Abstract: A method for maintaining a prepaid payment system comprises a user account that can be utilized to complete a purchase transaction with a merchant. A delayed processing window is introduced between a time when the merchant receives a payment approval notification from the payment system and a time when the payment system transmits a payment request to an issuer of a funding account associated with the user's payment system account. The payment system utilizes a user's stored value account maintained by the payment system to satisfy the requirements of a prepaid program, and therefore processes the payment request received from the merchant and transmits the payment approval notification without obtaining prior authorization from the issuer of the funding account. The payment system submits one or more payment requests for the funding transaction at a time after the completion of the purchase transaction between the user and the merchant.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for detecting an end of a query are disclosed. In one aspect, a method includes the actions of receiving audio data that corresponds to an utterance spoken by a user. The actions further include applying, to the audio data, an end of query model. The actions further include determining the confidence score that reflects a likelihood that the utterance is a complete utterance. The actions further include comparing the confidence score that reflects the likelihood that the utterance is a complete utterance to a confidence score threshold. The actions further include determining whether the utterance is likely complete or likely incomplete. The actions further include providing, for output, an instruction to (i) maintain a microphone that is receiving the utterance in an active state or (ii) deactivate the microphone that is receiving the utterance.

Abstract: Provided are methods and systems for using synthetic resource records to configure and manage web applications and various online services. A user is provided with the ability to setup a single synthetic resource record for their domain such that all of the associated dependent resource records are automatically configured and added to the domain as well. The methods and systems provided utilize synthetic resource records as a way of logically grouping resource records into “human readable” sets of resource records. Through synthetic resource records, users are able to address common configuration issues without having to know the technical specifications of associated zone files or resource records. Additionally, users can complete a variety of configuration tasks related to setting-up web applications without having to look-up resource record values from external sites (e.g., from the domain host).

Abstract: A system and method for ocular stabilization of video images is disclosed. While capturing video images in a forward field of view with a forward-facing video camera of a wearable head-mountable device (HMD), binocular eye-gaze directions of left and right eyes of a user of the HMD may be obtained with an eye-tracking device of the HMD. Based on the obtained binocular eye-gaze directions of left and right eyes of the user of the HMD, convergent gaze directions of the user may be determined as a function of time during an interval concurrent with the capturing of the video images. The captured video images may then be stabilized by compensating for motion of the forward-facing video camera with an intersection of the convergent gaze directions of the user with an image plane of the forward-facing video camera.

Abstract: Conditionally combining data from several sources may include retrieving data from several sources in response to a single request and using one or more redirection responses. The request from a client device may include a device identifier, a third-party identifier, a join identifier based, at least in part, on the device identifier, and first data from a first source, such as a first cookie, associated with the received request. A redirection destination may be determined based on configuration data associated with the third-party identifier, such as whether one or more services are selected and/or activated by a publisher of a resource. A redirection response is transmitted to the client device responsive to the received request and includes the redirection destination and the join identifier. The first data from the received request and second data received from the redirection destination may be associated based on the join identifier.

Abstract: A video delivery system alters information (such as time stamp information) associated with one or more frames of a first compressed video file to allow for accurate rendering of frames within the first compressed video file prior to a transition to a second compressed video file. For example, a time stamp of a particular frame of a compressed video file is altered prior to transmission of the particular frame by a video transmission system such that the altered time stamp indicates a time that has already passed. The particular frame can be decoded by a display system that receives the particular frame so that information included in the particular frame can be used in rendering of subsequently received frames. The display system can bypass display of the particular frame due to the altered time stamp.

Abstract: A user captures images on a user computing device. The user signs in to an application, which transmits the user's images to an account management system, which recognizes objects within the images and assigns one or more object categories to the images and recognizes multiple images comprising objects in a common object category. After receiving user consent, the application groups the images on the user computing device according to object category. The user computing device captures an image of another object. The application transmits the image to the account management system, which detects objects within the image, identifies the object category, and saves the image to the corresponding object category group on the user computing device. After receiving user consent, the account management system finds information for each image in the object category group of images and transmits the information to the user computing device.

Abstract: An apparatus and method for providing a virtual online live video channel is disclosed. The virtual video channel is generated without requiring a dedicated content provider and encoder for the channel. A schedule server stores a schedule for the virtual online live video channel, and the schedule may include one or more of live streaming video content, stored video content, and advertising. A video player establishes communication with a manifest server and requests playback of the video channel. The manifest server generates a per-user manifest file corresponding to the schedule of the video channel and delivers the per-user manifest file to the video player. The video player may then retrieve the desired video content from the content delivery network to produce the virtual online video channel.

