Abstract: One or more systems and/or methods for product similarity detection and recommendation are provided. Users may view articles and/or other content that includes images depicting products that may be of interest to the users. These images are processed using image processing functionality such as computer vision to identify the products depicted by the images. A vector embedding model is used to generate product vector representations of the products. Catalog items that are available from a catalog to supplement the articles and other content may be processed to generate catalog item vector representations. When content (an article) with an image depicting a product is to be displayed to the user, similarity between a product vector representation of the product and the catalog item vector representations is determined in order to identify and display catalog items depicting products that are similar to the product depicted by the image in the content.

Abstract: Systems and methods are disclosed for distributed file collection and processing. According to certain embodiments, a producer system may detect a producer configuration file that specifies a file type, identify at least one file of the specified file type, and send the at least one identified file to a collector system. A conductor system may receive a request from a consumer system and allocate the at least one identified file to the consumer system based on the request. The collector system may receive the at least one identified file from the producer system. The consumer system may send the request to the conductor system and pull the at least one identified file from the collector system.

Abstract: One or more systems and/or methods for combining vectors output by multiple different mechanisms for content item retrieval are provided. An image encoder may output a first set of vectors generated by an image model using an input image as input. A text encoder may output a second set of vectors generated by a text model using input text as input. A vector combination module may combine the first set of vectors and the second set of vectors to create a vector output. A weight is applied to the vector output to create a weighted output. An output vector is generated based upon a combination of the first set of vectors, the second set of vectors, and the weighted output. The output vector is used to query a catalog to identify a content item related to the input image and the input text.

Abstract: Systems and methods are disclosed for opting-out of targeted advertising in online advertising environments. One method includes receiving an opt-out verification request, the opt-out verification including an IP address of a user device and HTTP header fields of an HTTP request of the user device; determining a geographic area of the user device based on the IP address of the user device; accessing an opt-out database having entries of user devices that opted-out of receiving targeted advertising based on at least one HTTP header field of the HTTP request; determining whether the user device has opted-out of receiving targeted advertising based on the determined geographic area of the user device, the HTTP header fields of the HTTP quest, and the entries of user devices of the opt-out database; and transmitting a verification acknowledgement when the user device is determined to have opted-out of receiving targeted advertising.

Abstract: Disclosed frameworks for generating an image including a salient object and a staged background include extracting a salient object from a source image and applying a generative model to the salient object to generate the image. According to some embodiments, extracting a salient object from a source image involves using salient object detection method to identify the relevant portions of the source image corresponding to the salient object. In some embodiments, the generative model is a generative adversarial network trained using a domain relevant dataset.

Abstract: The disclosed systems and methods provide a cloud-based framework for the creation and/or enhancement of high quality media content for dissemination over a computerized network. The disclosed framework is configured to operate in accordance with at least one of a content creation, content enhancement and/or rendering template. Such templates can function as rules engines that call specific types of ML or AI algorithms or routines that modify the input data according to the type of template being used. Thus, the disclosed framework can be utilized for creating new high-quality media, improving upon existing media and/or reformatting media for rendering upon dissemination.

Abstract: One or more computing devices, systems, and/or methods are provided. A machine learning model may be trained using a plurality of sets of information. One or more pruning operations may be performed in association with the training to generate a machine learning model with sparse vector representations associated with features of the plurality of sets of information. A request for content associated with a client device may be received. A set of features associated with the request for content may be determined. A plurality of positive signal probabilities associated with a plurality of content items may be determined using the machine learning model based upon one or more sparse vector representations, of the machine learning model, associated with the set of features. A content item may be selected from the plurality of content items for presentation via the client device based upon the plurality of positive signal probabilities.

Abstract: One or more computing devices, systems, and/or methods are provided. A request for content associated with a device and/or a set of request information associated with the request for content may be received. A content item may be transmitted to the device. A set of client information associated with the device may be received. The set of client information may be analyzed to determine a fraudulence label associated with the request for content. Fraud detection information generated based upon the set of request information, the set of client information and/or the fraudulence label may be stored in a fraud detection database. A second request for content associated with a second device and/or a second set of request information associated with the second request for content may be received. A second fraudulence label may be determined based upon the second set of request information and/or the fraud detection database.

Abstract: The disclosed systems and methods provide a novel framework that enables cost-effective, accurate and scalable detection and recognition of key events in sporting or live events. The framework functions by creating a domain-specific video dataset with frame level annotations (i.e., deep domain datasets) and then training a lightweight camera view classifier to detect camera views for a given video. The disclosed framework uses pre-trained pose estimation and panoptic segmentation models along with geometric rules as labeling functions to define scene types and derive frame level classification training data. According to some embodiments, disclosed frameworks may be used to identify key persons or events, select a thumbnail corresponding to a key person or event, generate personalized highlights to enhance user experience and social media promotions for a team, sport or players, and predict and select the best camera view sequence for automatic highlights generation.

Abstract: The present teaching relates to tracking an event at a plurality of distributed servers. In one example, an event to be tracked is determined. A user associated with the event is identified. A script is generated to be embedded in a web page. The script triggers an event message when the user performs an online behavior related to the web page in accordance with the event. The event message triggered by the script is received. A tracing flag is determined from the event message. An instruction is provided to the plurality of distributed servers for executing one or more applications based on the event and the tracing flag.

