Proactive command framework

• US 10,395,655 B1
• Filed: 09/13/2017
• Issued: 08/27/2019
• Est. Priority Date: 09/13/2017
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Kumar, Anjishnu, Fan, Xing, Gupta, Arpit, Sarikaya, Ruhi
• Primary Examiner(s): Chawan, Vijay B
• Application Number:
  US15/703,268
• Time in Patent Office: 713 Days
• Field of Search: 704275, 704 9, 704 10, 704251, 704270, 704235, 704249, 704277, 704257
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/14 using statistical models, e.g. Hidden Markov Models [HMMs] G10L15/18 takes precedence
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G10L 2015/226 using non-speech characteristics

Abstract

Techniques for determining a command or intent likely to be subsequently invoked by a user of a system are described. A user inputs a command (either via a spoken utterance or textual input) to a system. The system determines content responsive to the command. The system also determines a second command or corresponding intent likely to be invoked by the user subsequent to the previous command. Such determination may involve analyzing pairs of intents, with each pair being associated with a probability that one intent of the pair will be invoked by a user subsequent to a second intent of the pair. The system then outputs first content responsive to the first command and second content soliciting the user as to whether the system to execute the second command.

Claims

1. A computer-implemented method comprising:

receiving first data corresponding to a user command;

determining the user command corresponds to first intent data;

based at least in part on the first intent data, determining first content data responsive to the user command;

identifying first pre-stored intent pair data representing the first intent data and second intent data;

based at least in part on a system interaction history associated with a plurality of users, determining a first probability that a future user command will correspond to the second intent data;

identifying second pre-stored intent pair data representing the first intent data and third intent data;

based at least in part on the system interaction history, determining a second probability that a future user command will correspond to the third intent data;

based at least in part on the first probability and the second probability, generating second data requesting an instruction to invoke the second intent data;

sending, to a first device, the first content data for output; and

sending, to the first device, the second data for output.

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

determining a first system interaction history associated with a user corresponding to the user command;

based at least in part on the first probability and the first system interaction history, determining an adjusted first probability;

based at least in part on the second probability and the first system interaction history, determining an adjusted second probability; and

based at least in part on the adjusted first probability and the adjusted second probability, generating the second data.

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

determining a first system interaction history associated with a user corresponding to the user command;

determining the first system interaction history lacks an instance of a first previous user command that (i) corresponded to the third intent data and that (ii) was received after a second previous user command corresponding to the first intent data;

based at least in part on the second probability and the first system interaction history, determining an adjusted second probability;

determining the adjusted second probability fails to satisfy a threshold probability; and

based at least in part on determining the adjusted second probability fails to satisfy the threshold probability, generating the second data.

4. A system comprising:

at least one processor; and

at least one memory including instructions that, when executed by the at least one processor, cause the system to;

receive first data corresponding to a user command;

determine first intent data representing the input data user command;

based at least in part on the first intent data, determine a probability that a further user command, corresponding to second intent data, will be received;

determine second data responsive to the user command;

based at least in part on the probability, generate third data corresponding to the second intent data;

send, to a first device, the second data for output; and

send, to the first device, the third data for output.

5. The system of claim 4, wherein the first data is audio data, and wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to:

perform speech recognition processing on the audio data to generate text data; and

perform natural language processing on the text data to determine the first intent data.

6. The system of claim 4, wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to:

determine a system interaction history associated with a user corresponding to the user command, the system interaction history including at least one instance of a first previous user command that (i) corresponded to the second intent data and that (ii) was received after a second previous user command corresponding to the first intent data;

based at least in part on the probability and the system interaction history, determine an adjusted probability;

determine the adjusted probability satisfies a threshold probability; and

based at least in part on determining the adjusted probability satisfies the threshold probability, generate the third data.

7. The system of claim 4, wherein the third data:

requests a user indicate the user wants fourth data, responsive to the second intent data, to be output;

or

is responsive to the second intent data.

8. The system of claim 4, wherein the first intent data is associated with a first domain and the second intent data is associated with a second domain.

9. The system of claim 4, wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to:

determine a second probability that the further user command will correspond to third intent data;

determine the second probability fails to satisfy a threshold probability; and

based at least in part on determining the second probability fails to satisfy the threshold probability, generate the third data.

10. The system of claim 4, wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to:

determine a second probability that the further user command will correspond to third intent data;

determine the probability is greater than the second probability; and

based at least in part on determining the probability is greater than the second probability, generate the third data.

11. The system of claim 4, wherein the at least one memory further includes instructions that, when executed by the at least one processor, further cause the system to:

prior to determining the second intent data, determine a domain to execute the second intent data.

12. A computer-implemented method comprising:

receiving first data corresponding to a user command;

determining first intent data representing the user command;

based at least in part on the first intent data, determining a probability that a further user command, corresponding to second intent data, will be received;

determining second data responsive to the user command;

based at least in part on the probability, generating third data corresponding to the second intent data;

sending, to a first device, the second data for output; and

sending, to the first device, the third data for output.

13. The computer-implemented method of claim 12, wherein the first data is audio data, and wherein computer-implemented method further comprises:

performing speech recognition processing on the audio data to generate text data; and

performing natural language processing on the text data to determine the first intent data.

14. The computer-implemented method of claim 12, further comprising:

determining a system interaction history associated with a user corresponding to the user command, the system interaction history including at least one instance of a first previous user command that (i) corresponded to the second intent data and that (ii) was received after a second previous user command corresponding to the first intent data;

based at least in part on the probability and the system interaction history, determining an adjusted probability;

determining the adjusted probability satisfies a threshold probability; and

based at least in part on determining the adjusted probability satisfies the threshold probability, generating the third data.

15. The computer-implemented method of claim 14, wherein the first data is audio data, and wherein the computer-implemented method further comprises:

determining speech characteristics represented in the audio data;

determining a user profile associated with the speech characteristics; and

determining the system interaction history is associated with the user profile.

16. The computer-implemented method of claim 12, wherein the third data:

requests a user indicate the user wants fourth data, responsive to the second intent data, to be output;

or

is responsive to the second intent data.

17. The computer-implemented method of claim 12, wherein the first intent data is associated with a first domain and the second intent data is associated with a second domain.

18. The computer-implemented method of claim 12, further comprising:

determining a second probability that the further user command will correspond to third intend data;

determining the second probability fails to satisfy a threshold probability; and

based at least in part on determining the second probability fails to satisfy the threshold probability, generating the third data.

19. The computer-implemented method of claim 12, further comprising:

determining a second probability that the further user command will correspond to third intent data;

determining the probability is greater than the second probability; and

based at least in part on determining the probability is greater than the second probability, generate the third data.

20. The computer-implemented method of claim 12, further comprising:

prior to determining the second intent data, determining a domain to execute the second intent data.