Abstract: A computer-implemented method of multisensory speech detection is disclosed. The method comprises determining an orientation of a mobile device and determining an operating mode of the mobile device based on the orientation of the mobile device. The method further includes identifying speech detection parameters that specify when speech detection begins or ends based on the determined operating mode and detecting speech from a user of the mobile device based on the speech detection parameters.

Abstract: An apparatus is described. The apparatus includes an image processing unit. The image processing unit includes a plurality of stencil processor circuits each comprising an array of execution unit lanes coupled to a two-dimensional shift register array structure to simultaneously process multiple overlapping stencils through execution of program code. The image processing unit includes a plurality of sheet generators respectively coupled between the plurality of stencil processors and the network. The sheet generators are to parse input line groups of image data into input sheets of image data for processing by the stencil processors, and, to form output line groups of image data from output sheets of image data received from the stencil processors. The image processing unit includes a plurality of line buffer units coupled to the network to pass line groups in a direction from producing stencil processors to consuming stencil processors to implement an overall program flow.

Abstract: Computer-implemented techniques include detecting, using a camera of a computing device having one or more processors, a digital watermark displayed by a display of a computing system. The digital watermark can be a visual indicator that is detectable by the camera of the computing device, and the computing system can further comprise a set of speakers and a set of cameras. The techniques can include determining, by the computing device, a unique identifier for the computing system based on the digital watermark. The techniques can also include automatically coordinating, by the computing device, an audio/video conference session between the computing device and the computing system using the unique identifier.

Abstract: A system can include a mobile computing device and a wearable computing device. The wearable computing device can include a sensor that outputs an indication that the wearable computing device is not being worn. Responsive to receiving the indication that the wearable computing device is being not being worn, one or both of the devices can be operable to change an access mode of computing environment provided by the respective device from an increased access mode to a reduced access mode.

Abstract: Systems and methods for model validation includes generating a first and a second time series of segmentation states for a data set representative of a simulated population, e.g., a collection of membership counts corresponding to respective segments of the simulated population. The first and second time series of segmentation states are generated by respectively processing the data set through a first and a second simulation each comprising iterative application of a plurality of event functions. The first and the second simulation differ in at least one capacity, e.g., one including a first event function configured with a first parameter, and the second not. Analysis of differences between the first and second time series may be compared to analysis of one of the time series using a subject model. The comparison is then used to validate the model or demonstrate accuracies, inaccuracies, and/or model bias with respect to a performance metric.

Abstract: A user interface for rotating imagery among a plurality of canonical views of a geographic area of interest is disclosed. The interface includes at least one rotation control icon for rotating the imagery among a plurality of canonical views of the geographic area, and a direction indicating portion that indicates a compass direction associated with the canonical view presented on the display device. In particular implementations, the interface can be configured to come into view over the imagery whenever a user navigates to a particular view of the imagery where a plurality of canonical views is available. The motion of the navigation control interface appearing into the imagery is intended to draw attention to the control interface to inform the user that additional canonical views are available.

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 a computer storage medium, for performing dynamic, stroke-based alignment of touch displays. In one aspect, a method includes obtaining a candidate transcription that an automated speech recognizer generates for an utterance, determining a particular context associated with the utterance, determining that a particular n-gram that is included in the candidate transcription is included among a set of undesirable n-grams that is associated with the context, adjusting a speech recognition confidence score associated with the transcription based on determining that the particular n-gram that is included in the candidate transcription is included among the set of undesirable n-grams that is associated with the context, and determining whether to provide the candidate transcription for output based at least on the adjusted speech recognition confidence score.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for providing a natural language based program interface to software applications. One of the methods includes, obtaining, via a natural language front end, a natural language query or a natural language update statement issued by a software application; converting the natural language query or natural language update statement into structured operations to be performed on APIs of a knowledge base; performing the structured operations on the APIs to produce a natural language output statement; and providing, via a natural language output interface, the natural language output statement to the software application. The knowledge base stores entity information according to a data schema and has structured APIs for use by software applications to query the knowledge base; the software applications are limited to communicating with the knowledge base through the interfaces provided by the natural language front end.

Abstract: An electronic device including a display panel is provided. In some embodiments, the display panel includes an active display region and an electrical component. The electrical component is distinct from a touch sensor device that is adapted to receive touch input across the active display region of the display panel. An electrical conductor connects the electrical component to another component of the electronic device via the display panel, such as via a bent portion of the display panel.

