Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for collaboration between multiple voice controlled devices are disclosed. In one aspect, a method includes the actions of identifying, by a first computing device, a second computing device that is configured to respond to a particular, predefined hotword; receiving audio data that corresponds to an utterance; receiving a transcription of additional audio data outputted by the second computing device in response to the utterance; based on the transcription of the additional audio data and based on the utterance, generating a transcription that corresponds to a response to the additional audio data; and providing, for output, the transcription that corresponds to the response.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for audio-visual speech separation. A method includes: obtaining, for each frame in a stream of frames from a video in which faces of one or more speakers have been detected, a respective per-frame face embedding of the face of each speaker; processing, for each speaker, the per-frame face embeddings of the face of the speaker to generate visual features for the face of the speaker; obtaining a spectrogram of an audio soundtrack for the video; processing the spectrogram to generate an audio embedding for the audio soundtrack; combining the visual features for the one or more speakers and the audio embedding for the audio soundtrack to generate an audio-visual embedding for the video; determining a respective spectrogram mask for each of the one or more speakers; and determining a respective isolated speech spectrogram for each speaker.

Abstract: Implementations set forth herein relate to rendering, at a message application, certain messages with a particular sorting scheme (e.g., a classification-based sorting scheme) based on the certain messages having transfer times that satisfy a transfer time threshold and, optionally, based on the certain messages being unread. Other messages rendered by the message application can be rendered with an alternative sorting scheme, such as chronologically. Utilization of the particular sorting scheme for the certain messages enables a user, viewing the messages, to review (e.g., reply-to, delete, and/or view) the certain messages with reduced latency, thereby allowing the certain messages to be more quickly reviewed and actioned when received.

Abstract: An electronic device comprises an image sensor, one or more processors, and memory storing instructions for receiving an event recording profile based on configuration data of the electronic device, the configuration data including a location type or a power type; receiving a plurality of images of a scene captured by the image sensor; detecting a trigger event based on one or more of the plurality of images of the scene; in response to detecting the trigger event, identifying an object of interest in one or more of the plurality of images of the scene; creating an event clip from the stored images that include the object of interest, wherein creating the event clip includes configuring a clip length based on the event recording profile; and providing the event clip for display.

Abstract: In aspects, a visitor interaction system is described for interacting with visitors. The visitor interaction system can obtain video data, identify an approach of a person, and determine an identity of the person. Further, the visitor interaction system can determine that the person belongs to a respective visitor group of a plurality of visitor groups. Each of the plurality of visitor groups corresponds to a response model that includes a plurality of response actions. At least one of the response actions includes an autonomous response action. The visitor interaction system can also identify a response model corresponding to the respective visitor group and initiate an autonomous response action associated with the response model. The visitor interaction system can then send a report message to the user via a client device.

Abstract: A method includes determining, by an assistant executing at one or more processors, a default group of actions that the assistant is configured to execute in response to receiving a particular audible command. The method includes determining, by the assistant, based on the default group of actions and a user profile associated with a particular user, a custom group of actions that the assistant is configured to execute in response to receiving the particular audible command from the particular user. The method also includes receiving, by the assistant, an indication of the particular audible command, and determining, by the assistant, whether the indication of particular audible command originated from the particular user. The method further includes, responsive to determining that the indication of particular audible command originated from the particular user, executing, by the assistant, each action from the custom group of actions.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating an output sequence from an input sequence. In one aspect, one of the systems includes an encoder neural network configured to receive the input sequence and generate encoded representations of the network inputs, the encoder neural network comprising a sequence of one or more encoder subnetworks, each encoder subnetwork configured to receive a respective encoder subnetwork input for each of the input positions and to generate a respective subnetwork output for each of the input positions, and each encoder subnetwork comprising: an encoder self-attention sub-layer that is configured to receive the subnetwork input for each of the input positions and, for each particular input position in the input order: apply an attention mechanism over the encoder subnetwork inputs using one or more queries derived from the encoder subnetwork input at the particular input position.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a fundus image processing machine learning models that is configured to process one or more fundus images captured by a fundus camera to generate a predicted label. One of the methods includes generating training data, comprising: receiving sets of one or more training fundus images captured by a fundus camera; receiving, for each of the sets, a ground truth label assigned to a different image of the eye of the patient corresponding to the set that has been captured using a different imaging modality; and generating, for each set of training fundus images, a training example that includes the set of training fundus images in association with the ground truth label assigned to the different image of the patients eye; and training the machine learning model on the training examples in the training data.

Abstract: The present disclosure describes systems and methods for a user equipment wirelessly communicating with another user equipment using dual connectivity (DC) with a terrestrial base station and a satellite or high-altitude platform. The described methods and systems include a principal routing manager assessing that different subsets of data, to be transmitted from the user equipment to the other user equipment, can use different, respective qualities of service (QoS) offered through different wireless-communication networks associated with the terrestrial base station and the satellite or high-altitude platform.

