Abstract: The various implementations described herein include methods and systems for determining device leadership among voice interface devices. In one aspect, a method is performed at a first electronic device of a plurality of electronic devices, each having microphones, a speaker, processors, and memory storing programs for execution by the processors. The first device detects a voice input. It determines a device state and a relevance of the voice input. It identifies a subset of electronic devices from the plurality to which the voice input is relevant. In accordance with a determination that the subset includes the first device, the first device determines a first score of a criterion associated with the voice input and receives second scores of the criterion from other devices in the subset. In accordance with a determination that the first score is higher than the second scores, the first device responds to the detected input.

Abstract: A method for rehearsal-free continual learning includes obtaining a set of training samples where training sample in the set of training samples is associated with a respective task of a plurality of different tasks. The method includes obtaining a task-invariant prompt representative of learned knowledge common to each respective task of the plurality of different tasks. The method includes, for each respective task of the plurality of different tasks, obtaining a respective task-specific prompt representative of learned knowledge specific to the respective task. The method includes, during each of one or more training iterations, for each respective training sample in the set of training samples, selecting the respective task-specific prompt representative of the respective task of the respective training sample and training a model using the task-invariant prompt and the selected respective task-specific prompt.

Abstract: A method for forecasting time-series data, when executed by data processing hardware, causes the data processing hardware to perform operations including receiving a time series forecasting query from a user requesting a time series forecast forecasting future data based on a set of current time-series data. The operations include obtaining, from the set of current time-series data, a set of training data. The operations include training, using a first portion of the set of training data, a first sub-model of a forecasting model and training, using a second portion of the set of training data, a second sub-model of the forecasting model. The second portion is different than the first portion. The operations include forecasting, using the forecasting model, the future data based on the set of current time-series data and returning, to the user, the forecasted future data for the time series forecast.

Abstract: Methods, systems, and computer-readable media for multi-model processing on resource-constrained devices. A resource-constrained device can determine, based on a battery-life for a battery of the device, whether to process input through a first model or a second model. The first model can be a gating model that is more energy efficient to execute, and the second model can be a main model that is more accurate than the gating model. Depending on the current battery-life and/or other criteria, the system can process, through the gating model, sensor input that can record activity performed by a user of the resource-constrained device. If the gating model predicts an activity performed by the user that is recorded by the sensor data, the device can process the same or additional input through the main model. Overall power consumption can be reduced with a minimum accuracy maintained over processing input only through the main model.

Abstract: A method for load balancing application requests across a multi-cluster container orchestration system includes receiving a load-balancing configuration for a multi-cluster service managing access to a set of destination clusters hosting a software application deployed by a user. The multi-cluster service uses the load-balancing configuration to load balance application level traffic across the set of destination clusters. Each destination cluster includes at least one container executing the software application and a respective geographical region. The method also includes receiving an application level request directed toward the software application hosted across the destination clusters. The application level request is received from a client and includes a host name and a geographical location associated with the client. The method also includes routing the request to one of the clusters based on the geographical location of the request and the respective geographical locations of the clusters.

Abstract: A method for training a neural network includes receiving a training input audio sequence including a sequence of input frames defining a hotword that initiates a wake-up process on a user device. The method further includes obtaining a first label and a second label for the training input audio sequence. The method includes generating, using a memorized neural network and the training input audio sequence, an output indicating a likelihood the training input audio sequence includes the hotword. The method further includes determining a first loss based on the first label and the output. The method includes determining a second loss based on the second label and the output. The method further includes optimizing the memorized neural network based on the first loss and the second loss associated with the training input audio sequence.

Abstract: A method for instantaneous learning in text-to-speech (TTS) during dialog includes receiving a user pronunciation of a particular word present in a query spoken by a user. The method also includes receiving a TTS pronunciation of the same particular word that is present in a TTS input where the TTS pronunciation of the particular word is different than the user pronunciation of the particular word. The method also includes obtaining user pronunciation-related features and TTS pronunciation related features associated with the particular word. The method also includes generating a pronunciation decision selecting one of the user pronunciation or the TTS pronunciation of the particular word that is associated with a highest confidence. The method also include providing the TTS audio that includes a synthesized speech representation of the response to the query using the user pronunciation or the TTS pronunciation for the particular word.

Abstract: Techniques are described herein that enable advanced gaming and virtual reality control using radar. These techniques enable small motions and displacements to be tracked, even in the millimeter or submillimeter scale, for user control actions even when those actions are optically occluded or obscured.

Abstract: This document describes methods and devices for a handover of a user equipment (110) from a Fifth Generation (5G) New Radio (NR) base station (121) to an Evolved Universal Terrestrial Radio Access (E-UTRA) base station (122). The user equipment (110) may determine to release or keep a Packet Data Convergence Protocol (PDCP) entity (312) depending on the whether the E-UTRA base station (122) is connected to an Evolved Packet Core (160) network or a Fifth Generation Core (150) network. The user equipment (110) may determine to release or keep the PDCP entity (312) depending based on a received NR Radio Resource Control (RRC) message or E-UTRA RRC message.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for speech recognition. One of the methods includes receiving first audio data corresponding to an utterance; obtaining a first transcription of the first audio data; receiving data indicating (i) a selection of one or more terms of the first transcription and (ii) one or more of replacement terms; determining that one or more of the replacement terms are classified as a correction of one or more of the selected terms; in response to determining that the one or more of the replacement terms are classified as a correction of the one or more of the selected terms, obtaining a first portion of the first audio data that corresponds to one or more terms of the first transcription; and using the first portion of the first audio data that is associated with the one or more terms of the first transcription to train an acoustic model for recognizing the one or more of the replacement terms.

