Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for partitioning videos. In one aspect, a method includes obtaining a partition of a video into one or more shots. Features are generated for each shot, including visual features and audio features. The generated features for each shot are provided as input to a partitioning neural network that is configured to process the generated features to generate a partitioning neural network output. The partition of the video into one or more chapters is determined based on the partitioning neural network output, where a chapter is a sequence of consecutive shots that are determined to be taken at one or more locations that are semantically related.

Abstract: Systems and methods of the present disclosure relate generally to facilitating temporal content selection. In one implementation, the system parses a first plurality of online documents to generate a first data structure comprising historical keyword co-occurrences. The system can generate, via parsing a second plurality of online documents different from the first plurality of online documents, a second data structure comprising keyword co-occurrences not in the first data structure. The system can temporally link a first keyword of the second data structure with a second keyword of the second data structure. The second keyword may co-occur with the first keyword on a subset of the second plurality of online documents. The system can the temporally linked second keyword to a content selection server.

Abstract: Multi-computing device network based cross-device conversion determination is described. A content selection computer server can identify cross-device conversions. A first computing device accesses third-party content via a first computer network connection. A second computing device accesses a webpage of the third-party content provider via a second computer network connection. The first and devices can be logged into a same online account. The content selection computer server can obtain, via a content tag that includes a script that executes on a webpage that includes the third-party content, data indicating that the first computing device accessed the third-party content. The selection computer server determines the conversion from the content tag and the second computing device having accessed the webpage.

Abstract: A thermostat for controlling an HVAC system is described, the thermostat having a user interface that is visually pleasing, approachable, and easy to use while also providing ready access to, and intuitive navigation within, a menuing system capable of receiving a variety of different types of user settings and/or control parameters. For some embodiments, the thermostat comprises a housing, a ring-shaped user-interface component configured to track a rotational input motion of a user, a processing system configured to identify a setpoint temperature value based on the tracked rotational input motion, and an electronic display coupled to the processing system. An interactive thermostat menuing system is accessible to the user by an inward pressing of the ring-shaped user interface component. User navigation within the interactive thermostat menuing system is achievable by virtue of respective rotational input motions and inward pressings of the ring-shaped user interface component.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing images or features of images using an image classification system that includes a batch normalization layer. One of the systems includes a convolutional neural network configured to receive an input comprising an image or image features of the image and to generate a network output that includes respective scores for each object category in a set of object categories, the score for each object category representing a likelihood that that the image contains an image of an object belonging to the category, and the convolutional neural network comprising: a plurality of neural network layers, the plurality of neural network layers comprising a first convolutional neural network layer and a second neural network layer; and a batch normalization layer between the first convolutional neural network layer and the second neural network layer.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for identifying a graphical user interface including a container having a first child element and a second child element; providing for display the container on the graphical user interface in a first state, the first state including display of the first child element, but not the second child element; applying a transition to the container to effectuate a state change of the container from the first state to a second state, the second state including display of the second child element, but not the first child element, applying the transition including applying a crossfade to the container to graphically fade the first child element to the second child element during application of the transition to the container; and after applying the transition, providing for display the container on the graphical user interface in the second state.

Abstract: A user conducts a wireless payment transaction with a merchant system by transmitting payment information from a user device to a terminal reader without accessing a secure element resident on the user device. A user taps a user device in a merchant system's terminal reader's radio frequency field. The terminal reader and the user device establish a communication channel and the terminal reader transmits a signal comprising a request for a payment processing response. The signal is received by the user device and converted by a controller to a request understandable by an application host processor. The controller transmits the request to the application host processor, where the request is processed, and a response is transmitted to the controller and then to the terminal reader. The response generated by the application host processor is identifiable by the merchant system as a payment response.

Abstract: A distributed mode loudspeaker (DML) includes a flat panel extending in a panel plane. The DML also includes a rigid, elongate member displaced from the flat panel and extending parallel to the panel plane, the elongate member being mechanically coupled to the flat panel at a first position along the elongate member and extending away from the first position to an end of the member free to vibrate in a direction perpendicular to the plane. The elongate member includes a soft magnetic material. The DML also includes an electromagnet system including at least one electrically-conducting coil having an axis perpendicular to the panel plane and displaced from the elongate member. The DML further includes an electronic control module electrically coupled to the electromagnet system and programmed to energize the electrically-conducting coil sufficient such that a magnetic field produced by the electrically-conducting coil displaces the free end of the elongate member.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, that determine whether to illuminate a scene with electromagnetic radiation at a first wavelength or electromagnetic radiation at a second wavelength. An illuminator that is configured to selectively emit electromagnetic radiation at the first wavelength and at the second wavelength is triggered to emit electromagnetic radiation at the first wavelength instead of the second wavelength. A first voltage is applied to a first electrically active filter. A third voltage is applied to a second electrically active filter. A first image is obtained from a first sensor that senses electromagnetic radiation that passes through the first filter. A second image is obtained from a second sensor that senses electromagnetic radiation that passes through the second filter. A depth map is generated from the first image and the second image.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for changing a distributed mode loudspeaker's fundamental frequency. One of the systems includes a distributed mode loudspeaker comprising an actuator that includes: a supported portion, and a cantilevered portion having a length, a first fundamental frequency, and adapted to create a force to cause vibration of a load to generate sound waves using the first fundamental frequency; a support element connected to the supported portion of the actuator and adapted to adjust, based on a change to a shape of the support element, a size of the length of the cantilevered portion to change the first fundamental frequency to a second fundamental frequency with which the load will generate sound waves; and a frequency selection module that provides a signal to the support element to cause the support element to change shape.

