Abstract: The MULTI-PURPOSE VIRTUAL CARD TRANSACTION APPARATUSES, METHODS AND SYSTEMS (“WIP”) transform wallet in proxy card generation requests and purchase inputs via WIP components into wallet in proxy card generation notifications and wallet in proxy card-based transaction purchase notifications. In one implementation, the WIP server may receive a transaction authentication request associated with a proxy payment identifier, and then determine that the proxy payment identifier is associated with an electronic wallet. The WIP sever may further obtain a payment identifier associated with the electronic wallet, and authenticate the transaction using the obtained payment identifier associated with the electronic wallet.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving, by a neural network (NN), a dataset for generating features from the dataset. A first set of features is computed from the dataset using at least a feature layer of the NN. The first set of features i) is characterized by a measure of informativeness; and ii) is computed such that a size of the first set of features is compressible into a second set of features that is smaller in size than the first set of features and that has a same measure of informativeness as the measure of informativeness of the first set of features. The second set of features if generated from the first set of features using a compression method that compresses the first set of features to generate the second set of features.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving, by a neural network (NN), a dataset for generating features from the dataset. A first set of features is computed from the dataset using at least a feature layer of the NN. The first set of features i) is characterized by a measure of informativeness; and ii) is computed such that a size of the first set of features is compressible into a second set of features that is smaller in size than the first set of features and that has a same measure of informativeness as the measure of informativeness of the first set of features. The second set of features if generated from the first set of features using a compression method that compresses the first set of features to generate the second set of features.

Abstract: Example aspects of the present disclosure are directed to systems and methods that learn a compressed representation of a machine-learned model (e.g., neural network) via representation of the model parameters within a reparameterization space during training of the model. In particular, the present disclosure describes an end-to-end model weight compression approach that employs a latent-variable data compression method. The model parameters (e.g., weights and biases) are represented in a “latent” or “reparameterization” space, amounting to a reparameterization. In some implementations, this space can be equipped with a learned probability model, which is used first to impose an entropy penalty on the parameter representation during training, and second to compress the representation using arithmetic coding after training. The proposed approach can thus maximize accuracy and model compressibility jointly, in an end-to-end fashion, with the rate-error trade-off specified by a hyperparameter.

Abstract: The SNAP MOBILE PAYMENT APPARATUSES, METHODS AND SYSTEMS (“SNAP”) transform real-time-generated merchant-product Quick Response codes via SNAP components into virtual wallet card-based transaction purchase notifications. In one embodiment, the SNAP obtains a snapshot of a QR code presented on a display screen of a point-of-sale device from a mobile device. The SNAP decodes the QR code to obtain product information included in a checkout request of the user, and merchant information for processing a user purchase transaction with a merchant providing the QR code. The SNAP accesses a user virtual wallet to obtain user account information to process the user purchase transaction with the merchant. Using the product information, merchant information and user account information, the SNAP generates a card authorization request, and which the SNAP provides to a payment network for transaction processing. Also, the SNAP obtains a purchase receipt confirming processing of the user purchase transaction.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing visual tracking. In one aspect, a method comprises receiving: (i) one or more reference video frames, (ii) respective reference labels for each of a plurality of reference pixels in the reference video frames, and (iii) a target video frame. The reference video frames and the target video frame are processed using a colorization machine learning model to generate respective pixel similarity measures between each of (i) a plurality of target pixels in the target video frame, and (ii) the reference pixels in the reference video frames. A respective target label is determined for each target pixel in the target video frame, comprising: combining (i) the reference labels for the reference pixels in the reference video frames, and (ii) the pixel similarity measures.

