Granted Patent B2

US 11,836,453 · App. 17/383,097 · Granted Dec 5, 2023

Inventors: Kamyar Mohajer (San Jose, CA); Keyvan Mohajer (Los Gatos, CA); Bernard Mont-Reynaud (Sunnyvale, CA); Pranav Singh (Sunnyvale, CA)

Assignee: SoundHound, Inc.

G06F40/40G06Q20/102G06Q30/0283G10L15/30G06F40/211

Abstract

Claims (50)

1. A method comprising:

receiving a natural language query as a text string including one or more words;

accessing a collection of natural language processing modules, natural language processing modules being configured to process distinct sets of natural language queries;

identifying a plurality of competing natural language processing modules, from the collection of natural language processing modules, to interpret the text string;

computing a plurality of interpretations of the text string using the identified natural language processing modules;

choosing the highest-ranking interpretation of the plurality of interpretations; and

returning a response to the natural language query using the highest-ranking interpretation.

2. The method of claim 1 , wherein returning a response to the natural language query comprises generating a composite response including:

performing a command corresponding to the highest-ranking interpretation; and

outputting a command response for a user.

3. The method of claim 2 , wherein outputting a command response comprises outputting a spoken response.

4. The method of claim 1 , wherein identifying a plurality of natural language processing modules comprises identifying a natural language processing module in a domain associated with the one or more words.

5. The method of claim 1 , wherein receiving a natural language query comprises receiving a natural language query in an automobile.

6. The method of claim 1 , wherein identifying a plurality of natural language processing modules comprises comparing the received natural language query to the sets of natural language queries corresponding to the natural language processing modules.

7. The method of claim 1 , further comprising accessing context information associated with the received natural language query; and

wherein identifying a plurality of natural language processing modules comprises identifying the natural language processing module based on the context information.

8. The method of claim 1 , wherein computing a plurality of interpretations of the text string comprises computing interpretations of the one or more words using the identified natural language processing modules.

9. A computer system comprising:

a processor; and

system memory coupled to the processor and storing instructions configured to cause the processor to:

receive a natural language query as a text string including one or more words;

access a collection of natural language processing modules, natural language processing modules being configured to process distinct sets of natural language queries;

identify a plurality of competing natural language processing modules, from the collection of natural language processing modules, to interpret the text string;

compute a plurality of interpretations of the text string using the identified natural language processing modules;

choose the highest-ranking interpretation of the plurality of interpretations; and

return a response to the natural language query using the highest-ranking interpretation.

10. The computer system of claim 9 , wherein instructions configured to return a response to the natural language query comprise instructions configured to generate a composite response including:

perform a command corresponding to the highest-ranking interpretation; and

output a command response for a user.

11. The computer system of claim 9 , wherein instructions configured to receive a natural language query comprises instructions configured to receive a natural language query at an automobile.

12. The computer system of claim 9 , wherein instructions configured to identify a plurality of natural language processing modules comprise instructions configured to compare the received natural language query to the sets of natural language queries corresponding to the natural language processing modules.

13. The computer system of claim 9 , further comprising instructions configured to access context information associated with the received natural language query; and

wherein instructions configured to identify a plurality of natural language processing modules comprise instructions configured to identify the natural language processing module based on the context information.

14. A method comprising:

receiving a natural language query as a text string including one or more words from a vehicle occupant;

accessing a collection of natural language processing modules at the vehicle, natural language processing modules being configured to process distinct sets of natural language queries;

identifying a plurality of natural language processing modules, from the collection of natural language processing modules, to interpret the text string;

computing a plurality of interpretations of the text string using the identified natural language processing modules;

choosing the highest-ranking interpretation of the plurality of interpretations; and

returning a response to the natural language query within the vehicle using the highest-ranking interpretation.

15. The method of claim 14 , wherein receiving a natural language query comprises receiving a query from the vehicle occupant requesting a 3 rd party service;

wherein computing a plurality of interpretations of the text string comprises interpreting the natural language query with a car manufacturer specific module; and

wherein using the highest-ranking interpretation comprises requesting the 3 rd party service.

16. The method of claim 14 , wherein using the highest-ranking interpretation to return a response to the natural language query comprises generating a composite response including:

performing a command corresponding to the highest-ranking interpretation; and

outputting a command response within the vehicle.

17. The method of claim 14 , wherein identifying a plurality of natural language processing modules comprises comparing the received natural language query to the sets of natural language queries corresponding to the natural language processing modules.

18. The method of claim 14 , further comprising accessing context information associated with the received natural language query; and

wherein identifying a plurality of natural language processing modules comprises identifying the natural language processing module based on the context information.

19. The method of claim 14 , wherein computing a plurality of interpretations of the text string comprises computing interpretations of the one or more words.

Assignments (7)