Application keywords

• US 9,740,751 B1
• Filed: 02/18/2016
• Issued: 08/22/2017
• Est. Priority Date: 02/18/2016
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Yeom, Jaehyun, Lee, Dong Ha
• Primary Examiner(s): Baker, Matthew
• Application Number:
  US15/046,831
• Time in Patent Office: 551 Days
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving audio data corresponding to an utterance of a user, determining that at least a portion of the audio data corresponds to an action keyword, obtaining candidate applications that correspond to the action keyword, which include applications associated with a uniform resource identifier (URI) that corresponds to the action keyword, ranking the candidate applications that correspond to the action keyword based at least on: (i) a bidding weight associated with the action keyword for each of the candidate applications, and (ii) a power score associated with each of the candidate applications, identifying a highest ranked application among the candidate applications that has a URI that corresponds to at least the action keyword, and executing the URI for the highest ranked application among the candidate applications that correspond to at least the action keyword.

Claims

1. A computer-implemented method comprising:

receiving audio data corresponding to an utterance of a user;

determining that at least a portion of the audio data corresponds to an action keyword, the action keyword corresponding to an action to be performed by a device associated with the user;

identifying candidate applications that correspond to the action keyword, the candidate applications that correspond to the action keyword comprising applications associated with a uniform resource identifier (URI) that corresponds to the action keyword;

selecting a candidate application, from the identified candidate applications that correspond to the action keyword, based at least on;

(i) a bidding weight associated with the action keyword for each of the candidate applications, and (ii) a power score associated with each of the candidate applications that is determined based at least on a popularity metric and a rating associated with each of the candidate applications,

wherein the bidding weight associated with the action keyword for each of the candidate applications is normalized for each of the candidate applications to generate a normalized bidding weight associated with the action keyword for each of the candidate applications; and

executing the URI for the selected candidate application.

2. The computer-implemented method of claim 1, further comprising:

determining that at least a portion of the audio data corresponds to an object keyword, the object keyword corresponding to an object of the action to be performed by the device associated with the user; and

identifying candidate applications that correspond to a combination of the action keyword and the object keyword, the candidate applications that correspond to the combination of the action keyword and the object keyword comprising applications associated with a URI that corresponds to the combination of the action keyword and the object keyword,

wherein selecting a candidate application comprises selecting the candidate application, from the identified candidate applications that correspond to a combination of the action keyword and the object keyword, based at least on;

(i) a bidding weight associated with the combination of the action keyword and the object keyword for each of the candidate applications, and (ii) a power score associated with each of the candidate applications, and

wherein executing the URI for the selected candidate application comprises executing the URI for the selected candidate application among the identified candidate applications that correspond to the combination of the action keyword and the object keyword.

3. The computer-implemented method of claim 1, further comprising:

determining at least one application, among the candidate applications, that is not installed on the device associated with the user; and

removing the at least one application that is determined to not be installed on the device associated with the user from the identified candidate applications.

4. The computer-implemented method of claim 1, wherein the bidding weight associated with the action keyword for each of the candidate applications is obtained from metadata associated with the application.

5. The computer-implemented method of claim 1, wherein the popularity metric is based on at least one of:

a number of downloads, a frequency of use, a user rating, a number of user reviews, a number of search queries, and a number of downloads in a given time period.

6. A computer-implemented method comprising:

receiving audio data corresponding to an utterance of a user;

determining that at least a portion of the audio data corresponds to an action keyword, the action keyword corresponding to an action to be performed by a device associated with the user;

identifying candidate applications that correspond to the action keyword, the candidate applications that correspond to the action keyword comprising applications associated with a uniform resource identifier (URI) that corresponds to the action keyword;

selecting a candidate application, from the identified candidate applications that correspond to the action keyword, based at least on;

(i) a bidding weight associated with the action keyword for each of the candidate applications, and (ii) a power score associated with each of the candidate applications that is determined based at least on a popularity metric and a rating associated with each of the candidate applications,

wherein selecting the candidate application, from the identified candidate applications that correspond to the action keyword, comprises;

determining a ranking score for each of the candidate applications, the ranking score being determined based at least on the bidding weight associated with the action keyword for each of the candidate applications, the power score associated with each of the candidate applications, and a feedback score for each URI that corresponds to the action keyword, the feedback score being determined for each URI based at least on a user action associated with the URI;

identifying a highest ranked application among the candidate applications that has a URI that corresponds to at least the action keyword; and

executing the URI for the highest ranked application.

