Using non-parallel voice conversion for speech conversion models

1. A computer-implemented method when executed by data processing hardware causes the data processing hardware to perform operations comprising:

receiving a set of training utterances each comprising a non-synthetic speech representation of a corresponding utterance;

for each training utterance, generating a corresponding voice conversion synthetic speech representation by using a voice conversion model to convert the non-synthetic speech representation into the corresponding voice conversion synthetic representation of the corresponding utterance, the voice conversion model comprising a content encoder configured to encode each training utterance, the content encoder comprising a pretrained automatic speech recognition (ASR) encoder, wherein the non-synthetic speech representation and the voice conversion synthetic speech representation form a corresponding training utterance pair;

at each of a plurality of output steps for each training utterance pair in the set of training utterance pairs:

generating, for output by a speech recognition model, a first probability distribution over possible non-synthetic speech recognition hypotheses for the corresponding non-synthetic speech representation of the corresponding utterance;

generating, for output by the speech recognition model, a second probability distribution over possible synthetic speech recognition hypotheses for the corresponding synthetic speech representation of the corresponding utterance; and

determining a consistent loss term for the corresponding training utterance pair based on the first probability distribution over possible non-synthetic speech recognition hypotheses and the second probability distribution over possible non-synthetic speech recognition hypotheses; and

updating parameters of the speech recognition model based on the consistent loss term determined at each of the plurality of output steps for each training utterance pair in the set of training utterance pairs,

wherein the voice conversion model comprises:

the content encoder further configured to:

receive, as input, input speech; and

generate, as output, an initial latent representation;

a vector quantization-variational autoencoder (VQ-VAE) layer configured to:

receive, as input, the initial latent representation; and

generate, as output, a latent representation of linguistic content extracted from the input speech that discards speaking style variations in the input speech; and

a decoder conditioned on a speaker embedding and configured to generate output speech based on the latent representation of linguistic content.

2. The method of claim 1 , wherein the operations further comprise, at each of the plurality of output steps for each training utterance pair in the set of training utterance pairs:

generating a non-synthetic speech loss term based on the first probability distribution over possible non-synthetic speech recognition hypotheses for the corresponding non-synthetic speech representation of the corresponding utterance and a ground-truth transcription of the corresponding utterance; and

generating a synthetic speech loss term based on the second probability distribution over possible synthetic speech recognition hypotheses for the corresponding synthetic speech representation of the corresponding utterance and the ground-truth transcription of the corresponding utterance.

3. The method of claim 2 , wherein the operations further comprise updating parameters of the speech recognition model based on the non-synthetic and synthetic speech loss terms generated at each of the plurality of output steps for each training utterance pair in the set of training utterance pairs.

4. The method of claim 1 , wherein:

the non-synthetic speech representation of the corresponding utterance comprises speech spoken by a source speaker and conveys an accent/locale associated with the source speaker; and

the voice conversion synthetic representation of the corresponding utterance comprises synthesized speech representing the corresponding utterance that conveys an accent/local associated with a target speaker.

5. The method of claim 4 , wherein the accent/local associated with the source speaker is different than the accent/local associated with the target speaker.

6. The method of claim 4 , wherein the voice conversion synthetic speech representation conveys the same linguistic content as the non-synthetic speech representation of the corresponding utterance.

7. The method of claim 1 , wherein the voice conversion model comprises a non-parallel voice conversion model.

8. The method of claim 1 , wherein the pretrained ASR encoder was previously trained on an ASR loss for a speech recognition task.

9. The method of claim 8 , wherein the parameters of the pretrained ASR encoder remain fixed while training the VQ-VAE layer and the decoder of the voice conversion model.

10. The method of claim 1 , wherein the VQ-VAE layer is trained using a VQ loss based on the latent representation of linguistic content generated for each timestep, the VQ loss encouraging the VQ-VAE layer to minimize a distance between an output and a nearest codebook.

11. The method of claim 1 , wherein the decoder is configured to:

receive, as input, the latent representation of linguistic content for the input speech and the speaker embedding; and

generate, as output, the output speech comprising a reconstruction of the input speech.

