first application separate from the computer system, wherein the first personalization data corresponds to at least a portion of first application data associated with the first user; and

one or more second data sets comprising second personalization data received from a second application separate from the computer system, wherein the second personalization data corresponds to at least a portion of second application data associated with the first user; and

one or more processors in communication with the data store, the one or more processors programmed to at least;

obtain a first notification regarding an update to the first application data associated with the first user, the first notification comprising first priority data;

store, in the data store, a first update data set based at least on the first notification, wherein the first update data set corresponds to at least a portion of the update to the first application data associated with the first user;

determine, based at least on the first priority data indicating a low priority for the first update data, not to update a spoken language understanding model, associated with the user, using the first update data;

obtain a second notification regarding an update to the second application data associated with the first user, the second notification comprising second priority data;

store, in the data store, a second update data set based at least on the second notification, wherein the second update data set corresponds to at least a portion of the update to the second application data associated with the first user;

determine, based at least on the second priority data indicating a high priority for the second update data, to update the spoken language understanding model using the second update data; and

update the spoken language understanding model associated with the first user using the second update data.

2. The computer system of claim 1, wherein the first update data set is stored such that it is accessible separately from the one or more first data sets.

3. The system of claim 1, wherein the one or more processors are further programmed to:

determine, based on the first priority data indicating a low priority for the first update data, to aggregate the first update data with third update data, wherein the third update data is associated with third priority data indicating a low priority for the third update data; and

aggregate the first update data with the third update data.

4. A computer-implemented method comprising:

as executed by one or more computing devices configured to execute specific instructions,

storing first personalization data received from an application separate from the one or more computing devices, wherein the first personalization data corresponds to at least a portion of application data associated with a user of the application;

determining, based on the first personalization data being associated with a low priority, not to update a spoken language understanding model associated with the user;

obtaining a notification regarding a change to the application data associated with the user;

storing second personalization data based at least on the notification, wherein the second personalization data corresponds to at least a portion of the change to the application data associated with the user;

determining, based on the second personalization data being associated with a high priority, to update the spoken language understanding model associated with the user; and

updating the spoken language understanding model using the second personalization data.

5. The computer-implemented method of claim 4, further comprising subscribing to notifications from the application.

6. The computer-implemented method of claim 4, further comprising subscribing to notifications from a plurality of applications that maintain application data associated with the user.

7. The computer-implemented method of claim 4, further comprising:

storing third personalization data received from a second application separate from the application, wherein the third personalization data corresponds to at least a portion of second application data associated with the user;

obtaining a second notification regarding a second a change to the second application data associated with the user; and

storing fourth personalization data based at least on the second notification, wherein fourth personalization data corresponds to at least a portion of the change to the second application data associated with the user.

8. The computer-implemented method of claim 4, wherein the first personalization data corresponds to a music library of the user, recent purchases by the user, or an address book of the user.

9. The computer-implemented method of claim 4, wherein the second personalization data is stored in a separate data object than the first personalization data.

10. The computer-implemented method of claim 9, wherein the separate data object comprises a separate electronic file, a separate data page, or a separate data partition.

11. The computer-implemented method of claim 4, wherein the notification comprises the second personalization data.

12. The computer-implemented method of claim 4, further comprising storing priority data in association with the second personalization data.

13. The computer-implemented method of claim 4, wherein the notification comprises priority data.

14. The computer-implemented method of claim 4, wherein the change to the application data comprises one of an addition, a deletion, or a modification.

15. The computer-implemented method of claim 4, wherein the second personalization data is stored such that it is accessible separately from the first personalization data.

16. The computer-implemented method of claim 4, further comprising:

determining, based on the first personalization data being associated with a low priority, to aggregate the first personalization data with third personalization data, wherein the third personalization data is also associated with a low priority; and

aggregating the first personalization data with the third personalization data.

17. One or more non-transitory computer readable media comprising executable code that, when executed, cause one or more computing devices to perform a process comprising:

storing first personalization data received from an application separate from the one or more computing devices, wherein the first personalization data corresponds to at least a portion of application data associated with a user of the application;

determining, based on the first personalization data being associated with a low priority, to aggregate the first personalization data with second personalization data, wherein the second personalization data is also associated with a low priority;

obtaining a notification regarding a change to the application data associated with the user;

storing third personalization data based at least on the notification, wherein the third personalization data corresponds to at least a portion of the change to the application data associated with the user;

determining, based on the third personalization data being associated with a high priority, to update the spoken language understanding model associated with the user; and

updating the spoken language understanding model using the third personalization data.

18. The one or more non-transitory computer readable media of claim 17, wherein the process further comprises subscribing to notifications from the application.

19. The one or more non-transitory computer readable media of claim 17, wherein the process further comprises subscribing to notifications from a plurality of applications that maintain application data associated with the user.

20. The one or more non-transitory computer readable media of claim 17, wherein the process further comprises:

storing fourth personalization data received from a second application separate from the application, wherein the fourth personalization data corresponds to at least a portion of second application data associated with the user;

obtaining a second notification regarding a second a change to the second application data associated with the user; and

storing fifth personalization data based at least on the second notification, wherein fifth personalization data corresponds to at least a portion of the change to the second application data associated with the user.

21. The one or more non-transitory computer readable media of claim 17, wherein the first personalization data corresponds to a music library of the user, recent purchases by the user, or an address book of the user.

22. The one or more non-transitory computer readable media of claim 17, wherein the second personalization data is stored in a separate data object than the first personalization data.

23. The one or more non-transitory computer readable media of claim 22, wherein the separate data object comprises a separate electronic file, a separate data page, or a separate data partition.

24. The one or more non-transitory computer readable media of claim 17, wherein the notification comprises the second personalization data.

25. The one or more non-transitory computer readable media of claim 17, wherein the process further comprises storing priority data in association with the second personalization data.

26. The one or more non-transitory computer readable media of claim 17, wherein the notification comprises priority data.

27. The one or more non-transitory computer readable media of claim 17, wherein the change to the application data comprises one of an addition, a deletion, or a modification.

28. The one or more non-transitory computer readable media of claim 17, wherein the second personalization data is stored such that it is accessible separately from the first personalization data.