Abstract: Embodiments are disclosed for providing an interface with a personalized virtual personal assistant (VPA) via a computing system. The example method comprises assigning a plurality of virtual personal assistant (VPA) instances to a plurality of users, each VPA instance of the plurality of VPA instances operating concurrently. For example, assigning the plurality of VPA instances to the plurality of users may include retrieving a personalized VPA configuration for each user of the plurality of users based on a plurality of audio samples, each of the plurality of audio samples corresponding to a user of the plurality of users.

Abstract: Disclosed embodiments include techniques for processing audio commands for a virtual personal assistant. A speech processing system segments speech detected in an audio input signal into a plurality of speech segments wherein the speech included in a given speech segment is uttered by a single speaker. The speech processing system clusters the speech segments included in the plurality of speech segments into a plurality of clusters, wherein each cluster included in the plurality of clusters is uttered by a different speaker. The speech processing system determines that a first cluster included in the plurality of clusters is uttered by a first authorized speaker. The speech processing system causes the virtual personal assistant to execute a first audio command included in the first cluster.

Abstract: Embodiments are disclosed for providing an interface with a personalized virtual personal assistant (VPA) via a computing system. The example method comprises assigning a plurality of virtual personal assistant (VPA) instances to a plurality of users, each VPA instance of the plurality of VPA instances operating concurrently. For example, assigning the plurality of VPA instances to the plurality of users may include retrieving a personalized VPA configuration for each user of the plurality of users based on a plurality of audio samples, each of the plurality of audio samples corresponding to a user of the plurality of users.

Abstract: Training a user-specific perturbation generator for an audio feature detection model includes receiving one or more positive audio samples of a user, each of the one or more positive audio samples including an audio feature; receiving one or more negative audio samples of the user, each of the one or more negative audio samples sharing an acoustic similarity with at least one of the one or more positive audio samples; and adversarially training a user-specific perturbation generator model to generate a user-specific perturbation, the training based on the one or more positive audio samples and the one or more negative audio samples. Perturbing audio samples of the user with the user-specific perturbation can cause an audio feature detection model to recognize the audio feature in audio samples that include the audio feature and/or to refrain from recognizing the audio feature in audio samples that do not include the audio feature.

Abstract: The current disclosure relates to systems and methods for wakeword or keyword detection in Virtual Personal Assistants (VPAs). In particular, systems and methods are provided for wakeword detection using deep neural networks including a parametric pooling layer, wherein the parametric pooling layer includes trainable parameters, enabling the layer to learn to distinguish between informative feature vectors and non-informative/noisy feature vectors extracted from a variable length acoustic signal. In one example, a parametric pooling layer may aggregate a variable length feature map, comprising a plurality of feature vectors extracted from an acoustic signal, into an embedding vector of pre-determined length, by weighting each of the plurality of feature vectors based on one or more learned parameters in a parametric pooling layer, and aggregating the plurality of weighted feature vectors into the embedding vector.