Abstract: The SNAP MOBILE PAYMENT APPARATUSES, METHODS AND SYSTEMS (“SNAP”) transform real-time-generated merchant-product Quick Response codes via SNAP components into virtual wallet card-based transaction purchase notifications. In one embodiment, the SNAP obtains a snapshot of a QR code presented on a display screen of a point-of-sale device from a mobile device. The SNAP decodes the QR code to obtain product information included in a checkout request of the user, and merchant information for processing a user purchase transaction with a merchant providing the QR code. The SNAP accesses a user virtual wallet to obtain user account information to process the user purchase transaction with the merchant. Using the product information, merchant information and user account information, the SNAP generates a card authorization request, and which the SNAP provides to a payment network for transaction processing. Also, the SNAP obtains a purchase receipt confirming processing of the user purchase transaction.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing action localization in images and videos. In one aspect, a system comprises a data processing apparatus; a memory in data communication with the data processing apparatus and storing instructions that cause the data processing apparatus to perform image processing and video processing operations comprising: receiving an input comprising an image depicting a person; identifying a plurality of context positions from the image; determining respective feature representations of each of the context positions; providing a feature representation of the person and the feature representations of each of the context positions to a context neural network to obtain relational features, wherein the relational features represent relationships between the person and the context positions; and determining an action performed by the person using the feature representation of the person and the relational features.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing action localization. In one aspect, a system comprises a data processing apparatus; a memory in data communication with the data processing apparatus and storing instructions that cause the data processing apparatus to perform operations comprising: receiving an input comprising an image depicting a person; identifying a plurality of context positions from the image; determining respective feature representations of each of the context positions; providing a feature representation of the person and the feature representations of each of the context positions to a context neural network to obtain relational features, wherein the relational features represent relationships between the person and the context positions; and determining an action performed by the person using the feature representation of the person and the relational features.

Abstract: The UNIVERSAL ELECTRONIC PAYMENT APPARATUSES, METHODS AND SYSTEMS (“UEP”) transform touchscreen inputs into a virtual wallet mobile application interface via UEP components into purchase transaction triggers and receipt notices. In one implementation the UEP provides, via a user device, a product information search request; and obtains, in response to the product information search request, information on a first product for sale by a first merchant and a second product for sale by a second merchant. The UEP generates a single purchase transaction request, using the information on the first product for sale by the first merchant and the second product for sale by the second merchant. The UEP provides, via the user device, the single purchase transaction request for payment processing. Also, the UEP obtains an electronic purchase receipt for the first product for sale by the first merchant and the second product for sale by the second merchant.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing visual tracking. In one aspect, a method comprises receiving: (i) one or more reference video frames, (ii) respective reference labels for each of a plurality of reference pixels in the reference video frames, and (iii) a target video frame. The reference video frames and the target video frame are processed using a colorization machine learning model to generate respective pixel similarity measures between each of (i) a plurality of target pixels in the target video frame, and (ii) the reference pixels in the reference video frames. A respective target label is determined for each target pixel in the target video frame, comprising: combining (i) the reference labels for the reference pixels in the reference video frames, and (ii) the pixel similarity measures.

Abstract: Example aspects of the present disclosure are directed to systems and methods that learn a compressed representation of a machine-learned model (e.g., neural network) via representation of the model parameters within a reparameterization space during training of the model. In particular, the present disclosure describes an end-to-end model weight compression approach that employs a latent-variable data compression method. The model parameters (e.g., weights and biases) are represented in a “latent” or “reparameterization” space, amounting to a reparameterization. In some implementations, this space can be equipped with a learned probability model, which is used first to impose an entropy penalty on the parameter representation during training, and second to compress the representation using arithmetic coding after training. The proposed approach can thus maximize accuracy and model compressibility jointly, in an end-to-end fashion, with the rate-error trade-off specified by a hyperparameter.

Abstract: The MULTI-PURPOSE VIRTUAL CARD TRANSACTION APPARATUSES, METHODS AND SYSTEMS (“WIP”) transform wallet in proxy card generation requests and purchase inputs via WIP components into wallet in proxy card generation notifications and wallet in proxy card-based transaction purchase notifications. In one implementation, the WIP server may receive a transaction authentication request associated with a proxy payment identifier, and then determine that the proxy payment identifier is associated with an electronic wallet. The WIP sever may further obtain a payment identifier associated with the electronic wallet, and authenticate the transaction using the obtained payment identifier associated with the electronic wallet.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving, by a neural network (NN), a dataset for generating features from the dataset. A first set of features is computed from the dataset using at least a feature layer of the NN. The first set of features i) is characterized by a measure of informativeness; and ii) is computed such that a size of the first set of features is compressible into a second set of features that is smaller in size than the first set of features and that has a same measure of informativeness as the measure of informativeness of the first set of features. The second set of features if generated from the first set of features using a compression method that compresses the first set of features to generate the second set of features.

