Abstract: Implementations disclosed herein are directed to systems and methods for evaluating new feature(s) for client device(s) based on performance measure(s) of the client device(s) and/or the new feature(s). The new feature(s) can include, for example, machine learning (ML) model(s), non-ML software-enabled functionality, non-ML hardware-enabled functionality, and/or ML or non-ML software application features for a given software application utilized by the client device(s). The client device(s) can generate the performance measure(s) by processing a plurality of testing instances for the new feature(s). The performance measure(s) can include, for example, latency measure(s), memory consumption measure(s), CPU usage measure(s), precision and/or recall measure(s), and/or other measures. In some implementations, the new feature(s) may be activated for use locally at the client device(s) based on the performance measure(s), and optionally at other client device(s) that share the same device characteristics.

Abstract: Implementations relate to ascertaining to what extent predictions, generated using a machine learning model, can be effectively reconstructed from model updates, where the model updates are generated based on those predictions and based on applying a particular loss technique (e.g., a particular cross-entropy loss technique). Some implementations disclosed generate measures that each indicate a degree of conformity between a corresponding reconstruction, generated using a corresponding model update, and a corresponding prediction. In some of those implementations, the measures are utilized in determining whether to utilize the particular loss technique (utilized in generating the model updates) in federated learning of the machine learning model and/or of additional machine learning model(s).

Abstract: Implementations described herein are directed to a framework for decentralized learning of large global machine learning (ML) model(s). In various implementations, remote processor(s) of a remote system can identify a global ML model, select client devices to participate in a given round of decentralized learning of the global ML model, and transmit, to each of the client devices, a processed version of the global ML model that is of a reduced transferrable size. Further, client device processor(s) of a client device can receive the processed version of the global ML model, obtain corresponding client data, perform partial model training, based on processing the corresponding client data, for the processed version of the global ML model to generate a corresponding update, and transmit the corresponding update back to the remote system. Moreover, the remote processor(s) can update, based on at least the corresponding update, the global ML model.

Abstract: An unintentional memorization measure can be used to determine whether an automatic speech recognition (ASR) model has unintentionally memorized one or more phrases during training of the ASR model. Various implementations include generating one or more candidate transcripts based on the vocabulary of the ASR model. For example, the system can generate a candidate transcript by appending a token of the vocabulary to a previous candidate transcript. Various implementations include processing the candidate transcript using a speech synthesis model to generate synthesized speech audio data that includes synthesized speech of the candidate transcript. Additionally or alternatively, the synthesized speech audio data can be processed using the ASR model to generate ASR output. Various implementations can include generating a loss based on comparing the ASR output and the candidate transcript.

Abstract: Implementations are directed to efficient federated learning of machine learning (ML) model(s) through on-the-fly decompression and compression of model parameters, of the ML model(s), when facilitating forward propagation and/or back propagation at client device(s). For example, implementations can transmit, from a remote system to a client device, a compressed on-device ML model that includes some compressed parameters. Further, the client device can, in performing forward propagation and/or back propagation using the on-device ML model, decompress those compressed parameters on-the-fly as the parameters are needed for the propagation. The propagation will utilize the decompressed parameters that were decompressed on the fly.

Abstract: Implementations disclosed herein are directed to systems and methods for evaluating new feature(s) for client device(s) based on performance measure(s) of the client device(s) and/or the new feature(s). The new feature(s) can include, for example, machine learning (ML) model(s), non-ML software-enabled functionality, non-ML hardware-enabled functionality, and/or ML or non-ML software application features for a given software application utilized by the client device(s). The client device(s) can generate the performance measure(s) by processing a plurality of testing instances for the new feature(s). The performance measure(s) can include, for example, latency measure(s), memory consumption measure(s), CPU usage measure(s), precision and/or recall measure(s), and/or other measures. In some implementations, the new feature(s) may be activated for use locally at the client device(s) based on the performance measure(s), and optionally at other client device(s) that share the same device characteristics.

Abstract: A method of a multilingual ASR model includes receiving a sequence of acoustic frames characterizing an utterance of speech. At a plurality of output steps, the method further includes generating a first higher order feature representation for an acoustic frame by a first encoder that includes a first plurality of multi-head attention layers; generating a second higher order feature representation for a corresponding first higher order feature representation by a second encoder that includes a second plurality of multi-head attention layers; and generating, by a first decoder, a first probability distribution over possible speech recognition hypotheses based on the second higher order feature representation and a sequence of N previous non-blank symbols. A gating layer of each respective MoE layer configured to dynamically route an output from a previous multi-head attention layer at each of the plurality of output steps to a respective pair of feed-forward expert networks.

