Abstract: Systems and methods are disclosed to implement a neural network training system to train a multitask neural network (MNN) to generate a low-dimensional entity representation based on a sequence of events associated with the entity. In embodiments, an encoder is combined with a group of decoders to form a MNN to perform different machine learning tasks on entities. During training, the encoder takes a sequence of events in and generates a low-dimensional representation of the entity. The decoders then take the representation and perform different tasks to predict various attributes of the entity. As the MNN is trained to perform the different tasks, the encoder is also trained to generate entity representations that capture different attribute signals of the entities. The trained encoder may then be used to generate semantically meaningful entity representations for use with other machine learning systems.

Abstract: The present disclosure presents systems and related methods for creating real-time predictions. One such method comprises receiving, by a computing device, a first set of data and a second set of data, wherein the first set of data comprises a plurality of items available from a first source for a first set of users and the second set of data comprises transaction purchase data for a second set of users that have reward accounts, utilizing a predictive data model that determines a propensity score for a user from only behavior data that is not attributed to the user; receiving a third set of data from a third source comprising social media channel data for a third set of users; and updating the predictive data model to determine the propensity score for the user based at least in part on the third set of data.

Abstract: The present disclosure presents systems and related methods for creating real-time predictions. One such method comprises receiving, by a computing device, a first set of data and a second set of data, wherein the first set of data comprises a plurality of items available from a first source for a first set of users and the second set of data comprises transaction purchase data for a second set of users that have reward accounts, utilizing a predictive data model that determines a propensity score for a user from only behavior data that is not attributed to the user; receiving a third set of data from a third source comprising social media channel data for a third set of users; and updating the predictive data model to determine the propensity score for the user based at least in part on the third set of data.

Abstract: The systems, methods and computer program products (collectively “system”) described herein relate to customized real time data delivery. The system may be configured to receive, by a performance marketing cluster, first data from a first data source. The system may also receive, by the performance marketing cluster, second data from a second data source. The system may determine, by the marketing cluster, an analysis scheme for the first data and the second data based on the first data source. The system may also determine, by the marketing cluster, at least one of a propensity to act or a recommendation selected from a predefined number of available options for a population based on the analysis scheme and the first data source.

Abstract: Systems and methods are disclosed to implement a machine learning system that is trained to assign annotations to text fragments in an unstructured sequence of text. The system employs a neural model that includes an encoder recurrent neural network (RNN) and a decoder RNN. The input text sequence is encoded by the encoder RNN into successive encoder hidden states. The encoder hidden states are then decoded by the decoder RNN to produce a sequence of annotations for text fragments within the text sequence. In embodiments, the system employs a fixed-attention window during the decoding phase to focus on a subset of encoder hidden states to generate the annotations. In embodiments, the system employs a beam search technique to track a set of candidate annotation sequences before the annotations are outputted. By using a decoder RNN, the neural model is better equipped to capture long-range annotation dependencies in the text sequence.

Abstract: A first cluster of web objects is identified from a click-through data structure. The click-through data structure can organize web objects into clusters based on query results of web objects selected by a user. Also, a second cluster of web objects can be identified from a metadata data structure. The metadata data structure can organize web objects into clusters based on metadata associated with the web objects. An output set of web objects is selected, in real time, from the identifier clusters.

Abstract: A method and a system for summarizing a concept are provided. A query corresponding to a concept is received from a user. A plurality of images and corresponding descriptive information may be collected based on the query. The plurality of images and the descriptive information may be processed to form feature vectors and processed descriptive information respectively. Further, one or more topics may be identified for the plurality of images. Each of the plurality of images may be assigned with one or more topic distribution values corresponding to the one or more topics. The one or more topics correspond to the processed descriptive information. A sparse set of images may be determined based on the feature vectors and the assigned topic distribution values, to summarize the concept. Also, a target summary may be built from the summarized concept, by regularizing one or more distribution constraints.

Abstract: The systems, methods and computer program products (collectively “system”) described herein relate to customized real time data delivery. The system may be configured to receive, by a performance marketing cluster, first data from a first data source. The system may also receive, by the performance marketing cluster, second data from a second data source. The system may determine, by the marketing cluster, an analysis scheme for the first data and the second data based on the first data source. The system may also determine, by the marketing cluster, at least one of a propensity to act or a recommendation selected from a predefined number of available options for a population based on the analysis scheme and the first data source.