Abstract: Systems and methods for generating a slate of ranked items are provided. In one example embodiment, a computer-implemented method includes inputting a sequence of candidate items into a machine-learned model, and obtaining, in response to inputting the sequence of candidate items into the machine-learned model, an output of the machine-learned model that includes a ranking of the candidate items that presents a diverse set of the candidate items at the top positions in the ranking such that one or more highly relevant candidate items can be demoted in the ranking.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for hotword detection on multiple devices are disclosed. In one aspect, a method includes the actions of receiving, by a first computing device, audio data that corresponds to an utterance. The actions further include determining a first value corresponding to a likelihood that the utterance includes a hotword. The actions further include receiving a second value corresponding to a likelihood that the utterance includes the hotword, the second value being determined by a second computing device. The actions further include comparing the first value and the second value. The actions further include based on comparing the first value to the second value, initiating speech recognition processing on the audio data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a dialog system using user feedback. According to one implementation, a method includes receiving, by a dialog engine and from a computing device, a voice input; determining, by the dialog engine, that the received voice input is classified as feedback to an answer of a question; identifying, by the dialog engine, a predetermined feedback score associated with the feedback; and adjusting a confidence score associated with the question and the answer based on the predetermined feedback score.

Abstract: A method for rejecting biased data using a machine learning model includes receiving a cluster training data set including a known unbiased population of data and training a clustering model to segment the received cluster training data set into clusters based on data characteristics of the known unbiased population of data. Each cluster of the cluster training data set includes a cluster weight. The method also includes receiving a training data set for a machine learning model and generating training data set weights corresponding to the training data set for the machine learning model based on the clustering model. The method also includes adjusting each training data set weight of the training data set weights to match a respective cluster weight and providing the adjusted training data set to the machine learning model as an unbiased training data set.

Abstract: A computing device outputs for display at a presence-sensitive display, a graphical keyboard comprising a plurality of keys, receives an indication of at least one gesture to select a group of keys of the plurality of keys, and determines at least one characteristic associated with the at least one gesture to select the group of keys of the plurality of keys. The computing device modifies a spatial model based at least in part on the at least one characteristic and determines a candidate word based at least in part on data provided by the spatial model and a language model, wherein the spatial model provides data based at least in part on the indication of the at least one gesture and wherein the language model provides data based at least in part on a lexicon. The computing device outputs for display at the presence-sensitive display, the candidate word.

Abstract: Systems and methods for optimizing content performance using creative extensions are provided. A content generation system receives request for a content item for presentation on a client device. The request includes an indication of a serving context for the content item. The content generation system uses a creative extension performance model and the serving context for the content item to calculate a predicted performance metric for the content item for multiple different potential creative extensions. Each of the potential creative extensions defines a different action that occurs in response to a user interaction with the content item. The content generation system selects one of the potential creative extensions based on the predicted performance metrics and generates a content item having the selected creative extension using data assets extracted from various data sources. The creative extension performance model is updated using event data from the client device.

Abstract: A method for rejecting biased data includes receiving a bias training data set based on a probability distribution of bias-sensitive variables of a target population and segmenting the bias training data set into clusters based on at least one respective bias-sensitive variable of the target population, each cluster including a bias cluster weight. The method also includes receiving a training data set for a machine learning model and segmenting the training data set into training clusters. Each training cluster is associated with at least one corresponding bias-sensitive variable of the target population and includes a corresponding training data set weight. The method also includes adjusting each training data set weight to match a respective bias cluster weight to form an adjusted training data set and providing the adjusted training data set to the machine learning model as an unbiased training data set.

Abstract: A computer-implemented method for delivering informational events to a user of a computing device includes a processor and a memory. The method includes receiving, into the memory from the user of the computing device, content configuration information defining a first informational element. The method also includes receiving, into the memory from the user of the computing device, trigger event configuration information associated with a trigger event. The method further includes receiving, into the memory from the user of the computing device, target display interface configuration information defining a target display interface. The method also includes determining, by the processor, that the trigger event is satisfied based on the trigger event configuration information. The method further includes automatically retrieving the first informational element based on the content configuration information when the trigger event is satisfied.

Abstract: A method includes executing an initial instance of a change log process for a distributed system, each instance of the change log process configured to store, on memory hardware in communication with the data processing hardware, a transaction history of transactions executed on the distributed system. The method also includes receiving transaction requests for executing corresponding transactions on the distributed system and determining a change log load based on the received transaction requests. The method includes executing at least one subsequent instance of the change log process when the change log load satisfied a threshold load. When multiple instances of the change log process are executing, the method includes ceasing execution of the at least one subsequent instance of the change log process and merging the transaction history of the initial instance of the change log process and the transaction history of the at least one subsequent instance of the change log process.

Abstract: A method and apparatus for selectively adjusting transmit diversity parameters in a mobile communication system including a mobile device and a base station. The mobile device transmits a signal set comprised of a plurality of signals differing only in phase from one another, receive in response an input parameter for each set, and adjust the phase difference of a subsequently transmitted signal set as a function of the input parameters.

Abstract: Systems and methods for extracting audiovisual features from online document elements are described herein. A computing device can identify a first audiovisual content element on an online document and can retrieve a second audiovisual content element from a content provider database. The computing device can extract an image, video, or audio feature from the first and the second audiovisual content elements by applying image feature, video frame feature, or audio fingerprint detection. The computing device can determine a match between the features extracted from the first and the second audiovisual content elements. The computing device can select the second audiovisual content element for display on the online document based on the match. The computing device can transmit the second audiovisual content element for insertion in a content slot of the online document.

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.