Speech recognition using device docking context

• US 8,396,709 B2
• Filed: 09/30/2011
• Issued: 03/12/2013
• Est. Priority Date: 01/21/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Lloyd, Matthew I., Risbood, Pankaj
• Primary Examiner(s): Neway, Samuel G
• Application Number:
  US13/250,466
• Publication Number:
  US 20120191449A1
• Time in Patent Office: 529 Days
• Field of Search: 704231-257
• US Class Current: 704/231
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: H04M 1/04 Supports for telephone transmitters or receivers
• CPC Class Codes: H04M 1/6075 adapted for handsfree use in a vehicle H04M1/6058 takes precedence; arrangements for holding telephones in a vehicle B60R11/0241
• CPC Class Codes: H04M 2250/74 with voice recognition means

Abstract

Methods, systems, and apparatuses, including computer programs encoded on a computer storage medium, for performing speech recognition using dock context. In one aspect, a method includes accessing audio data that includes encoded speech. Information that indicates a docking context of a client device is accessed, the docking context being associated with the audio data. A plurality of language models is identified. At least one of the plurality of language models is selected based on the docking context. Speech recognition is performed on the audio data using the selected language model to identify a transcription for a portion of the audio data.

Claims

1. A computer-implemented method, comprising:

accessing audio data that includes encoded speech;

accessing information that indicates a docking context of a client device, the docking context being associated with the audio data;

identifying a plurality of language models;

identifying multiple sets of weighting values for the plurality of language models, the multiple sets of weighting values comprising at least

a first set of multiple weighting values that correspond to multiple language models of the plurality of language models, the first set of multiple weighting values being associated with a first key phrase, wherein the first set of multiple weighting values is used to bias selection of a language model when a user utters the first key phrase, and

a second set of multiple weighting values that correspond to multiple language models of the plurality of language models, the second set of multiple weighting values being associated with a second key phrase, the second set of multiple weighting values being different from the first set of multiple weighting values, and the second key phrase being different from the first key phrase;

determining that the docking context indicates docking of the client device with a docking station of a first type;

based on determining that the docking context indicates docking of the client device with the docking station of the first type, selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase;

selecting at least a first language model of the plurality of language models using the first set of multiple weighting values associated with the first key phrase; and

performing speech recognition on the audio data using the first language model to identify a transcription for a portion of the audio data.

2. The computer-implemented method of claim 1, wherein the information that indicates a docking context of the client device indicates a connection between the client device and a second device with which the client device is physically connected.

3. The computer-implemented method of claim 1, wherein the information that indicates a docking context of the client device indicates a connection between the client device and a second device with which the client device is wirelessly connected.

4. The computer-implemented method of claim 1, wherein identifying the multiple sets of weighting values for the plurality of language models comprises identifying multiple sets of weighting values, each set of weighting values indicating probabilities that the respective language models of the plurality of language models will indicate a correct transcription for speech when a key phrase associated with the set of weighting values occurs in the speech.

5. The computer-implemented method of claim 1, wherein the speech encoded in the audio data was detected by the client device, and wherein the information that indicates a docking context indicates whether the client device was connected to a docking station while the speech encoded in the audio data was detected by the client device.

6. The computer-implemented method of claim 1, wherein the speech encoded in the audio data was detected by the client device, and wherein the information that indicates a docking context indicates a type of docking station to which the client device was connected while the speech encoded in the audio data was detected by the client device.

7. The computer-implemented method of claim 1, wherein the encoded speech includes one or more spoken query terms, and wherein the transcription includes a transcription of the spoken query terms, and wherein the method further comprises:

causing a search engine to perform a search using the transcription of the one or more spoken query terms; and

providing information indicating the results of the search query to the client device.

8. The computer-implemented method of claim 1, wherein each of the plurality of language models is trained for a particular topical category of words.