Abstract: Disclosed are methods and devices for converting mobile notifications to content using a local content exchange. In one embodiment, the method comprises receiving a request for content from a mobile application; querying a data store of available content, the data store of available content populated with one or more notifications received by other mobile applications installed on the mobile device; identifying a responsive notification for the request for content based on the querying, the content item identified based on bid parameters associated with the one or more notifications; and transmitting the responsive notification to the mobile application.

Abstract: Methods, systems, and computer-readable media are disclosed for utilizing unused network capacity for prefetch requests. One method includes: receiving, over a network, network traffic information from a network provider of the network; determining a threshold value for prefetch request fulfillment based on the received network traffic information; receiving, over the network, a plurality of prefetch requests from an application running on a mobile device connected to the network of the network provider; determining, for each prefetch request of the plurality of prefetch requests, a score for the prefetch request based on the received plurality of prefetch requests; and responding to, for each prefetch request of the plurality of prefetch requests, the prefetch request based on the determined threshold value and the determined score for the prefetch request.

Abstract: One or more computing devices, systems, and/or methods for determining whether requests for content are fraudulent are provided. A request for content may be received from a first device. A first user profile associated with the first device may be identified. The first user profile may comprise activity information associated with the first device, demographic information associated with the first device and/or interest information associated with the first device. A user profile database may be analyzed to identify a set of user profiles similar to the first user profile. A relevance score associated with the request for content may be generated based upon the resource, the set of user profiles and/or the first user profile. The relevance score may be compared with a threshold relevance to determine whether the request for content is fraudulent.

Abstract: One or more computing devices, systems, and/or methods are provided. Shaded bid values may be determined and/or submitted to one or more auction modules for participation in auctions. Auction information including at least one of minimum bid values to win associated with the auctions, sets of features associated with the auctions, the shaded bid values associated with the auctions, unshaded bid values associated with the auctions, etc. may be stored in a database. A machine learning model may be trained using a loss function and/or the auction information to generate a first machine learning model with feature parameters associated with features. A bid request, indicative of a second set of features, may be received. The first machine learning model may be used to determine a shaded bid value for submission based upon one or more first feature parameters, of the feature parameters, associated with the second set of features.

Abstract: Systems, apparatuses, and methods are provided for determining a bid value for placing an advertisement onto advertising space available through an electronic marketplace. A method is used for calculating the option value of maintaining the advertisement in the advertising space during one or more periods of time. The option value may be based on expected profits and the estimated future value of maintaining the advertisement. The option value may then be used to calculate the bid price for placing the advertisement.

Abstract: Shaded bid values may be determined and/or submitted to one or more auction modules for participation in auctions. Auction information including at least one of impression indications associated with the auctions, sets of features associated with the auctions, the shaded bid values associated with the auctions, etc. may be stored in a database. A machine learning model may be trained using the auction information to generate a first machine learning model with feature parameters associated with features. A bid request, indicative of a second set of features, may be received. The first machine learning model may be used to determine win probabilities and/or expected bid surpluses associated with multiple shaded bid values based upon one or more feature parameters, of the feature parameters, associated with the second set of features. A shaded bid value for submission may be determined based upon the win probabilities and/or the expected bid surpluses.

Abstract: The present disclosure relates to systems and methods customizing electronic communications. A future event associated with a first user may be determined, and a second user that is associated with the first user may be identified. A plurality of communications involving the first user and the second user may be analyzed. A selection rule may be applied based on the analyzed plurality of communications, the selection rule identifying content from the database. Content from the database may be selected based on the application of the selection rule. An electronic message may be provided to the first user identifying the future event, and the selected content may be provided to the first user.

Abstract: Systems and methods are for identifying online advertisements to display to a user based on historical user data collected across the user's use of a plurality of Internet devices. One method includes receiving, over a network, a request for an advertisement to display at a first device of the user, the request including a unique identifier stored on the user's first device; accessing, in a database, a demographic or browsing history information generated from the user's use of a second device, the demographic or browsing history information being stored in the database in relation to the unique identifier; and identifying, based on the demographic or browsing history information, an advertisement to display at the user's first device. The demographic or browsing history information is synchronized based on the unique identifier being stored on the user's first device and the user's second device.

Abstract: Systems and methods are disclosed for protecting consumer privacy in an online advertising environment. A request may be received from a browser for a webpage along with a unique browser identifier. The browser may be provided a first portion of the webpage that is locally available. The unique browser identifier may be provided to at least one advertising entity, wherein the advertising entity determines an advertisement based, at least in part, on the unique browser identifier. The advertisement may be received from the advertising entity, and provided to the browser as a second portion of the webpage.

Abstract: Systems, apparatuses, and methods are provided for determining a bid value for placing an advertisement onto advertising space available through an electronic marketplace. A method is used for calculating the option value of maintaining the advertisement in the advertising space during one or more periods of time. The option value may be based on expected profits and the estimated future value of maintaining the advertisement. The option value may then be used to calculate the bid price for placing the advertisement.