Biasing voice correction suggestions

• US 10,049,655 B1
• Filed: 01/05/2016
• Issued: 08/14/2018
• Est. Priority Date: 01/05/2016
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Cherepanov, Evgeny A., Foerster, Jakob Nicolaus, Sridar, Vikram, Rabinovitz, Ishai, Tabach, Omer
• Primary Examiner(s): Patel, Shreyans
• Application Number:
  US14/988,074
• Time in Patent Office: 952 Days
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/221 Announcement of recognition results

Abstract

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for natural language processing. One of the method includes receiving a voice input from a user device; generating a recognition output; receiving a user selection of one or more terms in the recognition output; receiving a user input of one or more letters replacing the user selected one or more terms; determining suggested correction candidates based in part on the user input and the voice input; and providing one or more suggested correction candidates to the user device as suggested corrected recognition outputs.

Claims

1. A method comprising:

receiving a voice input from a user device;

generating a recognition output;

receiving a user selection of one or more terms in the recognition output;

receiving a user input of one or more letters replacing the user selected one or more terms;

determining suggested correction candidates based in part on the user input and the voice input, wherein determining suggested correction candidates includes;

obtaining candidate suggestions for a corrected recognition output based on a collection of queries, wherein the candidate suggestions are ranked according to a respective score for each candidate suggestion,

re-ranking the candidate suggestions based on phonetic similarity between each candidate suggestion and the received voice input, and

determining one or more suggested correction candidates based on the re-ranking; and

providing one or more suggested correction candidates to the user device as suggested corrected recognition outputs.

2. The method of claim 1, further comprising:

receiving a user selection of one of the suggested corrected recognition outputs; and

providing one or more search results responsive to the selected suggested query.

3. The method of claim 1, wherein obtaining candidate suggestions comprises determining queries that match a portion of the recognition output with the user input to a collection of queries and ranking the queries based on a respective popularity.

4. The method of claim 1, wherein determining phonetic similarity includes calculating a phonetic distance between each candidate suggestion and the received voice input.

5. The method of claim 1, wherein re-raking the candidate suggestions includes modifying the respective scores of the candidate suggestions based on whether the corresponding phonetic similarity satisfies a threshold.

6. The method of claim 1, further comprising:

receiving additional user input letters for the one or more terms; and

in response to the additional user input, updating the suggested corrected recognition outputs.

7. A system comprising:

one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising;

receiving a voice input from a user device;

generating a recognition output;

receiving a user selection of one or more terms in the recognition output;

receiving a user input of one or more letters replacing the user selected one or more terms;

determining suggested correction candidates based in part on the user input and the voice input, wherein determining suggested correction candidates comprises;

obtaining candidate suggestions for a corrected recognition output based on a collection of queries, wherein the candidate suggestions are ranked according to a respective score for each candidate suggestion;

re-ranking the candidate suggestions based on phonetic similarity between each candidate suggestion and the received voice input; and

determining one or more suggested correction candidates based on the re-ranking; and

providing one or more suggested correction candidates to the user device as suggested corrected recognition outputs.

8. The system of claim 7, further operable to perform operations comprising:

receiving a user selection of one of the suggested corrected recognition outputs; and

providing one or more search results responsive to the selected suggested query.

9. The system of claim 7, wherein obtaining candidate suggestions comprises determining queries that match a portion of the recognition output with the user input to a collection of queries and ranking the queries based on a respective popularity.

10. The system of claim 7, wherein determining phonetic similarity includes calculating a phonetic distance between each candidate suggestion and the received voice input.

11. The system of claim 7, wherein re-raking the candidate suggestions includes modifying the respective scores of the candidate suggestions based on whether the corresponding phonetic similarity satisfies a threshold.

12. The system of claim 7, further comprising:

receiving additional user input letters for the one or more terms; and

in response to the additional user input, updating the suggested corrected recognition outputs.

13. One or more non-transitory computer-readable storage media encoded with instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising:

receiving a voice input from a user device;

generating a recognition output;

receiving a user selection of one or more terms in the recognition output;

receiving a user input of one or more letters replacing the user selected one or more terms;

determining suggested correction candidates based in part on the user input and the voice input, wherein determining suggested correction candidates comprises;

obtaining candidate suggestions for a corrected recognition output based on a collection of queries, wherein the candidate suggestions are ranked according to a respective score for each candidate suggestion;

re-ranking the candidate suggestions based on phonetic similarity between each candidate suggestion and the received voice input; and

determining one or more suggested correction candidates based on the re-ranking; and

providing one or more suggested correction candidates to the user device as suggested corrected recognition outputs.

14. The one or more non-transitory computer-readable storage media of claim 13, further configured to perform operations comprising:

receiving a user selection of one of the suggested corrected recognition outputs; and

providing one or more search results responsive to the selected suggested query.

15. The one or more non-transitory computer-readable storage media of claim 13, wherein obtaining candidate suggestions comprises determining queries that match a portion of the recognition output with the user input to a collection of queries and ranking the queries based on a respective popularity.

16. The one or more non-transitory computer-readable storage media of claim 13, wherein determining phonetic similarity includes calculating a phonetic distance between each candidate suggestion and the received voice input.

17. The one or more non-transitory computer-readable storage media of claim 13, wherein re-raking the candidate suggestions includes modifying the respective scores of the candidate suggestions based on whether the corresponding phonetic similarity satisfies a threshold.