9. The computer-implemented method of claim 1, wherein:

determining that the docking context indicates docking of the client device with a docking station of a first type comprises determining that the docking context indicates that the docking station is a vehicle docking station; and

based on determining that the docking context indicates docking of the client device with the docking station of the first type, selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase comprises;

in response to determining that the docking context indicates that the docking station is a vehicle docking station, selecting a set of weighting values that increases the probability that a navigation language model is selected relative to the other language models in the plurality of language models.

10. The method of claim 1, wherein the first type is selected from a group consisting of a media player docking station type, a vehicle docking station type, and a computer docking station type.

11. The method of claim 1, wherein:

determining that the docking context indicates docking of the client device with the docking station of the first type comprises determining that the docking station is a vehicle docking station;

selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase comprises selecting a set of weighting values associated with a key phrase that is associated with navigation; and

selecting at least the first language model of the plurality of language models comprises selecting a language model associated with navigation.

12. The method of claim 1, wherein:

determining that the docking context indicates docking of the client device with the docking station of the first type comprises determining that the docking station is a media playing device;

selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase comprises selecting a set of weighting values associated with a key phrase that is associated with media; and

selecting at least the first language model of the plurality of language models comprises selecting a language model associated with media.

13. The method of claim 1, wherein:

determining that the docking context indicates docking of the client device with the docking station of the first type comprises determining that the docking station is a computer;

selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase comprises selecting a set of weighting values associated with a key phrase that is associated with Internet searching; and

selecting at least the first language model of the plurality of language models comprises selecting a language model associated with Internet searching.

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

accessing second audio data that includes encoded speech;

accessing information that indicates a second docking context of the client device, the second docking context being associated with the second audio data;

determining that the second docking context indicates docking of the client device with a second docking station of a second type, the second type being different from the first type;

based on determining that the second docking context indicates docking of the client device with the second docking station of the second type, selecting, from among the multiple sets of weighting values, the second set of multiple weighting values associated with the second key phrase;

selecting at least a second language model of the plurality of language models using second set of multiple weighting values associated with the second key phrase, the second language model being different from the first language model; and

performing speech recognition on the second audio data using the second language model to identify a transcription for a portion of the second audio data,

wherein docking stations of the first type provide capabilities for one or more first manners of using the client device, and wherein docking stations of the second type provide capabilities for one or more second manners of using the client device that are different from the one or more first manners of using the client device.

15. The computer-implemented method of claim 1, wherein identifying multiple sets of weighting values for the plurality of language models comprises identifying multiple sets of weighting values that bias selection of language models, each of the multiple sets of weighting values biasing selection of language models for recognizing utterances in audio input when a key phrase associated with the set of weighting values is recognized in the audio input.

16. The computer-implemented method of claim 15, wherein selecting, from among the multiple sets of weighting values, the set of weighting values associated with the first key phrase comprises selecting the first set of multiple weighting values based on determining that the first docking context indicates docking of the client device with the first docking station of the first type, without recognizing the first key phrase associated with the first set of multiple weighting values.

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

accessing second audio data that includes encoded speech;

determining that the encoded speech included in the second audio data includes the first key phrase;

based on determining that the encoded speech included in the second audio data includes the first key phrase, selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase;

selecting at least a first language model of the plurality of language models using the first set of multiple weighting values associated with the first key phrase; and

performing speech recognition on the second audio data using the first language model to identify a transcription for a portion of the second audio data.

18. The computer-implemented method of claim 1, wherein identifying multiple sets of weighting values for the plurality of language models comprises identifying sets of weighting values that each comprise a weighting value corresponding to each language model of the plurality of language models.

19. The computer implemented method of claim 1, wherein identifying multiple sets of weighting values for the plurality of language models comprises identifying the second set of multiple weighting values, wherein the second set of multiple weighting values is used to bias selection of a language model when a user utters the second key phrase.

