Abstract: A method performed at an electronic device includes: executing an application; establishing a session between the executing application, a remote server associated with the application, and a media device that transmits output from the session to an output device, wherein the output from the session includes output from the executing application; and presenting to users of one or more second electronic devices on which the application is not executing and/or installed a session identifier associated with the session, wherein the session identifier identifies a controller configuration for enabling the users of the one or more second electronic devices to interact with the session.

Abstract: Systems and methods are described for improving endpoint detection of a voice query submitted by a user. In some implementations, a synchronized video data and audio data is received. A sequence of frames of the video data that includes images corresponding to lip movement on a face is determined. The audio data is endpointed based on first audio data that corresponds to a first frame of the sequence of frames and second audio data that corresponds to a last frame of the sequence of frames. A transcription of the endpointed audio data is generated by an automated speech recognizer. The generated transcription is then provided for output.

Abstract: Provided are systems and methods for synthesizing novel views of complex scenes (e.g., outdoor scenes). In some implementations, the systems and methods can include or use machine-learned models that are capable of learning from unstructured and/or unconstrained collections of imagery such as, for example, “in the wild” photographs. In particular, example implementations of the present disclosure can learn a volumetric scene density and radiance represented by a machine-learned model such as one or more multilayer perceptrons (MLPs).

Abstract: A user conducts activities, such as researching a product online or visiting a store that sells the product. A user then utilizes a risk analysis system to receive and store user activity data for the user's activities. When a purchase attempt is made with the user's financial account, a merchant sends a transaction request to the risk analysis system. The risk analysis system locates a record for the user and determines whether the product is identified in the user activity data. If the product is identified, the risk analysis system provides a risk score to the merchant indicating that the transaction is unlikely fraudulent. Alternatively, identification of the product provides a positive factor among multiple factors considered in a transaction risk analysis. An absence of the product in the activity can be used as a neutral or negative factor among multiple factors considered in a risk analysis of the transaction.

Abstract: At least one aspect of the present disclosure is directed to systems and methods of generating on-device content items to improve security and network utilization. A client device can access content resources and category information from a data structure in the memory the memory of the client device. The client device can generate on-device content items based on the content resources and the category information. The client device can receive an indication to display a content item in an application executing on the client device. The client device can scan the client device for context information to create a relevant device context. The client device can select an on-device content item based on the relevant device context. The client device can provide the selected on-device item to the application for display, detect an interaction with the displayed content item, and update a content selection model or a content generation model.

Abstract: A method for protecting against exposure to content violating a content policy, the method including receiving a number of content items including a first set of content items associated with a content group, determining a measurement associated with an amount of the first set of content items belonging to a specific content category, assigning one or more of the number of content items to be categorized by at least one of the machine learning algorithm or a manual review process, automatically applying the specific content category to one or more other content items of the content group such that the one or more other content items are not reviewed by the manual review process, and transmitting at least one of the number of content items, wherein the content category of each of the number of content items indicates whether the specific content item violates any content policies.

Abstract: Implementations are directed to utilizing an assistant to automatically navigate an interactive voice response (IVR) tree to arrive at a target state during an assisted telephone call. The assistant can receive input to initiate the assisted telephone call, identify an entity to engage with, on behalf of the user, and during the assisted telephone call, based on the input, and identify an IVR tree stored in association with the entity. In various implementations, navigation of the IVR tree can be modified based on interaction(s) detected at a client device subsequent to initiating the assisted telephone call. In various implementations, the assisted telephone call can be initiated from a search interface, and the target state can be associated with a given search result. In various implementations, the IVR tree can be dynamic in that only a subset of candidate state(s) of the IVR tree may be available as the target state.

Abstract: The technology provides for a system for determining a gesture provided by a user. In this regard, one or more processors of the system may receive image data from one or more visual sensors of the system capturing a motion of the user, and may receive motion data from one or more wearable computing devices worn by the user. The one or more processors may recognize, based on the image data, a portion of the user's body that corresponds to a gesture to perform a command. The one or more processors may also determine one or more correlations between the image data and the received motion data. Based on the recognized portion of the user's body and the one or more correlations between the image data and the received motion data, the one or more processors may detect the gesture.

Abstract: Example aspects of the present disclosure are directed to systems and methods for learning data augmentation strategies for improved object detection model performance. In particular, example aspects of the present disclosure are directed to iterative reinforcement learning approaches in which, at each of a plurality of iterations, a controller model selects a series of one or more augmentation operations to be applied to training images to generate augmented images. For example, the controller model can select the augmentation operations from a defined search space of available operations which can, for example, include operations that augment the training image without modification of the locations of a target object and corresponding bounding shape within the image and/or operations that do modify the locations of the target object and bounding shape within the training image.

Abstract: Methods, systems, and apparatus, including computer programs stored on a computer-readable storage medium, for asynchronous execution of client requests. In some implementations, data indicating a user request to a digital assistant is received. An action corresponding to the user request is determined. It is determined that the action is classified as an action to be performed asynchronously to the user request. A confirmation message is sent, for output, and the action is performed asynchronously to the user request.

Abstract: A companion advertisement is displayed on a control client which is also used to control the content being shown on a display client. The companion advertisement is related to a video or advertisement being displayed on the display client. Such a companion advertisement takes advantage of previously unused screen real estate and enables the user to interact with the advertisement or view further information without any interruption of the content being displayed on the display client. A server keeps the content displayed on the display client and control client synchronized. The server recognizes when a user is logged into both a display client and a control client and notifies the control client that an advertisement should be displayed. The control client then retrieves an advertisement related to the content on the display client.

Abstract: In a streaming application environment, input generated in a remote device may be synchronized with rendered content generated by a virtual streaming application in the streaming application environment in part by using frame refresh events passed between the remote device and the streaming application environment to enable the streaming application environment to effectively track a frame refresh rate of the remote device such that input events received from the remote device may be injected into the virtual streaming application at appropriate frame intervals.

Abstract: A method of generating search suggestions includes receiving an indication of a current location of the user device. After a user accesses a search function on the user device and before the user submits a search request, the method includes determining search results associated with locations in proximity to the current location of the user device, using query log data indicating selections of past search results by users after presentation of the past search results in response to respective past queries. The method also includes determining relative positioning of the search results based on the number of users that selected each search result, determining categories associated with the search results, grouping the search results by the determined categories; and sending the grouped search results and the one or more categories associated with the search results to the user device for display according to the determined relative positioning.

Abstract: The present disclosure is directed to associating computing devices with each other based on computer network activity for selection of content items as part of an online content item placement campaign. A first linking factor is identified based on a connection between a first device and the computer network via a first IP address during a first time period, and based on a connection between a second device and the computer network via the first IP address during the first time period. A number of devices that connect with the computer network via the first IP address is determined. A positive match probability is generated. A second and third linking factors are monitored. A negative match probability is determined based on the second and third linking factors. The first device is linked with the second device based on the positive and negative match probabilities.