Abstract: Systems and methods directed to returning personalized image-based search results are described. In examples, a query including an image may be received, and a personalized item embedding may be generated based on the image and user profile information associated with a user. Further, a plurality of candidate images may be obtained based on the personalized item embedding. The candidate images may then be ranked according to a predicted level of user engagement for a user, and then diversified to ensure visual diversity among the ranked images. A portion of the diversified images may then be returned in response to an image-based search.

Abstract: Aspects of the present disclosure relate to a providing related content recommendations in response to a user search query by supervising the training of pair embeddings using contrastive learning and pairwise co-click signals. The approach combines a two tower model architecture with a cascaded multilayer perceptron model to enable the adoption of variable combinations of input features and more representative learned pair embeddings. The learned embeddings undergo supervised contrastive loss training to generate a related content recommendation model, which is subsequently evaluated using both online and offline metrics. The related content recommendation model can provide results to search queries that improve recommendation quality and increase user engagement, thereby ultimately enhancing long term user experience.

Abstract: Systems and methods directed to returning personalized image-based search results are described. In examples, a query including an image may be received, and a personalized item embedding may be generated based on the image and user profile information associated with a user. Further, a plurality of candidate images may be obtained based on the personalized item embedding. The candidate images may then be ranked according to a predicted level of user engagement for a user, and then diversified to ensure visual diversity among the ranked images. A portion of the diversified images may then be returned in response to an image-based search.

Abstract: Systems and methods can be implemented to conduct searches based on images used as queries in a variety of applications. In various embodiments, a set of visual words representing a query image are generated from features extracted from the query image and are compared with visual words of index images. A set of candidate images is generated from the index images resulting from matching one or more visual words in the comparison. A multi-level ranking is conducted to sort the candidate images of the set of candidate images, and results of the multi-level ranking are returned to a user device that provided the query image. Additional systems and methods are disclosed.

Abstract: This document relates to using input images to selectively trigger skills. For example, the input images can be analyzed using a machine learning model, which can output image metadata characterizing content of the input images. Different skills can be selectively triggered based on the image metadata. For example, a given skill can register to be triggered when the image metadata matches one or more triggering conditions specified by that skill.

Abstract: This document relates to using input images to selectively trigger skills. For example, the input images can be analyzed using a machine learning model, which can output image metadata characterizing content of the input images. Different skills can be selectively triggered based on the image metadata. For example, a given skill can register to be triggered when the image metadata matches one or more triggering conditions specified by that skill.

Abstract: Systems and methods can be implemented to conduct a visual search as a service in a variety of applications. In various embodiments, a system is configured to provide searching capabilities of content provided by a first entity in response to a search request by a second entity. An image provided by the second entity can be used by the system as a query image to search the content of the first entity. In an embodiment, the first entity can be a commercial entity providing such a system with image related content regarding its products and services such that any number of individual consumers can search for particular products and services of the commercial entity via their communication enabled devices. In addition, such systems can be arranged for other embodiments to provide customized searches of a single source by many individual devices. Additional systems and methods are disclosed.

Abstract: Systems and methods can be implemented to conduct searches based on images used as queries in a variety of applications. In various embodiments, a set of visual words representing a query image are generated from features extracted from the query image and are compared with visual words of index images. A set of candidate images is generated from the index images resulting from matching one or more visual words in the comparison. A multi-level ranking is conducted to sort the candidate images of the set of candidate images, and results of the multi-level ranking are returned to a user device that provided the query image. Additional systems and methods are disclosed.

Abstract: A query processing system (QPS) for generating query suggestions to a user by matching the user's input query against a set of query-detection grammars defined by respective rule modules, and with respect to a structured knowledge base. In response to the matching operation, the QPS may generate a set of synthetic query suggestions to the user, representing refinements of the user's input query. The QPS can also present a refinement tool that conveys refinement options to the user. The QPS may also present one or more preview items to the user upon the user's selection of one of the synthetic query suggestions. Each synthetic query suggestion and each preview item is guaranteed to having a matching counterpart information item in the structured knowledge base.

Abstract: A computer-implemented training system is described herein for generating at least one model component based on labeled training data. The training system produces the labels in the training data by leveraging textual information expressed in already-evaluated documents. In another implementation, the training system generates a first model component and a second model component. In one implementation, in an application phase, a computer-implemented model-application system applies the first model component to identify an initial set of related items that are related to an input item (such as a query). The model-application system then applies the second model component to select a subset of related items from among the initial set of related items.

Abstract: A query processing system (QPS) for generating query suggestions to a user by matching the user's input query against a set of query-detection grammars defined by respective rule modules, and with respect to a structured knowledge base. In response to the matching operation, the QPS may generate a set of synthetic query suggestions to the user, representing refinements of the user's input query. The QPS can also present a refinement tool that conveys refinement options to the user. The QPS may also present one or more preview items to the user upon the user's selection of one of the synthetic query suggestions. Each synthetic query suggestion and each preview item is guaranteed to having a matching counterpart information item in the structured knowledge base.

