Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer-readable media, for speech recognition using multi-dialect and multilingual models. In some implementations, audio data indicating audio characteristics of an utterance is received. Input features determined based on the audio data are provided to a speech recognition model that has been trained to output score indicating the likelihood of linguistic units for each of multiple different language or dialects. The speech recognition model can be one that has been trained using cluster adaptive training. Output that the speech recognition model generated in response to receiving the input features determined based on the audio data is received. A transcription of the utterance generated based on the output of the speech recognition model is provided.

Abstract: A computer-implemented method for transcribing an utterance includes receiving, at a computing system, speech data that characterizes an utterance of a user. A first set of candidate transcriptions of the utterance can be generated using a static class-based language model that includes a plurality of classes that are each populated with class-based terms selected independently of the utterance or the user. The computing system can then determine whether the first set of candidate transcriptions includes class-based terms. Based on whether the first set of candidate transcriptions includes class-based terms, the computing system can determine whether to generate a dynamic class-based language model that includes at least one class that is populated with class-based terms selected based on a context associated with at least one of the utterance and the user.

Abstract: A method for training a generative adversarial network (GAN)-based text-to-speech (TTS) model and a speech recognition model in unison includes obtaining a plurality of training text utterances. At each of a plurality of output steps for each training text utterance, the method also includes generating, for output by the GAN-Based TTS model, a synthetic speech representation of the corresponding training text utterance, and determining, using an adversarial discriminator of the GAN, an adversarial loss term indicative of an amount of acoustic noise disparity in one of the non-synthetic speech representations selected from the set of spoken training utterances relative to the corresponding synthetic speech representation of the corresponding training text utterance. The method also includes updating parameters of the GAN-based TTS model based on the adversarial loss term determined at each of the plurality of output steps for each training text utterance of the plurality of training text utterances.

Abstract: A method includes receiving acoustic features of a first utterance spoken by a first user that speaks with typical speech and processing the acoustic features of the first utterance using a general speech recognizer to generate a first transcription of the first utterance. The operations also include analyzing the first transcription of the first utterance to identify one or more bias terms in the first transcription and biasing the alternative speech recognizer on the one or more bias terms identified in the first transcription. The operations also include receiving acoustic features of a second utterance spoken by a second user that speaks with atypical speech and processing, using the alternative speech recognizer biased on the one or more terms identified in the first transcription, the acoustic features of the second utterance to generate a second transcription of the second utterance.

Abstract: A computer-implemented method for transcribing an utterance includes receiving, at a computing system, speech data that characterizes an utterance of a user. A first set of candidate transcriptions of the utterance can be generated using a static class-based language model that includes a plurality of classes that are each populated with class-based terms selected independently of the utterance or the user. The computing system can then determine whether the first set of candidate transcriptions includes class-based terms. Based on whether the first set of candidate transcriptions includes class-based terms, the computing system can determine whether to generate a dynamic class-based language model that includes at least one class that is populated with class-based terms selected based on a context associated with at least one of the utterance and the user.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for an automated calling system are disclosed. In one aspect, a method includes the actions of receiving audio data of an utterance spoken by a user who is having a telephone conversation with a bot. The actions further include determining a context of the telephone conversation. The actions further include determining a user intent of a first previous portion of the telephone conversation spoken by the user and a bot intent of a second previous portion of the telephone conversation outputted by a speech synthesizer of the bot. The actions further include, based on the audio data of the utterance, the context of the telephone conversation, the user intent, and the bot intent, generating synthesized speech of a reply by the bot to the utterance. The actions further include, providing, for output, the synthesized speech.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for detecting hotwords using a server. One of the methods includes receiving an audio signal encoding one or more utterances including a first utterance; determining whether at least a portion of the first utterance satisfies a first threshold of being at least a portion of a key phrase; in response to determining that at least the portion of the first utterance satisfies the first threshold of being at least a portion of a key phrase, sending the audio signal to a server system that determines whether the first utterance satisfies a second threshold of being the key phrase, the second threshold being more restrictive than the first threshold; and receiving tagged text data representing the one or more utterances encoded in the audio signal when the server system determines that the first utterance satisfies the second threshold.

Abstract: A method for contextual biasing for speech recognition includes obtaining a base automatic speech recognition (ASR) model trained on non-biased data and a sub-model trained on biased data representative of a particular domain. The method includes receiving a speech recognition request including audio data characterizing an utterance captured in streaming audio. The method further includes determining whether the speech recognition request includes a contextual indicator indicating the particular domain. When the speech recognition request does not include the contextual indicator, the method includes generating, using the base ASR model, a first speech recognition result of the utterance by processing the audio data.

Abstract: A method includes receiving training data that includes unspoken textual utterances, un-transcribed non-synthetic speech utterances, and transcribed non-synthetic speech utterances. Each unspoken textual utterance is not paired with any corresponding spoken utterance of non-synthetic speech. Each un-transcribed non-synthetic speech utterance not paired with a corresponding transcription. Each transcribed non-synthetic speech utterance paired with a corresponding transcription. The method also includes generating a corresponding alignment output for each unspoken textual utterance of the received training data using an alignment model. The method also includes pre-training an audio encoder on the alignment outputs generated for corresponding to the unspoken textual utterances, the un-transcribed non-synthetic speech utterances, and the transcribed non-synthetic speech utterances to teach the audio encoder to jointly learn shared speech and text representations.

Abstract: A method for speech conversion includes obtaining a speech conversion model configured to convert input utterances of human speech directly into corresponding output utterances of synthesized speech. The method further includes receiving a speech conversion request including input audio data corresponding to an utterance spoken by a target speaker associated with atypical speech and a speaker identifier uniquely identifying the target speaker. The method includes activating, using the speaker identifier, a particular sub-model for biasing the speech conversion model to recognize a type of the atypical speech associated with the target speaker identified by the speaker identifier.

