Abstract: Aspects of the disclosure relate to determining relevant content in response to a request for information. One or more computing devices (170) may load data elements into registers (385A-385B), wherein each register is associated with at least one parallel processor in a group of parallel processors (380A-380B). For each of the parallel processors, the data elements loaded in its associated registers may be sorted, in parallel, in descending order. The sorted data elements, for each of the parallel processors, may be merged with the sorted data elements of other processors in the group. The merged and sorted data elements may be transposed and stored.

Abstract: A transformer-transducer model for unifying streaming and non-streaming speech recognition includes an audio encoder, a label encoder, and a joint network. The audio encoder receives a sequence of acoustic frames, and generates, at each of a plurality of time steps, a higher order feature representation for a corresponding acoustic frame. The label encoder receives a sequence of non-blank symbols output by a final softmax layer, and generates, at each of the plurality of time steps, a dense representation. The joint network receives the higher order feature representation and the dense representation at each of the plurality of time steps, and generates a probability distribution over possible speech recognition hypothesis. The audio encoder of the model further includes a neural network having an initial stack of transformer layers trained with zero look ahead audio context, and a final stack of transformer layers trained with a variable look ahead audio context.

Abstract: Aspects of the present disclosure are directed to a computer-implemented method including receiving, by a user computing device, data that describes a destination for the media content item. Example destinations can include a location of a recipient of message including the media content item and a digital location (e.g., website, social networking page, etc.). The method can include selecting, by a computing system comprising the user computing device, one or more media content items based on the data that describes the destination for the media content item. Media content items that are more relevant and/or appropriate can be selected by considering the destination of the media content item. The selected media content item(s) can be provided for display by the user computing device in a dynamic keyboard interface.

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: 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: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing requests for digital components. A method includes receiving a request from a client device for a digital component generated by a first SDK installed at a client device, examining data fields of the request by servers, generating real-time requests by the servers, and transmitting each real-time request to a third-party digital component provider by the servers. The method further includes receiving responses to the real-time requests from the third party digital component provider by the servers, selecting a particular response to transmit to the client device responsive to the request for the digital component by the servers, and transmitting the particular response to the first SDK with instructions specifying which SDK installed at the client device is required to render a digital component included in the particular response by the servers.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining hotword suitability. In one aspect, a method includes receiving speech data that encodes a candidate hotword spoken by a user, evaluating the speech data or a transcription of the candidate hotword, using one or more predetermined criteria, generating a hotword suitability score for the candidate hotword based on evaluating the speech data or a transcription of the candidate hotword, using one or more predetermined criteria, and providing a representation of the hotword suitability score for display to the user.

Abstract: This document describes methods, devices, systems, and means for a switching scheme for opting in and out of multi-user orthogonal frequency-division multiple access (MU-OFDMA). In one aspect, an electronic device enters the MU-OFDMA mode to communicate via a wireless network over a shared-channel bandwidth. During the MU-OFDMA mode, the electronic device determines that an uplink-queue size is greater than a first threshold size. Responsive to the determining, the electronic device opts out of the MU-OFDMA mode and enters a single-user mode to contend for a transmit channel for transmitting uplink data.

Abstract: A computerized method, system for, and computer-readable medium operable to: retrieve program details data about at least one media program. Retrieve social media posts involving the at least one media program. Match the retrieved program details data with the retrieved social media posts to find matching locations. Store time offset data corresponding to the matched locations within the at least one media program into at least one micropost. Rank the at least one media program by the number of microposts corresponding to the at least one media program. Present the at least one media program, each of the at least one media program with at least one micropost, wherein the at least one micropost comprises a link to the original social media page hosting the social media post and a video deeplink that plays a time off-set video, the time off-set video starting from the stored time offset data.

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: Geographic data corresponding to a certain geographic area of offline coverage is stored in a memory of a computing device. An indication of a geographic boundary delimiting a region smaller than, and contained within, the geographic area is obtained. A current location of the computing device relative to the geographic boundary is determined and, in response to determining that the current location is outside the region, additional geographic data is retrieved via a wireless communication network to expand the geographic area of offline coverage. At least some of the geographic data corresponding to the expanded geographic area is provided via a user interface of the computing device when the computing device is offline.

Abstract: A method for processing a cryptographic operation request includes receiving, at a hardware security module (HSM), the cryptographic operation request including a cryptographic key and at least one authorization token, determining, by the HSM, whether an access control list (ACL) associated with the cryptographic key of the cryptographic operation request is authorized to govern access to the cryptographic key, and validating, by the HSM, the at least one authorization token. When the at least one authorization token is valid and the ACL is authorized to govern access to the cryptographic key of the cryptographic operation request, the method includes processing, by the HSM, the cryptographic operation request.

Abstract: A control system includes an energy management system in operation with intelligent, network-connected thermostats located in structures. The thermostats are operable to control heating, ventilation, and air conditioning (HVAC) systems. Control during a demand response (DR) event period may be performed based on an optimal control trajectory of the HVAC system, where the control trajectory is optimal in that it minimizes a cost function.

Abstract: A process identifies a search query spike from queries submitted by users during a first span of time, which is less than a predefined duration. The spike corresponds to a set of queries identified as equivalent. The frequency of submitting queries from the set during the first time span exceeds the frequency of submitting queries from the set during an average span of time. The process correlates the spike to a broadcast video program by matching terms from the set of search queries to subtitle terms appearing in the video program at a first location. The first location in the video program was broadcast within a predefined time before the first span of time. The process receives notification from a user device indicating user interest in the video program. The process transmits to the user device search results corresponding to some search queries from the set of search queries.

Abstract: The present disclosure describes systems and methods directed to a user equipment (UE) limited-service mode for wireless communications. For a user equipment (UE) that is wirelessly communicating with a base station, a service-mode manager determines that a thermal, power, or battery condition local to the UE violates a threshold and causes the UE to transmit a message that indicates a request by the UE to enter a UE limited-service mode. The base station allocates a set of resources of the air interface to be used for wireless communications upon the UE entering the UE limited-service mode. The base station then transmits a message to the UE, directing the UE to enter the UE limited-service mode and wirelessly communicate with the base station using the allocated set of resources of the air interface.

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: 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: Systems and methods for facilitating cross device resumable media content consumption. In an aspect, a system includes an update component that receives a media statistics update where the media statistics update includes information identifying an account, media content identification information, and a first playback time position associated with media content. A monitoring component is also included that identifies when the account accesses the media content associated with the media content identification information. A resume component is also included that plays back the media content at the first playback time position in response to identifying an access of the media content associated with the account.

Abstract: Particular techniques for generating synthetic images and/or for training machine learning model(s) based on the generated synthetic images. For example, training a machine learning model based on training instances that each include a generated synthetic image, and ground truth label(s) for the generated synthetic image. After training of the machine learning model is complete, the trained machine learning model can be deployed on one or more robots and/or one or more computing devices.