12. A system comprising:

data processing hardware; and

memory hardware in communication with the data processing hardware and storing instructions that when executed by the data processing hardware causes the data processing hardware to perform operations comprising:

receiving a set of training utterances each comprising a non-synthetic speech representation of a corresponding utterance;

for each training utterance, generating a corresponding voice conversion synthetic speech representation by using a voice conversion model to convert the non-synthetic speech representation into the corresponding voice conversion synthetic representation of the corresponding utterance, the voice conversion model comprising a content encoder configured to encode each training utterance, the content encoder comprising a pretrained automatic speech recognition (ASR) encoder, wherein the non-synthetic speech representation and the voice conversion synthetic speech representation form a corresponding training utterance pair;

at each of a plurality of output steps for each training utterance pair in the set of training utterance pairs:

generating, for output by a speech recognition model, a first probability distribution over possible non-synthetic speech recognition hypotheses for the corresponding non-synthetic speech representation of the corresponding utterance;

generating, for output by the speech recognition model, a second probability distribution over possible synthetic speech recognition hypotheses for the corresponding synthetic speech representation of the corresponding utterance; and

determining a consistent loss term for the corresponding training utterance pair based on the first probability distribution over possible non-synthetic speech recognition hypotheses and the second probability distribution over possible non-synthetic speech recognition hypotheses; and

updating parameters of the speech recognition model based on the consistent loss term determined at each of the plurality of output steps for each training utterance pair in the set of training utterance pairs,

wherein the voice conversion model comprises:

the content encoder further configured to:

receive, as input, input speech; and

generate, as output, an initial latent representation;

a vector quantization-variational autoencoder (VQ-VAE) layer configured to:

receive, as input, the initial latent representation; and

generate, as output, a latent representation of linguistic content extracted from the input speech that discards speaking style variations in the input speech; and

a decoder conditioned on a speaker embedding and configured to generate output speech based on the latent representation of linguistic content.

13. The system of claim 12 , wherein the operations further comprise, at each of the plurality of output steps for each training utterance pair in the set of training utterance pairs:

generating a non-synthetic speech loss term based on the first probability distribution over possible non-synthetic speech recognition hypotheses for the corresponding non-synthetic speech representation of the corresponding utterance and a ground-truth transcription of the corresponding utterance; and

generating a synthetic speech loss term based on the second probability distribution over possible synthetic speech recognition hypotheses for the corresponding synthetic speech representation of the corresponding utterance and the ground-truth transcription of the corresponding utterance.

14. The system of claim 13 , wherein the operations further comprise updating parameters of the speech recognition model based on the non-synthetic and synthetic speech loss terms generated at each of the plurality of output steps for each training utterance pair in the set of training utterance pairs.

15. The system of claim 12 , wherein:

the non-synthetic speech representation of the corresponding utterance comprises speech spoken by a source speaker and conveys an accent/locale associated with the source speaker; and

the voice conversion synthetic representation of the corresponding utterance comprises synthesized speech representing the corresponding utterance that conveys an accent/local associated with a target speaker.

16. The system of claim 15 , wherein the accent/local associated with the source speaker is different than the accent/local associated with the target speaker.

17. The system of claim 15 , wherein the voice conversion synthetic speech representation conveys the same linguistic content as the non-synthetic speech representation of the corresponding utterance.

18. The system of claim 12 , wherein the voice conversion model comprises a non-parallel voice conversion model.

19. The system of claim 12 , wherein the ASR encoder was previously trained on ASR loss for a speech recognition task.

20. The system of claim 19 , wherein the parameters of the pretrained ASR encoder remain fixed while training the VQ-VAE layer and the decoder of the voice conversion model.

21. The system of claim 12 , wherein the VQ-VAE layer is trained using a VQ loss based on the latent representation of linguistic content generated for each timestep, the VQ loss encouraging the VQ-VAE layer to minimize a distance between an output and a nearest codebook.

22. The system of claim 12 , wherein the decoder is configured to:

receive, as input, the latent representation of linguistic content for the input speech and the speaker embedding; and

generate, as output, the output speech comprising a reconstruction of the input speech.