Abstract: The system may receive an individual transaction record of a purchase at a merchant location. The individual transaction record may be collected at a point of sale. The individual transaction record may also include at least one of a first purchase time, a first purchase location, a first purchase amount, a first merchant category, or a first purchase item. The system may further predict a future purchase based on the individual transaction record and transmit an offer based on the future purchase.

Abstract: A first cluster of web objects is identified from a click-through data structure. The click-through data structure can organize web objects into clusters based on query results of web objects selected by a user. Also, a second cluster of web objects can be identified from a metadata data structure. The metadata data structure can organize web objects into clusters based on metadata associated with the web objects. An output set of web objects is selected, in real time, from the identifier clusters.

Abstract: A first cluster of web objects is identified from a click-through data structure. The click-through data structure can organize web objects into clusters based on query results of web objects selected by a user. Also, a second cluster of web objects can be identified from a metadata data structure. The metadata data structure can organize web objects into clusters based on metadata associated with the web objects. An output set of web objects is selected, in real time, from the identifier clusters.

Abstract: A first cluster of web objects is identified from a click-through data structure. The click-through data structure can organize web objects into clusters based on query results of web objects selected by a user. Also, a second cluster of web objects can be identified from a metadata data structure. The metadata data structure can organize web objects into clusters based on metadata associated with the web objects. An output set of web objects is selected, in real time, from the identifier clusters.

Abstract: A method and a system for summarizing a concept are provided. A query corresponding to a concept is received from a user. A plurality of images and corresponding descriptive information may be collected based on the query. The plurality of images and the descriptive information may be processed to form feature vectors and processed descriptive information respectively. Further, one or more topics may be identified for the plurality of images. Each of the plurality of images may be assigned with one or more topic distribution values corresponding to the one or more topics. The one or more topics correspond to the processed descriptive information. A sparse set of images may be determined based on the feature vectors and the assigned topic distribution values, to summarize the concept. Also, a target summary may be built from the summarized concept, by regularizing one or more distribution constraints.

Abstract: The present invention is directed towards systems and methods for filtering unsafe content items within a multimedia repository. The method of the present invention comprises automatically identifying a first plurality of unsafe content items from the content items and automatically identifying a second plurality of unclassifiable content items from the content items. The method identifies a third plurality of unsafe content items from the content items, the third plurality of unsafe content items identified according to a similarity to the first plurality and manually identifying a fourth plurality of unsafe content items, the fourth plurality of unsafe content items comprising a subset of the second plurality of unclassifiable items. A fifth plurality of unsafe content items is then identified, the fifth plurality of unsafe content items identified according to a similarity to the fourth plurality.

Abstract: The present invention is directed towards systems and methods for improving image labeling using content-based image retrieval. The method according to one embodiment of the present invention comprises selecting an image query and identifying a first image related to the image query. The method then selects a plurality of second images related to the first image and assigns weights to the plurality of second images. Finally, the method displays the first images and the second images on a client device.

Abstract: The present invention is directed towards systems and methods for improving image labeling using content-based image retrieval. The method according to one embodiment of the present invention comprises selecting an image query and identifying a first image related to the image query. The method then selects a plurality of second images related to the first image and assigns weights to the plurality of second images. Finally, the method displays the first images and the second images on a client device.

Abstract: The present invention is directed towards systems and methods for filtering unsafe content items within a multimedia repository. The method of the present invention comprises automatically identifying a first plurality of unsafe content items from the content items and automatically identifying a second plurality of unclassifiable content items from the content items. The method identifies a third plurality of unsafe content items from the content items, the third plurality of unsafe content items identified according to a similarity to the first plurality and manually identifying a fourth plurality of unsafe content items, the fourth plurality of unsafe content items comprising a subset of the second plurality of unclassifiable items. A fifth plurality of unsafe content items is then identified, the fifth plurality of unsafe content items identified according to a similarity to the fourth plurality.

Abstract: Geometric tagging is described. A method for transforming an image into a three dimensional (3D) representation includes receiving a first user input that specifies selection of a category from a set of categories of geometric objects. Each category of the set is associated with one or more taggable features. A list of user controls is presented that correspond to the taggable features of the category. A second user input is received via the list of user controls that associates tags within an image feature of an image. Each of the tags is associated with one of the taggable features. The image is processed according to the tags of the second user input. A 3D representation of the image is presented based on the processing. The image can include structured scenes, with planar and/or non-planar surfaces, and/or free-form surfaces.