Dynamic voice assistant interaction

• US 10,714,081 B1
• Filed: 03/07/2016
• Issued: 07/14/2020
• Est. Priority Date: 03/07/2016
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Miller, John Martin, Loritsch, Michael Lee, Tucker, Ross
• Primary Examiner(s): Serrou, Abdelali
• Application Number:
  US15/062,760
• Time in Patent Office: 1,590 Days
• CPC Class Codes: G06Q 30/0275 Auctions
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/08 Speech classification or search
• 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 2015/088 Word spotting
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

Systems, methods, and computer-readable media are disclosed for dynamic voice assistant interaction. Example methods may include receiving first voice data, determining a first meaning of the first voice data, conducting an auction for an audio segment to present in response to the first voice data, wherein the auction is based at least in part on the first meaning, and determining a first audio response for presentation via a speaker in response to the first voice data.

Claims

1. A method comprising:

receiving, by one or more computer processors coupled to at least one memory, first voice data comprising a user utterance;

converting the first voice data to first text data, wherein the first text data comprises a textual representation of the user utterance;

determining a first keyword in the first text data;

identifying a user request using the first keyword;

determining a first follow-up inquiry associated with the user request;

presenting a first audible response to the user request, wherein the first audible response comprises the first follow-up inquiry;

receiving second voice data in response to the first audible response;

converting the second voice data to second text data;

determining that the second text data does not include a keyword;

sending a first bid request to an ad server, wherein the first bid request comprises a request for bids for audio content to be presented in response to the second voice data;

receiving a bid from the ad server, wherein the bid comprises a bid amount, text of candidate audio content, and a content category identifier;

determining that the candidate audio content meets a relevance threshold using the content category identifier;

determining that the bid is a winning bid;

converting the text of the candidate audio content to audio data;

presenting a second audible response comprising the audio data, wherein the audio data represents a second follow-up inquiry;

receiving third voice data comprising an answer to the second follow-up inquiry;

determining a user interest level using the answer, wherein the user interest level increases responsive to an affirmative answer and decreases responsive to a negative answer; and

sending a second bid request based at least in part on the user interest level; and

presenting a third audible response indicating that an audio advertisement is forthcoming.

2. The method of claim 1, further comprising:

presenting a fourth audible response, wherein the third audible response comprises an offer to purchase an item; and

initiating purchase of the item.

3. The method of claim 1, wherein the first voice data is received during a user interaction, and wherein the bid request comprises a predetermined pricing schedule for subsequent audio content to be presented during the user interaction.

4. A method comprising:

receiving, by one or more computer processors coupled to at least one memory, first voice data;

determining a first meaning of the first voice data;

determining a set of candidate advertisers based at least in part on the first meaning, the set of candidate advertisers comprising a first advertiser;

determining a predetermined follow-up inquiry to the first voice data based at least in part on the first meaning and a category associated with the first advertiser;

causing audible presentation of a first audio response to the first voice data, wherein the first audio response comprises the predetermined follow-up inquiry;

conducting an auction for an audio segment for an available audio segment slot, wherein the set of candidate advertisers is eligible to participate in the auction;

determining that the first advertiser is a winner of the auction;

determining a winning segment from the first advertiser, the winning segment comprising text for presentation as audio;

causing audible presentation of an indication that an audio advertisement is forthcoming; and

causing audible presentation of the text.

5. The method of claim 4, further comprising presenting the first audio response via one or more speaker devices.

6. The method of claim 4, further comprising:

determining a keyword associated with the first meaning;

wherein determining the predetermined follow-up inquiry to the first voice data based at least in part on the first meaning and the category associated with the first advertiser comprises determining the predetermined follow-up inquiry to the first voice data based at least in part on the keyword and the category associated with the first advertiser.

7. The method of claim 4, further comprising determining that there is no secondary predetermined follow-up inquiry associated with the first meaning.

8. The method of claim 4, wherein the audio segment is a first audio segment, the method further comprising:

receiving second voice data;

sending a bid request to the first advertiser for a second audio segment in response to the second voice data, wherein the bid request comprises a predetermined price floor; and

sending a bid request to the set of candidate advertisers after sending the bid request to the first advertiser.

9. The method of claim 8, wherein the segment auction is a first segment auction, the method further comprising:

receiving a bid amount from the first advertiser;

determining that the bid amount does not meet the predetermined price floor; and

conducting a second segment auction for the second audio segment.

10. The method of claim 4, further comprising:

receiving a bid from the first advertiser, the bid comprising a bid amount, and text of a candidate audio segment; and

determining that the candidate audio segment is eligible.

11. The method of claim 4, further comprising:

determining that the first advertiser is associated with the first meaning;

receiving a bid from a second advertiser, the bid comprising a bid amount, and text of a candidate audio segment;

determining that the second advertiser is a blacklisted advertiser; and

canceling the bid.

12. The method of claim 4, wherein determining the first meaning of the first voice data comprises:

converting the first voice data to first text data, wherein the first text data comprises a textual representation of a user utterance;

determining a first keyword associated with the first text data; and

identifying a user request using the first keyword.

17. The method of claim 4, further comprising determining that the category associated with the first advertiser corresponds to a category associated with the first meaning.

13. A device comprising:

a speaker device;

at least one memory that stores computer-executable instructions; and

at least one processor configured to access the at least one memory and execute the computer-executable instructions to;

receive first voice data;

determine a first meaning of the first voice data;

determine a set of candidate advertisers based at least in part on the first meaning, the set of candidate advertisers comprising a first advertiser;

determine a predetermined follow-up inquiry to the first voice data based at least in part on the first meaning and a category associated with the first advertiser;

cause audible presentation of a first audio response to the first voice data, wherein the first audio response comprises the predetermined follow-up inquiry;

conduct an auction for an audio segment for an available audio segment slot, wherein the set of candidate advertisers is eligible to participate in the auction;

determine a winning segment from an advertiser, the winning segment comprising text for presentation as audio;

determine that the first advertiser is a winner of the auction;

cause audible presentation of an indication that an audio advertisement is forthcoming; and

cause audible presentation of the text.

14. The device of claim 13, wherein the at least one processor is further configured to access the at least one memory and execute the computer-executable instructions to present the first audio response via the speaker device.

15. The device of claim 13, wherein the at least one processor is further configured to access the at least one memory and execute the computer-executable instructions to:

determine a keyword associated with the first meaning;

wherein the at least one processor is configured to determine the predetermined follow-up inquiry to the first voice data based at least in part on the first meaning and the category associated with the first advertiser by determining the predetermined follow-up inquiry to the first voice data based at least in part on the keyword and the category associated with the first advertiser.

16. The device of claim 13, wherein the audio segment is a first audio segment, and wherein the at least one processor is further configured to access the at least one memory and execute the computer-executable instructions to:

receive second voice data; and

send a bid request to the first advertiser for a second audio segment in response to the second voice data, wherein the bid request comprises a predetermined price floor.

18. The device of claim 13, wherein the at least one processor is further configured to access the at least one memory and execute the computer-executable instructions to:

determine that the category associated with the first advertiser corresponds to a category associated with the first meaning.