Abstract: Generally, the present disclosure is directed to enhanced federated learning (FL) that employs a set of clients with varying amounts of computational resources (e.g., system memory, storage, and processing bandwidth). To overcome limitations of conventional FL methods that employ a set of clients with varying amounts of computational resources, the embodiments run multi-directional knowledge distillation between the server models produced by each federated averaging (FedAvg) pool, using unlabeled server data as the distillation dataset. By co-distilling the two (or more) models frequently over the course of FedAvg rounds, information is shared between the pools without sharing model parameters. This leads to increased performance and faster convergence (in fewer federated rounds).

Abstract: A method includes receiving distillation data including a plurality of out-of-domain training utterances. For each particular out-of-domain training utterance of the distillation data, the method includes generating a corresponding augmented out-of-domain training utterance, and generating, using a teacher ASR model trained on training data corresponding to a target domain, a pseudo-label corresponding to the corresponding augmented out-of-domain training utterance. The method also includes distilling a student ASR model from the teacher ASR model by training the student ASR model using the corresponding augmented out-of-domain training utterances paired with the corresponding pseudo-labels generated by the teacher ASR model.

Abstract: A method of phrase extraction for ASR models includes obtaining audio data characterizing an utterance and a corresponding ground-truth transcription of the utterance and modifying the audio data to obfuscate a particular phrase recited in the utterance. The method also includes processing, using a trained ASR model, the modified audio data to generate a predicted transcription of the utterance, and determining whether the predicted transcription includes the particular phrase by comparing the predicted transcription of the utterance to the ground-truth transcription of the utterance. When the predicted transcription includes the particular phrase, the method includes generating an output indicating that the trained ASR model leaked the particular phrase from a training data set used to train the ASR model.

Abstract: Information can be distilled from a global automatic speech recognition (ASR) model to a client ASR model. Many implementations include using an RNN-T model as the ASR model, where the global ASR model includes a global encoder, a joint network, a prediction network, and where the client ASR model includes a client encoder, the joint network, and the prediction network. Various implementations include using principal component analysis (PCA) while training the global ASR model to learn a mean vector and a set of principal components corresponding to the global ASR model. Additional or alternative implementations include training the client ASR model to generate one or more predicted coefficients of the global ASR model.

Abstract: A method includes receiving distillation data including a plurality of out-of-domain training utterances. For each particular out-of-domain training utterance of the distillation data, the method includes generating a corresponding augmented out-of-domain training utterance, and generating, using a teacher ASR model trained on training data corresponding to a target domain, a pseudo-label corresponding to the corresponding augmented out-of-domain training utterance. The method also includes distilling a student ASR model from the teacher ASR model by training the student ASR model using the corresponding augmented out-of-domain training utterances paired with the corresponding pseudo-labels generated by the teacher ASR model.

Abstract: A method of phrase extraction for ASR models includes obtaining audio data characterizing an utterance and a corresponding ground-truth transcription of the utterance and modifying the audio data to obfuscate a particular phrase recited in the utterance. The method also includes processing, using a trained ASR model, the modified audio data to generate a predicted transcription of the utterance, and determining whether the predicted transcription includes the particular phrase by comparing the predicted transcription of the utterance to the ground-truth transcription of the utterance. When the predicted transcription includes the particular phrase, the method includes generating an output indicating that the trained ASR model leaked the particular phrase from a training data set used to train the ASR model.

Abstract: During a round of decentralized learning for updating of a global machine learning (ML) model, remote processor(s) of a remote system may transmit, to a population of computing devices, primary weights for a primary version of the global ML model, and cause each of the computing devices to generate a corresponding update for the primary version of the global ML model. Further, the remote processor(s) may cause the primary version of the global ML model to be updated based on the corresponding updates that are received during the round of decentralized learning. However, the remote processor(s) may receive other corresponding updates subsequent to the round of decentralized learning. Accordingly, various techniques described herein (e.g., FARe-DUST, FeAST on MSG, and/or other techniques) enable the other corresponding updates to be utilized in achieving a final version of the global ML model.

Abstract: Implementations disclosed herein are directed to a hybrid federated learning (FL) technique that utilizes both federated averaging (FA) and federated distillation (FD) during a given round of FL of a given global machine learning (ML) model. Implementations may identify a population of client devices to participate in the given round of FL, determine a corresponding quantity of instances of client data available at each of the client devices that may be utilized during the given round of FL, and select different subsets of the client devices based on the corresponding quantity of instances of client data. Further, implementations may cause a first subset of the client devices to generate a corresponding FA update and a second subset of client devices to generate a corresponding FD update. Moreover, implementations may subsequently update the given global ML model based on the corresponding FA updates and the corresponding FD updates.

Abstract: A method of phrase extraction for ASR models includes obtaining audio data characterizing an utterance and a corresponding ground-truth transcription of the utterance and modifying the audio data to obfuscate a particular phrase recited in the utterance. The method also includes processing, using a trained ASR model, the modified audio data to generate a predicted transcription of the utterance, and determining whether the predicted transcription includes the particular phrase by comparing the predicted transcription of the utterance to the ground-truth transcription of the utterance. When the predicted transcription includes the particular phrase, the method includes generating an output indicating that the trained ASR model leaked the particular phrase from a training data set used to train the ASR model.

Abstract: A method include obtaining a set of differentially private (DP) gradients each generated based on processing corresponding private data, and obtaining a set of public gradients each generated based on processing corresponding public data. The method also includes applying mirror descent to the set of public gradients to learn a geometry for the set of DP gradients, and reshaping the set of DP gradients based on the learned geometry. The method further includes training a machine learning model based on the reshaped set of DP gradients.

Abstract: A method includes receiving training data for a machine learning model, the training data comprising a plurality of training examples and a corresponding plurality of labels. The method further includes dividing the training data into a plurality of training batches. For each training batch of the plurality of training batches, the method additionally includes learning a weight for each training example in the training batch that minimizes a sum of weighted losses for the training batch subject to a divergence constraint, where the divergence constraint limits a divergence of the learned weights for the training batch from a reference distribution, where the divergence is determined according to a chosen divergence measure. The method also includes training the machine learning model with each training batch of the plurality of training batches using the learned weight for each training example in the training batch. The method additionally includes providing the trained machine learning model.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network using local layer-wise losses.

Abstract: The present disclosure relates to a triplet embedding system that improves dimensionality reduction through exponential triplet embedding. In particular, the triplet embedding system employs heavy-tailed properties of t-exponential distributions and robust non-convex loss functions to improve visualizations in the presence of noisy data. In addition, the triplet embedding system uses triplet similarity weighting and improved sampling to improve and accelerate triplet embedding in large datasets. Overall, the triplet embedding system produces improved dimensionality reduction visualizations, which accurately reveal the underlying structure of the real-world high-dimensional datasets in lower-dimensional space.

Abstract: The present disclosure relates to a triplet embedding system that improves dimensionality reduction through exponential triplet embedding. In particular, the triplet embedding system employs heavy-tailed properties of t-exponential distributions and robust non-convex loss functions to improve visualizations in the presence of noisy data. In addition, the triplet embedding system uses triplet similarity weighting and improved sampling to improve and accelerate triplet embedding in large datasets. Overall, the triplet embedding system produces improved dimensionality reduction visualizations, which accurately reveal the underlying structure of the real-world high-dimensional datasets in lower-dimensional space.