14. The computer-implemented method of claim 6, further comprising:

determining at least one application, among the candidate applications, that is not installed on the device associated with the user; and

removing the at least one application that is determined to not be installed on the device associated with the user from the identified candidate applications.

15. The computer-implemented method of claim 6, wherein the bidding weight associated with the action keyword for each of the candidate applications is obtained from metadata associated with the application.

16. The computer-implemented method of claim 6, wherein the popularity metric is based on at least one of:

a number of downloads, a frequency of use, a user rating, a number of user reviews, a number of search queries, and a number of downloads in a given time period.

17. The computer-implemented method of claim 6, wherein the bidding weight associated with the action keyword for each of the candidate applications is normalized for each of the candidate applications to generate a normalized bidding weight associated with the action keyword for each of the candidate applications.

7. A system comprising one or more processors and one or more storage devices storing instructions that are operable, when executed by the one or more processors, to cause the one or more processors to perform operations comprising:

receiving audio data corresponding to an utterance of a user;

determining that at least a portion of the audio data corresponds to an action keyword, the action keyword corresponding to an action to be performed by a device associated with the user;

identifying candidate applications that correspond to the action keyword, the candidate applications that correspond to the action keyword comprising applications associated with a uniform resource identifier (URI) that corresponds to the action keyword;

selecting a candidate application, from the identified candidate applications that correspond to the action keyword, based at least on;

(i) a bidding weight associated with the action keyword for each of the candidate applications, and (ii) a power score associated with each of the candidate applications that is determined based at least on a popularity metric and a rating associated with each of the candidate applications,

wherein the bidding weight associated with the action keyword for each of the candidate applications is normalized for each of the candidate applications to generate a normalized bidding weight associated with the action keyword for each of the candidate applications; and

executing the URI for the selected candidate application.

8. The system of claim 7, wherein the operations comprise:

determining that at least a portion of the audio data corresponds to an object keyword, the object keyword corresponding to an object of the action to be performed by the device associated with the user; and

identifying candidate applications that correspond to a combination of the action keyword and the object keyword, the candidate applications that correspond to the combination of the action keyword and the object keyword comprising applications associated with a URI that corresponds to the combination of the action keyword and the object keyword,

wherein selecting a candidate application comprises selecting the candidate application, from the identified candidate applications that correspond to a combination of the action keyword and the object keyword, based at least on;

(i) a bidding weight associated with the combination of the action keyword and the object keyword for each of the candidate applications, and (ii) a power score associated with each of the candidate applications, and

wherein executing the URI for the selected candidate application comprises executing the URI for the selected candidate application among the identified candidate applications that correspond to the combination of the action keyword and the object keyword.

9. The system of claim 7, wherein the operations comprise:

determining at least one application, among the candidate applications, that is not installed on the device associated with the user; and

removing the at least one application that is determined to not be installed on the device associated with the user from the identified candidate applications.

10. The system of claim 7, wherein the bidding weight associated with the action keyword for each of the candidate applications is obtained from metadata associated with the application.

11. The system of claim 7, wherein the popularity metric is based on at least one of:

a number of downloads, a frequency of use, a user rating, a number of user reviews, a number of search queries, and a number of downloads in a given time period.

12. The system of claim 7, wherein selecting the candidate application, from the identified candidate applications that correspond to the action keyword, is further based at least on a feedback score for each URI that corresponds to the action keyword, the feedback score being determined for each URI based at least on a user action associated with the URI.

13. The system of claim 12, wherein selecting the candidate application, from the identified candidate applications that correspond to the action keyword, comprises:

determining a ranking score for each of the candidate applications, the ranking score being determined as a product of the bidding weight associated with the action keyword for each of the candidate applications, the power score associated with each of the candidate applications, and the feedback score for each URI that corresponds to the action keyword; and

identifying a highest ranked application among the candidate applications that has a URI that corresponds to at least the action keyword.