Abstract: The present disclosure provides for role switch with perceptible continuity of audio streaming during the role switch. The role switch occurs in two stages, with a first stage including a host link role switch and a second stage including a relay link role switch. For example, the host link role switch concerns the relationship of each audio receiver device to the host device, where the master device receives audio directly from the host for relay to the slave. The relay link role switch concerns a relationship between the audio receiver devices. For example, the communication master controls timing, such as when to send packets between the master and slave devices. Each stage of the two stage procedure may take approximately 100 ms or less. Between each stage, the audio buffers of the master and slave devices have an opportunity to refill.

Abstract: On-demand application permissions is described, including obtaining one or more consents associated with one or more functions of an application, where the application does not allow using the functions without associated consents; receiving, from a user, a consent associated with a function; activating to allow that function to be executed; when a request from the user to use another function is received, determining that the another function requires an associated consent to activate; determining that there is no consent from the user associated with the another function; prompting for a consent associated with the another function; receiving the consent associated with another function; and activating to allow the another function to be executed.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for selecting distributed mode loudspeaker electrodes using an output frequency. One of the methods includes determining, for a piezoelectric transducer, a subset of frequencies, from a range of frequencies, at which to output a sound; selecting, based on the subset of frequencies, one or more electrode pairs from two or more electrode pairs included in the piezoelectric transducer to generate the sound; and providing, by a drive module connected to each of the two or more electrode pairs, current to each of the selected one or more electrode pairs to cause the piezoelectric transducer to generate a force that, when provided to a load, causes the load to generate the sound within the subset of frequencies.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for providing contextual information for a notification. Logical relationships between panels for a user interface of the one or more computing devices are identified. A set of interface patterns that correspond to different logical relationships between panels and describe different spatial relationships for arranging panels having the corresponding logical relationships are obtained. At least one of the logical relationships between the panels for the user interface is determined to match a logical relationship corresponding to a particular interface pattern. The particular interface pattern is selected from among the set of interface patterns to specify the arrangement of the panels of the user interface. The user interface having the panels arranged in the user interface according to the particular interface pattern that is selected is provided.

Abstract: A system includes a panel extending in a plane, an actuator attached to a surface of the panel, and an electronic control module to activate the actuator to cause vibration of the panel. The actuator includes: a plate to create a force to cause vibration of the panel to generate sound waves, having a width, WT, at a first edge; a stub extending from the first edge of the plate, having a width at a region of connection to the plate that is less than WT, the stub being attached to the surface of the panel to transfer the force received from the plate to the panel and cause the panel to vibrate; and a damper supported by a surface of the plate facing the panel coupling the plate to the panel, the damper having a having a width greater than WS.

Abstract: Various aspects of the subject technology related to systems and methods for detecting and correlating anomalous time-series data. A system may be configured to receive and process time-series data associated with one or more network data streams to generate sets of aligned time-series data. The system may detect anomalous time-stamped data points in the sets of aligned time series data and generate groups of annotated time-series data. The annotation identifies specific time-stamped data points as anomalous. The system may determine the number of anomalous groups of annotated time-series data within all groups of annotated time-series data and may further determine the probability that one or more anomalous groups belong to at least one of the groups of annotated time-series data using a generative statistical model and outputting one or more correlated anomalous groups.

Abstract: Systems and methods for providing third-party content can include a data processing system receiving criteria defining online activities of an online activity type related to a product or service context from a computing device of a third-party content provider. The data processing system can identify a cluster of client devices based on the defined online activities, and determine a subset of keywords associated with the defined online activities and the identified cluster of client devices. The data processing system can determine a performance metric of the subset of keywords based on the cluster of client devices, and provide the subset of keywords and the performance metric to the computing device. The data processing system can receive from the computing device a parameter value of a first keyword of the subset of keywords to use in selecting content items of the third-party content provider associated with the first keyword.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, relating to synthetic call initiation. In some implementations, a method includes receiving, by a call triggering module of a call initiating system for placing calls and conducing conversations between a bot of the call initiating system and human callee during the calls, data indicating a first event, determining, by the call triggering module and using the data indicating the first event, that the first event is a trigger event that triggers a workflow for the call initiating system that begins with initiating a phone call, selecting, based on the determined trigger event, a particular workflow, and in response to the selecting, initiating a phone call to a callee specified by the particular workflow.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage medium, for identifying a set of items of digital content displayed to a user; processing the set of items to identify a set of boost items, items within the set of boost items to be prominently displayed, processing comprising: receiving a close-ties score associated with a respective item, the close-ties score representing a relationship between the user and other users associated with the respective item and an importance of a social circle associated with the item to the user, determining that the close-ties score associated with the respective item exceeds a threshold close-ties score, and in response to determining that the close-ties score exceeds the threshold close-ties score, adding the respective item to the set of boost items; providing instructions for boosting a display of items in the set of boost items in a page displayed to the user.

Abstract: A method for providing a cross-application content player is disclosed. The method includes receiving, during playback of a content item in a first content player within a first user interface of a first mobile application, an indication of a user request pertaining to a second mobile application, wherein the first mobile application is associated with a first platform, and the second mobile application is associated with a second platform. The method further includes identifying a play state of the content item corresponding to the user request, and causing playback of the content item in a second content player within a second user interface of the second mobile application based on the play state.