Language model biasing modulation

• US 10,714,075 B2
• Filed: 04/11/2019
• Issued: 07/14/2020
• Est. Priority Date: 03/30/2015
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Moreno Mengibar, Pedro J., Aleksic, Petar
• Primary Examiner(s): Singh, Satwant K
• Application Number:
  US16/381,167
• Publication Number:
  US 20190237063A1
• Time in Patent Office: 460 Days
• Field of Search: None
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams
• CPC Class Codes: G10L 15/24 Speech recognition using non-acoustical features

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for modulating language model biasing. In some implementations, context data is received. A likely context associated with a user is determined based on at least a portion of the context data. One or more language model biasing parameters based at least on the likely context associated with the user is selected. A context confidence score associated with the likely context based on at least a portion of the context data is determined. One or more language model biasing parameters based at least on the context confidence score is adjusted. A baseline language model based at least on the one or more of the adjusted language model biasing parameters is biased. The baseline language model is provided for use by an automated speech recognizer (ASR).

Claims

1. A computer-implemented method comprising:

receiving, by an automated speech recognizer (ASR) that is configured to use a language model that has previously been biased for use in transcribing utterances, audio data corresponding to an utterance;

determining that, when the audio data was received, a particular context that is associated with biasing the language model was still applicable;

in response to determining that the particular context that is associated with biasing the language mode was still applicable when the audio data was received, generating, by the ASR, a transcription of the utterance using the language model that has previously been biased for use in transcribing utterances; and

providing a representation of the transcription for output.

2. The method of claim 1, wherein generating, by the ASR, a transcription of the utterance using the language model that has previously been biased for use in transcribing utterances comprises:

not returning the biased language model to a baseline state, or reducing an effect of the previous biasing of the language model.

3. The method of claim 1, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a predetermined period of time has not elapsed.

4. The method of claim 1, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a user that spoke the utterance is likely in a same context.

5. The method of claim 1, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a confidence score associated with the particular context does satisfy a threshold.

6. The method of claim 1, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a user that spoke the utterance is likely associated with another context that is consistent with the particular context.

7. The method of claim 1, wherein determining that, when the audio data was received, a particular context that is associated with biasing the language model was still applicable comprises:

determining that interactions from a user indicate a change in behavior since the language model was biased where the change is within a threshold amount.

8. A computer storage medium storing instructions that, when executed by data processing apparatus, cause one or more computers to perform operations comprising:

receiving, by an automated speech recognizer (ASR) that is configured to use a language model that has previously been biased for use in transcribing utterances, audio data corresponding to an utterance;

determining that, when the audio data was received, a particular context that is associated with biasing the language model was still applicable;

in response to determining that the particular context that is associated with biasing the language mode was still applicable when the audio data was received, generating, by the ASR, a transcription of the utterance using the language model that has previously been biased for use in transcribing utterances; and

providing a representation of the transcription for output.

9. The medium of claim 8, wherein generating, by the ASR, a transcription of the utterance using the language model that has previously been biased for use in transcribing utterances comprises:

not returning the biased language model to a baseline state, or reducing an effect of the previous biasing of the language model.

10. The medium of claim 8, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a predetermined period of time has not elapsed.

11. The medium of claim 8, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a user that spoke the utterance is likely in a same context.

12. The medium of claim 8, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a confidence score associated with the particular context does satisfy a threshold.

13. The medium of claim 8, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a user that spoke the utterance is likely associated with another context that is consistent with the particular context.

14. The medium of claim 8, wherein determining that, when the audio data was received, a particular context that is associated with biasing the language model was still applicable comprises:

determining that interactions from a user indicate a change in behavior since the language model was biased where the change is within a threshold amount.

15. A system comprising:

one or more computers including one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to perform operations comprising;

receiving, by an automated speech recognizer (ASR) that is configured to use a language model that has previously been biased for use in transcribing utterances, audio data corresponding to an utterance;

determining that, when the audio data was received, a particular context that is associated with biasing the language model was still applicable;

in response to determining that the particular context that is associated with biasing the language mode was still applicable when the audio data was received,

generating, by the ASR, a transcription of the utterance using the language model that has previously been biased for use in transcribing utterances; and

providing a representation of the transcription for output.

16. The system of claim 15, wherein generating, by the ASR, a transcription of the utterance using the language model that has previously been biased for use in transcribing utterances comprises:

not returning the biased language model to a baseline state, or reducing an effect of the previous biasing of the language model.

17. The system of claim 15, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a predetermined period of time has not elapsed.

18. The system of claim 15, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a user that spoke the utterance is likely in a same context.

19. The system of claim 15, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a confidence score associated with the particular context does satisfy a threshold.

20. The system of claim 15, wherein determining that the particular context that is associated with biasing the language model was still applicable comprises determining that a user that spoke the utterance is likely associated with another context that is consistent with the particular context.