Abstract: The present disclosure provides systems, methods, and computer program products for providing efficient embedding table storage and lookup in machine-learning models.

Abstract: Methods, systems and apparatus for correcting a result of a quantum computation.

Abstract: Systems and methods are provided for including information in an invitation from a first to connect with a second user that identifies a context regarding how the first user found the second user. A system can include a request component configured to receive, at a first network source, information identifying a second user that a first user has selected to send an invitation to connect with at the first network source, a context component configured to determine a context regarding selection of the second user by the first user, including where the first user found the information identifying the second user, and a generation component configured to generate an invitation that invites the second user to connect with the first user at the first network source. The invitation comprises first user identification information and context information identifying the context regarding the selection of the second user by the first user.

Abstract: Methods, systems, devices, and computer-readable storage media for activity detection of a user of a computing device, using multi-modal sensing. A device can be configured to receive sensor data corresponding to multiple modalities and process the sensor data to predict an activity performed by a user of a computing device. The device in response to the detected activity can perform a response action, such as muting or pausing audio playback from the computing device. Different modalities can be combined, such as body vibration data, air vibration data, and image data, which can be processed to distinguish user activity, e.g., speaking versus not speaking, to allow the computing device to perform the correct corresponding action.

Abstract: Implementations set forth herein relate to management of casting requests and user inputs at a rechargeable device, which provides access to an automated assistant and is capable of rendering data that is cast from a separate device. Casting requests can be handled by the rechargeable device despite a device SoC of the rechargeable device operating in a sleep mode. Furthermore, spoken utterances provided by a user for invoking the automated assistant can also be adaptively managed by the rechargeable device in order mitigate idle power consumption by the device SoC. Such spoken utterances can be initially processed by a digital signal processor (DSP), and, based on one or more features (e.g., voice characteristic, conformity to a particular invocation phrase, etc.) of the spoken utterance, the device SoC can be initialized for an amount of time that is selected based on the features of the spoken utterance.

Abstract: Methods and apparatus for generating waveforms for application in a quantum computing device. In one aspect, a system comprises a batch generator that receives experiment data sets defining respective experiments, each experiment data set comprising a set of waveforms defined by respective waveform data; determines unique waveforms; generates a corresponding set of respective waveform data that includes the respective waveform data for each unique waveform; generates, for each of the experiments, a waveform list that references the respective waveform data in the set of respective waveform data that corresponds to the waveforms in the set; and batch instructions that are executable by waveform generator hardware and that cause the waveform generator hardware to process each waveform list by selecting each referenced waveform data in the waveform list; and generate, in response to the selected waveform data, a waveform that is suitable for application in a quantum computing device.

Abstract: Implementations are described herein for leveraging machine learning to automate source code refactoring and/or rearchitecting. In various implementations, one or more ground truth boundaries may be removed from one or more boundaried source code files to produce one or more boundary-less source code files. One or more of the boundary-less source code files may be processed using a machine learning model to predict one or more candidate boundaries for reintroduction into the one or more boundary-less source code files. The one or more ground truth boundaries may be compared with the one or more predicted candidate boundaries. The machine learning model may be trained based on the comparing.

Abstract: A method of postponing an electronic message is performed at a computing device having one or more processors and memory. The memory stores one or more programs configured for execution by the one or more processors. The device displays a first electronic message for a user of the computing device and receives user input to postpone the first electronic message. In response to the user input, the device displays a plurality of postponement locations in a user interface window and receives user selection of a postponement location from the displayed plurality of postponement locations. The device then inhibits display of the first electronic message at the computing device. When it is subsequently determined that the computing device is at the first postponement location, the device redisplays the first electronic message.

Abstract: In response to a user interacting with a tangible peripheral assistant control device (e.g., depressing a button of the device), causing an automated assistant to perform one or more actions. The action(s) performed can be based on input previously provided by the user in configuring the peripheral assistant control device. The action(s) performed in response to interaction with the peripheral assistant control device can vary based on one or more conditions, such as which user is currently active, where the peripheral assistant control device is currently located (which can optionally be inferred based on which of multiple assistant computing devices the button is paired with), and/or the current state of one or more smart devices and/or other devices (e.g., as determined based on a device topology). A utility of the peripheral assistant control device can be automatically extended beyond what was specifically requested by a user during configuration.

Abstract: A method for accessing one or more service processes of service includes executing at least one service enclave and executing an enclave sandbox that wraps the at least one service enclave. The at least one service enclave provides an interface to the one or more service processes. The enclave sandbox is configured to establish an encrypted communication tunnel to the at least one service enclave interfacing with the one or more service processes, and communicate program calls to/from the one or more service processes as encrypted communications through the encrypted communication tunnel.