Abstract: A method for arbitrarily grouping computer system resources includes receiving a resource policy to be applied to a group of resources associated with a user. The resource policy includes a unique user-defined label assigned to each resource of the group of resources. The method also includes identifying the group of resources among a plurality of resources using the user-defined label. Here, the plurality of resources is associated with the user and distributed across a plurality of different computing environments. For each identified resource of the group of resources, the method also includes applying the resource policy to a corresponding identified resource by accessing the respective computing environment where the corresponding identified resource resides.

Abstract: Apparatus for quantum error correction is disclosed. The apparatus includes an array of processing cores, each processing core comprising: a processor on a first chip; and a processor cache on the first chip; and a bus for interconnecting neighbouring processing cores in the array of processing cores; wherein each processing core includes: control code which, when executed by the processor, causes the processor to access a processor cache of at least one neighbouring processing core.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for predicting future patient health using neural networks. One of the methods includes receiving electronic health record data for a patient; generating a respective observation embedding for each of the observations, comprising, for each clinical note: processing the sequence of tokens in the clinical note using a clinical note embedding LSTM to generate a respective token embedding for each of the tokens; and generating the observation embedding for the clinical note from the token embeddings; generating an embedded representation, comprising, for each time window: combining the observation embeddings of observations occurring during the time window to generate a patient record embedding; and processing the embedded representation of the electronic health record data using a prediction recurrent neural network to generate a neural network output that characterizes a future health status of the patient.

Abstract: A method includes presenting, in a user interface of an electronic calendar application, a plurality of calendar events and a plurality of calendar event categories, wherein each calendar event of the plurality of calendar events is associated with one or more of the plurality of categories; receiving, via the user interface of the electronic calendar application, a user selection of a first calendar event category of the plurality of calendar event categories; identifying among the plurality of calendar events, one or more calendar events associated with the first calendar event category; and adjusting an appearance of each the identified calendar events associated with the first calendar event category to distinguish the identified calendar events from other calendar events presented in the user interface of the electronic calendar application.

Abstract: Technology for characterizing internet application performance is described. An example method may involve, analyzing client requests from a plurality of internet service providers, the client requests comprising requests to access media items and being associated with internet service provider identities; identifying a set of the client requests that are associated with a first internet service provider; determining for at least one client request of the set: a request time, a payload data quantity, and an elapsed time; calculating a transfer rate comprising an application level throughput for the at least one client request, the transfer rate being based on the payload data quantity and the elapsed time; and calculating a performance measurement of the first and second internet service providers, the performance measurements being based on the transfer rate and on a portion of the set of client requests that have a transfer rate meeting a transfer threshold.

Abstract: This document describes techniques and systems that enable a mobile device-based radar system for applying different power modes to a multi-mode interface. The techniques and systems include a user device having a radar system, and an interaction manager. The radar system generates a radar field, provides radar data, and operates at one of various different radar-power states. The user device analyzes the radar data to detect a presence or movement of a user within the radar field. Responsive to the detection, the radar system changes from a first radar-power state to a second radar-power state. Based on this change, the interaction manager selects a power mode, for a multi-mode interface, that corresponds to the second radar-power state, and applies the selected power mode to the multi-mode interface to provide a corresponding display via a display device.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for suppressing search results to personally objectionable content. One of the methods includes receiving an identifier of a resource that has image content. A first classifier classifies the image content as including objectionable content or not including objectionable content. A second classifier classifies the image content as including professionally produced content or not including professionally produced content. Whenever the image content is classified as including objectionable content and as not including professionally produced content, the resource is designated as having personally objectionable content.

Abstract: The disclosed subject matter relates generally to display technology, including but not limited to methods and systems for providing knowledge-based media content recommendations, displaying a unified user interface with content-based media recommendations, mixing multiple media content items to be displayed on a focus area, and/or casting controlled media content associated with a network-connected television device. More particularly, the disclosed subject matter relates to displaying a unified user interface having tabbed interfaces in which upcoming media content items from multiple content providers can be displayed in one of the tabbed interfaces.

Abstract: The various embodiments described herein include methods, devices, and systems for power-management on camera devices. In one aspect, a method is performed at a camera device having memory, one or more processors, and an image sensor. The method includes: (1) while a wireless communication component of the camera device is deactivated: (a) capturing a plurality of images containing a motion event; (b) characterizing the motion event; and (c) determining, based on the characterization of the motion event, whether to send video data to a remote computing system; and (2) in accordance with a determination to send video data to the remote computing system: (i) activating the wireless communication component of the camera device; (ii) establishing a wireless connection to the remote computing system via the wireless communication component; and (iii) sending video information to the remote computing system via the established wireless connection.

Abstract: A method of training a speech model includes receiving, at a voice-enabled device, a fixed set of training utterances where each training utterance in the fixed set of training utterances includes a transcription paired with a speech representation of the corresponding training utterance. The method also includes sampling noisy audio data from an environment of the voice-enabled device. For each training utterance in the fixed set of training utterances, the method further includes augmenting, using the noisy audio data sampled from the environment of the voice-enabled device, the speech representation of the corresponding training utterance to generate noisy audio samples and pairing each of the noisy audio samples with the corresponding transcription of the corresponding training utterance. The method additionally includes training a speech model on the noisy audio samples generated for each speech representation in the fixed set of training utterances.