Abstract: Methods, systems, and apparatus for receiving audio data corresponding to a user utterance and context data, identifying an initial set of one or more n-grams from the context data, generating an expanded set of one or more n-grams based on the initial set of n-grams, adjusting a language model based at least on the expanded set of n-grams, determining one or more speech recognition candidates for at least a portion of the user utterance using the adjusted language model, adjusting a score for a particular speech recognition candidate determined to be included in the expanded set of n-grams, determining a transcription of user utterance that includes at least one of the one or more speech recognition candidates, and providing the transcription of the user utterance for output.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for an automated calling system are disclosed. In one aspect, a method includes the actions of receiving audio data of an utterance spoken by a user who is having a telephone conversation with a bot. The actions further include determining a context of the telephone conversation. The actions further include determining a user intent of a first previous portion of the telephone conversation spoken by the user and a bot intent of a second previous portion of the telephone conversation outputted by a speech synthesizer of the bot. The actions further include, based on the audio data of the utterance, the context of the telephone conversation, the user intent, and the bot intent, generating synthesized speech of a reply by the bot to the utterance. The actions further include, providing, for output, the synthesized speech.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for end to end speech conversion are disclosed. In one aspect, a method includes the actions of receiving first audio data of a first utterance of one or more first terms spoken by a user. The actions further include providing the first audio data as an input to a model that is configured to receive first given audio data in a first voice and output second given audio data in a synthesized voice without performing speech recognition on the first given audio data. The actions further include receiving second audio data of a second utterance of the one or more first terms spoken in the synthesized voice. The actions further include providing, for output, the second audio data of the second utterance of the one or more first terms spoken in the synthesized voice.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for detecting hotwords using a server. One of the methods includes receiving an audio signal encoding one or more utterances including a first utterance; determining whether at least a portion of the first utterance satisfies a first threshold of being at least a portion of a key phrase; in response to determining that at least the portion of the first utterance satisfies the first threshold of being at least a portion of a key phrase, sending the audio signal to a server system that determines whether the first utterance satisfies a second threshold of being the key phrase, the second threshold being more restrictive than the first threshold; and receiving tagged text data representing the one or more utterances encoded in the audio signal when the server system determines that the first utterance satisfies the second threshold.

Abstract: A method for training a generative adversarial network (GAN)-based text-to-speech (TTS) model and a speech recognition model in unison includes obtaining a plurality of training text utterances. At each of a plurality of output steps for each training text utterance, the method also includes generating, for output by the GAN-Based TTS model, a synthetic speech representation of the corresponding training text utterance, and determining, using an adversarial discriminator of the GAN, an adversarial loss term indicative of an amount of acoustic noise disparity in one of the non-synthetic speech representations selected from the set of spoken training utterances relative to the corresponding synthetic speech representation of the corresponding training text utterance. The method also includes updating parameters of the GAN-based TTS model based on the adversarial loss term determined at each of the plurality of output steps for each training text utterance of the plurality of training text utterances.

Abstract: Methods, systems, and apparatus for receiving audio data corresponding to a user utterance and context data, identifying an initial set of one or more n-grams from the context data, generating an expanded set of one or more n-grams based on the initial set of n-grams, adjusting a language model based at least on the expanded set of n-grams, determining one or more speech recognition candidates for at least a portion of the user utterance using the adjusted language model, adjusting a score for a particular speech recognition candidate determined to be included in the expanded set of n-grams, determining a transcription of user utterance that includes at least one of the one or more speech recognition candidates, and providing the transcription of the user utterance for output.

Abstract: A method for denormalizing raw speech recognition results. The method includes receiving a speech input from a user and obtaining context metadata associated with the speech input. The context metadata indicates that the speech input includes dictated speech directed to a messaging application that is currently executing on a user device for inclusion in an electronic message. The method further includes generating, using a speech recognizer, a raw speech recognition result including an explicit punctuation term spoken by the user and corresponding to the speech input. Based on the context metadata, the method includes denormalizing the generated raw speech recognition result into denormalized text by applying an explicit punctuation denormalizer to convert the explicit punctuation term in the raw speech recognition result into a corresponding punctuation symbol and displaying the denormalized text including the corresponding punctuation symbol on a display screen of the user device.

Abstract: A method includes receiving acoustic features of a first utterance spoken by a first user that speaks with typical speech and processing the acoustic features of the first utterance using a general speech recognizer to generate a first transcription of the first utterance. The operations also include analyzing the first transcription of the first utterance to identify one or more bias terms in the first transcription and biasing the alternative speech recognizer on the one or more bias terms identified in the first transcription. The operations also include receiving acoustic features of a second utterance spoken by a second user that speaks with atypical speech and processing, using the alternative speech recognizer biased on the one or more terms identified in the first transcription, the acoustic features of the second utterance to generate a second transcription of the second utterance.

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).

Abstract: A method for training a generative adversarial network (GAN)-based text-to-speech (TTS) model and a speech recognition model in unison includes obtaining a plurality of training text utterances. At each of a plurality of output steps for each training text utterance, the method also includes generating, for output by the GAN-Based TTS model, a synthetic speech representation of the corresponding training text utterance, and determining, using an adversarial discriminator of the GAN, an adversarial loss term indicative of an amount of acoustic noise disparity in one of the non-synthetic speech representations selected from the set of spoken training utterances relative to the corresponding synthetic speech representation of the corresponding training text utterance. The method also includes updating parameters of the GAN-based TTS model based on the adversarial loss term determined at each of the plurality of output steps for each training text utterance of the plurality of training text utterances.