Abstract: A service provider system may include an application fulfillment platform that delivers desktop applications on demand to desktops on physical computing devices or virtual desktop instances of end users. An application delivery agent installed on an end user's computing resource instance may store application state data (e.g., configuration data, runtime settings, or application templates) or scratch data that is generated by an application executing on the computing resource instance to a secure location on service provider storage resources. After a machine failure or change, or a rebuilding of a virtualized computing resource instance or virtual desktop instance, an application delivery agent installed on the new machine or instance may reinstall the application, retrieve the stored application state or scratch data from service provider resources, and restore the application to the last known persisted state.

Abstract: Methods, systems, and computer-readable media for application streaming using access and execution environments are disclosed. User input comprising a selection of an application by a user is received. The user input is received at an access environment from a computing device associated with a client. The access environment permits access by the user, and the access environment is associated with an execution environment for the application. The execution environment specifies a computing resource description for the application. Execution of the application is initiated using one or more computing resources compatible with the computing resource description. A pixel display is streamed to the computing device associated with the client. The pixel display is associated with the application executing using the one or more computing resources.

Abstract: The SNAP MOBILE PAYMENT APPARATUSES, METHODS AND SYSTEMS (“SNAP”) transform real-time-generated merchant-product Quick Response codes via SNAP components into virtual wallet card-based transaction purchase notifications. Payment information and VAS data can also be provided based on location. A request for payment information can be received. A location can be determined, and a merchant associated with the location can also be determined. Payment information and/or VAS data can be selected based on the merchant and/or location, and can be provided for a payment transaction.

Abstract: The UNIVERSAL ELECTRONIC PAYMENT APPARATUSES, METHODS AND SYSTEMS (“UEP”) transform touchscreen inputs into a virtual wallet mobile application interface via UEP components into purchase transaction triggers and receipt notices. In one implementation the UEP provides, via a user device, a product information search request; and obtains, in response to the product information search request, information on a first product for sale by a first merchant and a second product for sale by a second merchant. The UEP generates a single purchase transaction request, using the information on the first product for sale by the first merchant and the second product for sale by the second merchant. The UEP provides, via the user device, the single purchase transaction request for payment processing. Also, the UEP obtains an electronic purchase receipt for the first product for sale by the first merchant and the second product for sale by the second merchant.

Abstract: A service provider system may include an application fulfillment platform that delivers desktop applications to desktops on physical computing devices or virtual desktop instances. A computing resource instance may be registered with the platform, which generates a unique identifier and a security token for the computing resource instance using multiple authentication mechanisms. An end user of a customer organization may be registered with the platform, which generates a unique identifier and a security token for the end user using multiple authentication mechanisms. An application delivery agent may submit service requests to the platform on behalf of itself or the given user. The identity and security credentials included in the requests may be dependent on the request type and the entities on whose behalf they are submitted. A proxy service on the platform may receive the requests and validate the credentials, then dispatch the requests to other services on the platform.

Abstract: The UNIVERSAL ELECTRONIC PAYMENT APPARATUS, METHODS AND SYSTEMS (“UEP”) transform touchscreen inputs into a virtual wallet mobile application interface via UEP components into purchase transaction triggers and receipt notices. In one implementation the UEP provides, via a user device, a product information search request; and obtains, in response to the product information search request, information on a first product for sale by a first merchant and a second product for sale by a second merchant. The UEP generates a single purchase transaction request, using the information on the first product for sale by the first merchant and the second product for sale by the second merchant. The UEP provides, via the user device, the single purchase transaction request for payment processing. Also, the UEP obtains an electronic purchase receipt for the first product for sale by the first merchant and the second product for sale by the second merchant.