Performing speech recognition over a network and using speech recognition results based on determining that a network connection exists

1. A computer-implemented method comprising:

receiving, by a mobile device, a speech data representation of an utterance;

obtaining, by the mobile device, feature data extracted from the speech data representation of the utterance;

obtaining, by the mobile device, a transcription of the utterance using the feature data;

identifying, by the mobile device, contact information associated with the transcription;

determining, by the mobile device, that a network connection exists; and

initiating, by the mobile device, communication using the contact information based on determining that a network connection exists.

2. The method of claim 1 , comprising determining, by the mobile device, that the mobile device supports local feature extraction, wherein the feature data is extracted from the speech data representation of the mobile device after the mobile device determines that local feature extraction is supported.

3. The method of claim 1 , comprising determining, by the mobile device, that the mobile device supports local speech recognition, wherein the transcription is obtained based on determining that local speech recognition is supported.

4. The method of claim 1 , wherein obtaining the transcription comprises generating the transcription using one or more speech recognition models that are locally stored by mobile device.

5. The method of claim 1 , wherein identifying contact information associated with the transcription comprises:

determining that a portion of the transcription matches a name of a contact; and

determining that the contact information is associated with the contact.

6. The method of claim 1 , wherein determining that a network connection exists comprises determining whether a computer-to-telephone connection exists.

7. A non-transitory computer-readable medium storing software comprising

instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

receiving, by a mobile device, a speech data representation of an utterance;

obtaining, by the mobile device, feature data extracted from the speech data representation of the utterance;

obtaining, by the mobile device, a transcription of the utterance using the feature data;

identifying, by the mobile device, contact information associated with the transcription;

determining, by the mobile device, that a network connection exists; and

initiating, by the mobile device, communication using the contact information based on determining that a network connection exists.

8. The medium of claim 7 , the operations further comprising determining, by the mobile device, that the mobile device supports local feature extraction, wherein the feature data is extracted from the speech data representation of the mobile device after the mobile device determines that local feature extraction is supported.

9. The medium of claim 7 , the operations further comprising determining, by the mobile device, that the mobile device supports local speech recognition, wherein the transcription is obtained based on determining that local speech recognition is supported.

10. The medium of claim 7 , wherein obtaining the transcription comprises generating the transcription using one or more speech recognition models that are locally stored by mobile device.

11. The medium of claim 7 , wherein identifying contact information associated with the transcription comprises:

determining that a portion of the transcription matches a name of a contact; and

determining that the contact information is associated with the contact.

12. The medium of claim 7 , wherein determining that a network connection exists comprises determining whether a computer-to-telephone connection exists.

13. 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, by a mobile device, a speech data representation of an utterance;

obtaining, by the mobile device, feature data extracted from the speech data representation of the utterance;

obtaining, by the mobile device, a transcription of the utterance using the feature data;

identifying, by the mobile device, contact information associated with the transcription;

determining, by the mobile device, that a network connection exists; and

initiating, by the mobile device, communication using the contact information based on determining that a network connection exists.

14. The system of claim 13 , the operations further comprising determining, by the mobile device, that the mobile device supports local feature extraction, wherein the feature data is extracted from the speech data representation of the mobile device after the mobile device determines that local feature extraction is supported.

15. The system of claim 13 , the operations further comprising determining, by the mobile device, that the mobile device supports local speech recognition, wherein the transcription is obtained based on determining that local speech recognition is supported.

16. The system of claim 13 , wherein obtaining the transcription comprises generating the transcription using one or more speech recognition models that are locally stored by mobile device.

17. The system of claim 13 , wherein identifying contact information associated with the transcription comprises:

determining that a portion of the transcription matches a name of a contact; and

determining that the contact information is associated with the contact.

18. The system of claim 13 , wherein determining that a network connection exists comprises determining whether a computer-to-telephone connection exists.