Multi-user authentication on a device
In some implementations, an utterance is determined to include a particular user speaking a hotword based at least on a first set of samples of the particular user speaking the hotword. In response to determining that an utterance includes a particular user speaking a hotword based at least on a first set of samples of the particular user speaking the hotword, at least a portion of the utterance is stored as a new sample. A second set of samples of the particular user speaking the utterance is obtained, where the second set of samples includes the new sample and less than all the samples in the first set of samples. A second utterance is determined to include the particular user speaking the hotword based at least on the second set of samples of the user speaking the hotword.
1. A computer-implemented method comprising:
determining that an utterance captured by a first speech-enabled device includes a particular user speaking a hotword based at least on a first hotword detection model generated from a first set of samples of the particular user speaking the hotword;
in response to determining that the utterance includes the particular user speaking the hotword based at least on the first hotword detection model generated from the first set of samples of the particular user speaking the hotword, selecting at least a portion of the utterance as a new sample for inclusion in a second set of samples of the particular user speaking the hotword; and
providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples.
2. The method of claim 1 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples comprises:
determining that the second speech-enabled device is used by the particular user.
3. The method of claim 1 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples comprises:
providing the new sample to the second speech-enabled device to generate the second hotword detection model.
4. The method of claim 3 , wherein providing the new sample to the second speech-enabled device to train the second hotword detection model comprises:
determining the second speech-enable device does not already store the new sample.
5. The method of claim 4 , wherein determining the second speech-enable device does not already store the new sample comprises:
receiving an identifier of a set of samples that the second speech-enable device has stored; and
determining that the set of samples does not include the new sample.
6. The method of claim 4 , wherein determining the second speech-enable device does not already store the new sample comprises:
receiving, for each sample stored by the second speech-enabled device, an identifier of the sample; and
determining that none of the identifiers received match an identifier of the new sample.
7. The method of claim 1 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples is in response to selecting at least the portion of the utterance as the new sample for inclusion in the second set of samples of the particular user speaking the hotword.
8. The method of claim 1 , wherein selecting at least a portion of the utterance as a new sample for inclusion in a second set of samples of the particular user speaking the hotword comprises:
determining that a type of the first speech-enabled device matches a type of the second speech-enabled device; and
in response to determining that the type of the first speech-enabled device matches the type of the second speech-enabled device, selecting at least the portion of the utterance as the new sample for inclusion in the second set of samples of the particular user speaking the hotword.
9. 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:
determining that an utterance captured by a first speech-enabled device includes a particular user speaking a hotword based at least on a first hotword detection model generated from a first set of samples of the particular user speaking the hotword;
in response to determining that the utterance includes the particular user speaking the hotword based at least on the first hotword detection model generated from the first set of samples of the particular user speaking the hotword, selecting at least a portion of the utterance as a new sample for inclusion in a second set of samples of the particular user speaking the hotword; and
providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples.
10. The system of claim 9 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples comprises:
determining that the second speech-enabled device is used by the particular user.
11. The system of claim 9 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples comprises:
providing the new sample to the second speech-enabled device to generate the second hotword detection model.
12. The system of claim 11 , wherein providing the new sample to the second speech-enabled device to generate the second hotword detection model comprises:
determining the second speech-enable device does not already store the new sample.
13. The system of claim 12 , wherein determining the second speech-enable device does not already store the new sample comprises:
receiving an identifier of a set of samples that the second speech-enable device has stored; and
determining that the set of samples does not include the new sample.
14. The system of claim 12 , wherein determining the second speech-enable device does not already store the new sample comprises:
receiving, for each sample stored by the second speech-enabled device, an identifier of the sample; and
determining that none of the identifiers received match an identifier of the new sample.
15. The system of claim 9 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples is in response to selecting at least the portion of the utterance as the new sample for inclusion in the second set of samples of the particular user speaking the hotword.
16. The system of claim 9 , wherein selecting at least a portion of the utterance as a new sample for inclusion in a second set of samples of the particular user speaking the hotword comprises:
determining that a type of the first speech-enabled device matches a type of the second speech-enabled device; and
in response to determining that the type of the first speech-enabled device matches the type of the second speech-enabled device, selecting at least the portion of the utterance as the new sample for inclusion in the second set of samples of the particular user speaking the hotword.
17. 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:
determining that an utterance captured by a first speech-enabled device includes a particular user speaking a hotword based at least on a first hotword detection model generated from a first set of samples of the particular user speaking the hotword;
in response to determining that the utterance includes the particular user speaking the hotword based at least on the first hotword detection model generated from the first set of samples of the particular user speaking the hotword, selecting at least a portion of the utterance as a new sample for inclusion in a second set of samples of the particular user speaking the hotword; and
providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples.
18. The medium of claim 17 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples comprises:
determining that the second speech-enabled device is used by the particular user.
19. The medium of claim 17 , wherein providing the second set of samples of the particular user speaking the hotword to a second speech-enabled device, that is used by the particular user, to train a second hotword detection model from the second set of samples comprises:
providing the new sample to the second speech-enabled device to generate the second hotword detection model.
20. The medium of claim 19 , wherein providing the new sample to the second speech-enabled device to generate the second hotword detection model comprises:
determining the second speech-enable device does not already store the new sample.