Abstract: Methods, systems, and apparatus for generation, distribution and verification of tokens are described. In an implementation, a method is described in which a value of an offer is determined and a token for representing the offer is generated. The token has a number of characters based on the determination of the value of the offer.

Abstract: Methods, systems, and computer-readable media for presenting discount offers to a user. The discount offer may be based on a discount offered by a retailer or supplier and may be combined with a discount offered by one or more additional retailers or suppliers. The discount program may be coordinated by a centralized discount server that may track user browsing behavior and purchases. The discount server may present discounts based on a product previously purchased.

Abstract: Systems and/or methods are presented that facilitate determining the contextual influence of a user in referrals of products or services, and determining an amount of compensation for successful referrals based on the contextual influence of the user. A central service component can track activity and receive data associated with referrals, including data related to the number and type of referrals, and the number of successful referrals. An evaluation component can analyze referral data associated with the user and can determine a conversion ratio as the number of successful referrals compared to the total number of referrals. The evaluation component can utilize the conversion ratio to facilitate determining an amount of compensation that can be awarded and distributed to the user based on the influence of the user within the context of the product or service referred by the user.

Abstract: Receipts may be received at a location that a customer normally uses to receive electronic correspondence, such as an e-mail address, an instant messaging address, etc. Among the items that are sent to that location, those items that contain a receipt may be identified. The identified receipts, or information extracted from such receipts, may be sent to a receipt store. An action, such as displaying the receipts to a customer, may be taken based on content stored in the receipt store. The information that is stored in a receipt store and/or that is displayed to customers may have an arbitrary level of detail.

Abstract: The disclosed subject matter pertains to systems and/or methods that facilitate associating user activity, such as offline user activity, with an online advertisement. An ad link component can receive information associated with online activity and offline activity of a user, online advertisements, and other information, and can analyze such information. The online activity can include viewing, clicking on, pledging interest in, or obtaining a token related to, an online advertisement. The ad link component can employ ad link criteria to determine whether offline user activity, such as a purchase or other desired activity, can be associated with the online advertisement. If the ad link component determines the offline user activity is associated with the online advertisement, an incentive, which can be based on incentive criteria, can be awarded to the user. A distribution component can distribute the incentive to the account of the user.

Abstract: Systems and/or methods are presented that can employ an ad object comprising a data packet(s) to facilitate providing an advertising platform that is ecosystem independent and can support various advertising models, such as affiliate, viral, and/or word-of-mouth (WOM) advertising. The data packet can be associated with an advertisement for a product(s) and/or service(s) and can include a mechanism(s) that facilitates enabling users to reproduce the data packet and associated advertisement in different sites, such as user sites. As the data packet propagates through a community network, the data packet can receive and accumulate data, such as data relating to purchases, comments, or ratings, storing such data in the data packet and displaying a portion of such data with the advertisement to facilitate WOM advertising. The data packet facilitates providing incentives to users who perform desired actions (e.g., purchases, syndication) related to the advertisement.

Abstract: Systems and/or methods are presented that facilitate creating clusters of users that can be linked to each other based on common interactions of such users with an object associated with an advertisement for a product or service. A central service component can track activity and receive data associated with objects, including data related to interactions with such objects by users in a community network. An evaluation component can analyze received data, and can create links between users and/or clusters of users based on common interactions of users with a particular object. The evaluation component can also link clusters that have a particular user in common. The evaluation component can assign a rank or weight level to descriptive content associated with an object, and an associated product or service, based on common object interactions between users.

Abstract: Systems and/or methods are presented that facilitate determining the contextual influence of a user in referrals of products or services, and determining an amount of compensation for successful referrals based on the contextual influence of the user. A central service component can track activity and receive data associated with referrals, including data related to the number and type of referrals, and the number of successful referrals. An evaluation component can analyze referral data associated with the user and can determine a conversion ratio as the number of successful referrals compared to the total number of referrals. The evaluation component can utilize the conversion ratio to facilitate determining an amount of compensation that can be awarded and distributed to the user based on the influence of the user within the context of the product or service referred by the user.

Abstract: The claimed subject invention presents systems and methods to compensate directly a user of a service platform through advertising spending in exchange for knowledge of the user intent. Advertising spending contributions are allocated to a compensation account that can be accessed by an advertising platform, affiliated with the service platform or not, to compensate users and to update compensation records and transactions with the advertiser. Compensation fraud is actively and reactively mitigated. Compensation of the user is conducted securely and includes compensation points, direct payments and instant rebates, which can be used in transactions with online and offline service platforms. The intent-compensation proposition funded through advertising spending creates a price incentive that differentiates both service platform and advertiser from their competitors, and creates value for user and advertiser.