20. A system comprising:

one or more processors; and

a computer-readable medium coupled to the one or more processors having instructions stored thereon which, when executed by the one or more processors, cause the system to perform operations comprising;

accessing audio data that includes encoded speech;

accessing information that indicates a docking context of a client device, the docking context being associated with the audio data;

identifying a plurality of language models;

identifying multiple sets of weighting values for the plurality of language models, the multiple sets of weighting values comprising at least

a first set of multiple weighting values that correspond to multiple language models of the plurality of language models, the first set of multiple weighting values being associated with a first key phrase, wherein the first set of multiple weighting values is used to bias selection of a language model when a user utters the first key phrase, and

a second set of multiple weighting values that correspond to multiple language models of the plurality of language models, the second set of multiple weighting values being associated with a second key phrase, the second set of multiple weighting values being different from the first set of multiple weighting values, and the second key phrase being different from the first key phrase;

determining that the docking context indicates docking of the client device with a docking station of a first type;

based on determining that the docking context indicates docking of the client device with the docking station of the first type, selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase;

selecting at least a first language model of the plurality of language models using the first set of multiple weighting values associated with the first key phrase; and

performing speech recognition on the audio data using the first language model to identify a transcription for a portion of the audio data.

21. The system of claim 20, wherein identifying the multiple sets of weighting values for the plurality of language models comprises identifying multiple sets of weighting values, each set of weighting values indicating probabilities that the respective language models of the plurality of language models will indicate a correct transcription for speech when a key phrase associated with the set of weighting values occurs in the speech.

22. The system of claim 20, wherein the speech encoded in the audio data was detected by the client device, and wherein the information that indicates a docking context indicates whether the client device was connected to a docking station while the speech encoded in the audio data was detected by the client device.

23. The system of claim 20, wherein the speech encoded in the audio data was detected by the client device, and wherein the information that indicates a docking context indicates a type of docking station to which the client device was connected while the speech encoded in the audio data was detected by the client device.

24. A non-transitory computer storage medium encoded with a computer program, the program comprising instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:

accessing audio data that includes encoded speech;

accessing information that indicates a docking context of a client device, the docking context being associated with the audio data;

identifying a plurality of language models;

identifying multiple sets of weighting values for the plurality of language models, the multiple sets of weighting values comprising at least

a first set of multiple weighting values that correspond to multiple language models of the plurality of language models, the first set of multiple weighting values being associated with a first key phrase, wherein the first set of multiple weighting values is used to bias selection of a language model when a user utters the first key phrase, and

a second set of multiple weighting values that correspond to multiple language models of the plurality of language models, the second set of multiple weighting values being associated with a second key phrase, the second set of multiple weighting values being different from the first set of multiple weighting values, and the second key phrase being different from the first key phrase;

determining that the docking context indicates docking of the client device with a docking station of a first type;

based on determining that the docking context indicates docking of the client device with the docking station of the first type, selecting, from among the multiple sets of weighting values, the first set of multiple weighting values associated with the first key phrase;

selecting at least a first language model of the plurality of language models using the first set of multiple weighting values associated with the first key phrase; and

performing speech recognition on the audio data using the first language model to identify a transcription for a portion of the audio data.

25. The computer storage medium of claim 24, wherein identifying the multiple sets of weighting values for the plurality of language models comprises identifying multiple sets of weighting values, each set of weighting values indicating probabilities that the respective language models of the plurality of language models will indicate a correct transcription for speech when a key phrase associated with the set of weighting values occurs in the speech.

26. The computer storage medium of claim 24, wherein the speech encoded in the audio data was detected by the client device, and wherein the information that indicates a docking context indicates whether the client device was connected to a docking station while the speech encoded in the audio data was detected by the client device.

27. The computer storage medium of claim 24, wherein the speech encoded in the audio data was detected by the client device, and wherein the information that indicates a docking context indicates a type of docking station to which the client device was connected while the speech encoded in the audio data was detected by the client device.

28. The computer storage medium of claim 24, wherein the encoded speech includes one or more spoken query terms, and wherein the transcription includes a transcription of the spoken query terms, and wherein the operations further comprise:

causing a search engine to perform a search using the transcription of the one or more spoken query terms; and

providing information indicating the results of the search query to the client device.