Abstract: Processor(s) of a client device can: identify a textual segment stored locally at the client device; process the textual segment, using an on-device TTS generator model, to generate synthesized speech audio data that includes synthesized speech of the textual segment; process the synthesized speech, using an on-device ASR model to generate predicted ASR output; and generate a gradient based on comparing the predicted ASR output to ground truth output corresponding to the textual segment. Processor(s) of the client device can also: process the synthesized speech audio data using an on-device TTS generator model to make a prediction; and generate a gradient based on the prediction. In these implementations, the generated gradient(s) can be used to update weight(s) of the respective on-device model(s) and/or transmitted to a remote system for use in remote updating of respective global model(s). The updated weight(s) and/or the updated model(s) can be transmitted to client device(s).

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for cross input modality learning in a mobile device are disclosed. In one aspect, a method includes activating a first modality user input mode in which user inputs by way of a first modality are recognized using a first modality recognizer; and receiving a user input by way of the first modality. The method includes, obtaining, as a result of the first modality recognizer recognizing the user input, a transcription that includes a particular term; and generating an input context data structure that references at least the particular term. The method further includes, transmitting, by the first modality recognizer, the input context data structure to a second modality recognizer for use in updating a second modality recognition model associated with the second modality recognizer.

Abstract: An unintentional memorization measure can be used to determine whether an automatic speech recognition (ASR) model has unintentionally memorized one or more phrases during training of the ASR model. Various implementations include generating one or more candidate transcripts based on the vocabulary of the ASR model. For example, the system can generate a candidate transcript by appending a token of the vocabulary to a previous candidate transcript. Various implementations include processing the candidate transcript using a speech synthesis model to generate synthesized speech audio data that includes synthesized speech of the candidate transcript. Additionally or alternatively, the synthesized speech audio data can be processed using the ASR model to generate ASR output. Various implementations can include generating a loss based on comparing the ASR output and the candidate transcript.

Abstract: Processor(s) of a client device can: identify a textual segment stored locally at the client device; process the textual segment, using an on-device TTS generator model, to generate synthesized speech audio data that includes synthesized speech of the textual segment; process the synthesized speech, using an on-device ASR model to generate predicted ASR output; and generate a gradient based on comparing the predicted ASR output to ground truth output corresponding to the textual segment. Processor(s) of the client device can also: process the synthesized speech audio data using an on-device TTS generator model to make a prediction; and generate a gradient based on the prediction. In these implementations, the generated gradient(s) can be used to update weight(s) of the respective on-device model(s) and/or transmitted to a remote system for use in remote updating of respective global model(s). The updated weight(s) and/or the updated model(s) can be transmitted to client device(s).

Abstract: Techniques are disclosed that enable generating jointly probable output by processing input using a multi-stream recurrent neural network transducer (MS RNN-T) model. Various implementations include generating a first output sequence and a second output sequence by processing a single input sequence using the MS RNN-T, where the first output sequence is jointly probable with the second output sequence. Additional or alternative techniques are disclosed that enable generating output by processing multiple input sequences using the MS RNN-T. Various implementations include processing a first input sequence and a second input sequence using the MS RNN-T to generate output. In some implementations, the MS RNN-T can be used to process two or more input sequences to generate two or more jointly probable output sequences.

Abstract: Implementations disclosed herein are directed to techniques for enabling decentralized learning of global language models (LMs). Remote processor(s) of a remote system can obtain a global LM that includes a global embedding matrix, generate a global embedding mask for the global embedding matrix using a masking technique, apply the global embedding mask to global embedding matrix to generate a sparsified global LM that includes a masked global embedding matrix that is a masked version of the global embedding matrix, transmit the sparsified global LM to computing device(s) that are participating in a given round of decentralized learning for the global language model, receive corresponding updates from the computing device(s), and cause the global LM to be updated based on the corresponding updates. By generating the global embedding mask and applying it to the global embedding matrix, the transferable size of the global LM is reduced thereby enabling decentralized learning thereof.

Abstract: Techniques are disclosed that enable generating jointly probable output by processing input using a multi-stream recurrent neural network transducer (MS RNN-T) model. Various implementations include generating a first output sequence and a second output sequence by processing a single input sequence using the MS RNN-T, where the first output sequence is jointly probable with the second output sequence. Additional or alternative techniques are disclosed that enable generating output by processing multiple input sequences using the MS RNN-T. Various implementations include processing a first input sequence and a second input sequence using the MS RNN-T to generate output. In some implementations, the MS RNN-T can be used to process two or more input sequences to generate two or more jointly probable output sequences.

