Selecting from multiple content sources

• US 9,691,379 B1
• Filed: 06/26/2014
• Issued: 06/27/2017
• Est. Priority Date: 06/26/2014
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Mathias, Lambert, Potter, Stephen Frederick
• Primary Examiner(s): Augustin, Marcellus
• Application Number:
  US14/316,613
• Time in Patent Office: 1,097 Days
• Field of Search: 704263-2701, 704275
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning

Abstract

A speech-based system may be configured to receive and act upon spoken requests from a user. In some cases, a spoken request may ask the system to play content such as music without specifying from which of multiple available content sources the music is to be obtained. In response to such a request, the system analyzes feature scores for each of the content sources. The feature scores indicate usage characteristics of the different sources by a current user or groups of users. The features scores for a particular source may be averaged or otherwise combined to create a composite score, and the source having the highest composite score is selected as the source of the requested content.

Claims

1. A system, comprising:

one or more processors;

one or more computer-readable media storing computer-executable instructions that, when executed on the one or more processors, cause the one or more processors to perform acts comprising;

receiving an audio signal corresponding to a spoken request that expresses a play content intent, wherein the spoken request indicates a semantic role and an entity;

analyzing the audio signal to detect the semantic role and the entity;

identifying multiple content collections that have matching content for which the semantic role is filled by the entity;

identifying feature scores of each content collection, wherein an individual feature score of a particular content collection indicates an amount of usage of the content collection by one or more users;

calculating a composite score for each content collection based at least in part on the feature scores of the content collection;

selecting one of the multiple content collections by analyzing the calculated composite scores of the content collections;

receiving the matching content from the selected content collection; and

playing the matching content.

2. The system of claim 1, wherein:

the play content intent comprises a play music intent;

the semantic role comprises at least one of an artist, an album, or a track; and

the content collections comprise music collections.

3. The system of claim 1, wherein the content collections comprise at least one of video, music, audio books, or spoken word.

4. The system of claim 1, wherein calculating the composite scores comprises weighting the feature scores.

5. The system of claim 1, wherein the content collections comprise at least one of:

network-accessible content services;

local content repositories;

or

user-owned content libraries.

6. A method comprising:

receiving, by one or more processors, an audio signal corresponding to a natural language request that expresses an intent, wherein the request indicates an entity that fills a semantic role of the intent;

analyzing, by the one or more processors, the audio signal to detect the semantic role and the entity;

identifying, by the one or more processors, multiple content sources that have matching content for which the semantic role is filled by the entity;

identifying, by the one or more processors, at least one feature score for each of the content sources, wherein the at least one feature score characterizes usage of the content sources by one or more users;

selecting, by the one or more processors, one of the content sources based at least in part on the at least one feature score; and

receiving, by the one or more processors, the matching content from the selected content source.

7. The method of claim 6, wherein the at least one feature score of a particular content source characterizes a usage relationship of the particular content source to at least one of (a) the semantic role and entity or (b) one or more users.

8. The method of claim 6, wherein the at least one feature score characterizes at least one of:

usage of the content sources by a current user;

usage of the content sources by users other than the current user;

or

usage of the content sources by a category of users.

9. The method of claim 6, wherein the at least one feature score characterizes usage of the content sources for obtaining the matching content.

10. The method of claim 6, wherein the at least one feature score characterizes usage of the content sources for obtaining content that is similar to the matching content.

11. The method of claim 6, wherein the at least one feature score characterizes at least one of:

usage of the content sources by a current user for obtaining the matching content;

usage of the content sources by users other than the current user for obtaining the matching content;

or

usage of the content sources by a category of users for obtaining the matching content.

12. The method of claim 6, wherein the content sources comprise at least one of:

network-accessible services;

local content repositories;

or

user-owned content libraries.

13. A method comprising:

receiving, by one or more processors, an audio signal corresponding to a natural language request from a current user, wherein the natural language request indicates requested content;

identifying, by the one or more processors, multiple sources that have the requested content;

identifying, by the one or more processors, at least one feature score for each of the sources, wherein the feature scores characterize at least one of (a) usage of the sources by one or more users;

or (b) usage of the sources for providing the requested content; and

selecting, by the one or more processors, one of the sources based at least in part on the at least one feature score.

14. The method of claim 13, wherein the at least one feature score for a particular one of the sources comprises a level of confidence that the particular one of the sources has the requested content.

15. The method of claim 13, wherein the at least one feature score characterizes at least one of:

usage of the sources by a current user;

usage of the sources by users other than the current user;

or

usage of the sources by a category of users.

16. The method of claim 13, wherein the at least one feature score characterizes usage of the sources for obtaining the requested content.

17. The method of claim 13, wherein the at least one feature score characterizes usage of the sources for obtaining content that is similar to the requested content.

18. The method of claim 13, wherein the at least one feature score characterizes at least one of:

usage of the sources by a current user for obtaining the requested content;

usage of the sources by users other than the current user for obtaining the requested content;

or

usage of the sources by a category of users for obtaining the requested content.

19. The method of claim 13, wherein the sources comprise at least one of:

network-accessible services;

local content repositories;

or

user-owned content libraries.