AUTOMATIC LANGUAGE IDENTIFICATION FOR SPEECH

• US 20180211650A1
• Filed: 01/24/2017
• Published: 07/26/2018
• Est. Priority Date: 01/24/2017
• Status: Active Grant

• Current Assignee: Lenovo PC International Limited (Lenovo Group Ltd.)
• Original Assignee: Lenovo Singapore Pte Limited (Lenovo Group Ltd.)
• Inventors: Knudson, Ryan Charles, Waltermann, Rod D.
• Granted Patent: US 10,741,174 B2
• Time in Patent Office: Days
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 2015/226 using non-speech characteristics

Abstract

One embodiment provides a method, including: receiving, at an information handling device, voice input from at least one user; accessing at least one language model; determining, based upon the accessed at least one language model, at least one language of the voice input; and providing output associated with the voice input, based upon the determined language. Other aspects are described and claimed.

Claims

1. A method, comprising:

determining, using GPS data, a user'"'"'s geographic location;

accessing, based upon the user'"'"'s geographic location, at least one language model associated with the user'"'"'s geographic location;

receiving, at an information handling device, voice input from at least one user;

determining, based upon the accessed at least one language model, at least one language of the voice input; and

providing output associated with the voice input, based upon the determined language.

2. The method of claim 1, wherein the determining comprises comparing the voice input to the at least one language model.

3. The method of claim 2, further comprising generating a confidence score related to a match between the voice input and the at least one language model.

4. The method of claim 2, wherein each of the at least one language models are associated with a different language.

5. The method of claim 1, wherein the accessing comprises accessing a default language model.

6. The method of claim 5, responsive to determining the voice input is not associated with the default language model, further comprising prompting the at least one user to identify the at least one language associated with the voice input.

7. The method of claim 1, wherein the voice input comprises a wakeup indication.

8. The method of claim 7, wherein the accessing comprises accessing at least one language model associated with the wakeup indication.

9. The method of claim 1, wherein the output comprises a textual transcription of the voice input in a script associated with the determined at least one language.

10. The method of claim 1, wherein the output comprises a vocal recitation of generated results responsive to the voice input.

11. An information handling device, comprising:

a processor;

a memory device that stores instructions executable by the processor to;

determine, using GPS data, a user'"'"'s geographic location;

access, based upon the user'"'"'s geographic location, at least one language model associated with the user'"'"'s geographic location;

receive voice input from at least one user;

determine, based upon the accessed at least one language model, at least one language of the voice input; and

provide output associated with the voice input, based upon the determined language.

12. The information handling device of claim 11, wherein the instructions executable by the processor to determine comprises instructions executable by the processor to compare the voice input to the at least one language model.

13. The information handling device of claim 12, wherein instructions executable by the processor further comprise instructions to generate a confidence score related to a match between the voice input and the at least one language model.

14. The information handling device of claim 12, wherein each of the at least one language models are associated with a different language.

15. The information handling device of claim 11, wherein the instructions executable by the processor to access comprises instructions executable by the processor to access a default language model.

16. The information handling device of claim 11, wherein the voice input comprises a wakeup indication.

17. The information handling device of claim 16, wherein the instructions executable by the processor to access comprises accessing at least one language model associated with the wakeup indication.

18. The information handling device of claim 11, wherein the output comprises a textual transcription of the voice input in a script associated with the determined at least one language.

19. The information handling device of claim 11, wherein the output comprises a vocal recitation of generated results responsive to the voice input.

20. A product, comprising:

a storage device that stores code, the code being executable by a processor and comprising;

code that determines, using GPS data, a user'"'"'s geographic location;

code that access, based upon the user'"'"'s geographic location, at least one language model associated with the user'"'"'s geographic location;

code that receives voice input from at least one user;

code that determines, based upon the accessed at least one language model, at least one language of the voice input; and

code that provides output associated with the voice input, based upon the determined language.