Abstract: Implementations disclosed herein are directed to systems and methods for evaluating on-device machine learning (ML) model(s) based on performance measure(s) of client device(s) and/or the on-device ML model(s). The client device(s) can include on-device memory that stores the on-device ML model(s) and a plurality of testing instances for the on-device ML model(s). When certain condition(s) are satisfied, the client device(s) can process, using the on-device ML model(s), the plurality of testing instances to generate the performance measure(s). The performance measure(s) can include, for example, latency measure(s), memory consumption measure(s), CPU usage measure(s), ML model measure(s) (e.g., precision and/or recall), and/or other measures. In some implementations, the on-device ML model(s) can be activated (or kept active) for use locally at the client device(s) based on the performance measure(s). In other implementations, the on-device ML model(s) can be sparsified based on the performance measure(s).

Abstract: Processor(s) of a client device can: identify a textual segment stored locally at the client device; process the textual segment, using a speech synthesis model stored locally at the client device, to generate synthesized speech audio data that includes synthesized speech of the identified textual segment; process the synthesized speech, using an on-device speech recognition model that is stored locally at the client device, to generate predicted output; and generate a gradient based on comparing the predicted output to ground truth output that corresponds to the textual segment. In some implementations, the generated gradient is used, by processor(s) of the client device, to update weights of the on-device speech recognition model. In some implementations, the generated gradient is additionally or alternatively transmitted to a remote system for use in remote updating of global weights of a global speech recognition model.

Abstract: Implementations disclosed herein are directed to unsupervised federated training of global machine learning (“ML”) model layers that, after the federated training, can be combined with additional layer(s), thereby resulting in a combined ML model. Processor(s) can: detect audio data that captures a spoken utterance of a user of a client device; process, using a local ML model, the audio data to generate predicted output(s); generate, using unsupervised learning locally at the client device, a gradient based on the predicted output(s); transmit the gradient to a remote system; update weight(s) of the global ML model layers based on the gradient; subsequent to updating the weight(s), train, using supervised learning remotely at the remote system, a combined ML model that includes the updated global ML model layers and additional layer(s); transmit the combined ML model to the client device; and use the combined ML model to make prediction(s) at the client device.

Abstract: Implementations described herein identify and correct automatic speech recognition (ASR) misrecognitions. For example, on-device processor(s) of a client device may generate a predicted textual segment that is predicted to correspond to spoken utterance of a user of the client device, and may receive further input that modifies the predicted textual segment to an alternate textual segment. Further, the on-device processor(s) may store these textual segments in on-device storage as a candidate correction pair, and transmit the candidate correction pair to a remote system. Moreover, remote processor(s) of the remote system may determine that the candidate correction pair is an actual correction pair, and may cause client devices to generate updates for a global ASR model for the candidate correction pair. Additionally, the remote processor(s) may distribute the global ASR model to the client devices and/or additional client devices.

Abstract: Processor(s) of a client device can: receive sensor data that captures environmental attributes of an environment of the client device; process the sensor data using a machine learning model to generate a predicted output that dictates whether one or more currently dormant automated assistant functions are activated; making a decision as to whether to trigger the one or more currently dormant automated assistant functions; subsequent to making the decision, determining that the decision was incorrect; and in response to determining that the determination was incorrect, generating a gradient based on comparing the predicted output to ground truth output. In some implementations, the generated gradient is used, by processor(s) of the client device, to update weights of the on-device speech recognition model. In some implementations, the generated gradient is additionally or alternatively transmitted to a remote system for use in remote updating of global weights of a global speech recognition model.

Abstract: Implementations described herein are directed to a framework for decentralized learning of large global machine learning (ML) model(s). In various implementations, remote processor(s) of a remote system can identify a global ML model, select client devices to participate in a given round of decentralized learning of the global ML model, and transmit, to each of the client devices, a processed version of the global ML model that is of a reduced transferrable size. Further, client device processor(s) of a client device can receive the processed version of the global ML model, obtain corresponding client data, perform partial model training, based on processing the corresponding client data, for the processed version of the global ML model to generate a corresponding update, and transmit the corresponding update back to the remote system. Moreover, the remote processor(s) can update, based on at least the corresponding update, the global ML model.