Abstract: Disclosed herein are systems and methods for a sound enhancement system coupled to a mobile device. In one aspect, the present disclosure includes the system having first bidirectional communication link configured to receive an audio signal, transmit an audio signal, and receive configuration parameters. The system also includes a memory configured to store the configuration parameters. Additionally, the system has an audio input configured to receive ambient audio. Yet further, the is configured to operate in one of two operation modes. In the first mode, the system is configured to receive an audio signal from the bidirectional communication link and output an audio signal based on the received audio signal by way of the audio output pathway, and in the second mode, the system is configured to process the ambient audio signal by the audio processor based on stored configuration parameters and create an output audio signal.

Abstract: Methods, systems, and apparatus, including a system for transforming sparse elements to a dense matrix. The system is configured to receive a request for an output matrix based on sparse elements including sparse elements associated with a first dense matrix and sparse elements associated with a second dense matrix; obtain the sparse elements associated with the first dense matrix fetched by a first group of sparse element access units; obtain the sparse elements associated with the second dense matrix fetched by a second group of sparse element access units; and transform the sparse elements associated with the first dense matrix and the sparse elements associated with the second dense matrix to generate the output dense matrix that includes the sparse elements associated with the first dense matrix and the sparse elements associated with the second dense matrix.

Abstract: The present disclosure is generally directed a data processing system for authenticating packetized audio signals in a voice activated computer network environment. The data processing system can improve the efficiency and effectiveness of auditory data packet transmission over one or more computer networks by, for example, disabling malicious transmissions prior to their transmission across the network. The present solution can also improve computational efficiency by disabling remote computer processes possibly affected by or caused by the malicious audio signal transmissions. By disabling the transmission of malicious audio signals, the system can reduce bandwidth utilization by not transmitting the data packets carrying the malicious audio signal across the networks.

Abstract: A system includes a neural network that includes a Mixture of Experts (MoE) subnetwork between a first neural network layer and a second neural network layer. The MoE subnetwork includes multiple expert neural networks. Each expert neural network is configured to process a first layer output generated by the first neural network layer to generate a respective expert output. The MoE subnetwork further includes a gating subsystem that selects, based on the first layer output, one or more of the expert neural networks and determine a respective weight for each selected expert neural network, provides the first layer output as input to each of the selected expert neural networks, combines the expert outputs generated by the selected expert neural networks in accordance with the weights for the selected expert neural networks to generate an MoE output, and provides the MoE output as input to the second neural network layer.

Abstract: A method for generating an application recommendation includes receiving a signal from an electronic device indicating that uninstallation of a first application has been initiated by a user on the electronic device, obtaining contextual information of the application, the contextual information including data indicating when the application was installed and frequency of use of the application, obtaining utilization data, the utilization data including data indicating applications that have been installed on the electronic device, determining a plurality of applications that are similar to the first application; generating a first predictive model using the contextual information and the utilization data, the first predictive model being configured to predict a likelihood of uninstallation of an application; and automatically recommending, at the time of uninstallation of the application, one or more of the plurality of applications, the recommending being based at least in part on the first predictive model

Abstract: This specification describes technologies relating to inferring interactions with advertisers. In general, one aspect of the subject matter described in this specification can be embodied in methods that include providing an online advertisement to a system, receiving data from the system indicating an interaction of the system with the online advertisement and data indicating a subsequent interaction of the system with an online resource of an advertiser associated with the online advertisement, and inferring that an advertisement conversion has occurred based on the received data. Other embodiments of this aspect include corresponding systems, apparatus, and computer program products.

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: Systems and methods for classifying messages are provided. Each message in a plurality of messages is classified, thereby independently identifying a message category in a set of message categories for each respective message in the plurality. The plurality of messages is delivered to a plurality of recipients with a designation of the message category of each respective message in the first plurality of messages. A plurality of recipient initiated message interaction events for messages in the first plurality of messages over a predetermined period of time is collected from the plurality of recipients. A message categorization dataset is then constructed from (i) the first plurality of messages, (ii) the designation of the message category of each respective message in the subset of the first plurality of messages, and (iii) the plurality of recipient initiated message interaction events. This message categorization dataset is used to train or evaluate a message classifier.

Abstract: Methods and systems are provided for determining the intent of a recommendation made by a user of a mobile application where the application includes a plurality of separable components, any one or more of which the recommendation can apply to. An application in which a user recommendation control is provided for presentation to a user also includes a tag indicating how a recommendation of the application should be interpreted with respect to the components included therein. The tag can be set by the application developer and can be in the form of text (e.g., a keyword or term) or a uniform resource locator (URL). Where a tag references multiple components of an application, a recommending user can be presented with a recommendation intent query. The recommendation intent query allows a user to designate one or more components of the application to which the user's recommendation should be attributed.