Abstract: This disclosure relates to systems and methods for identifying risky merchants associated with an electronic payment service. In some implementations, a risk assessment system receives a set of features indicative of one or more risks posed by a merchant enrolled in the electronic payment service, where each feature of the set of features indicative of one or more financial attributes of the merchant. The risk assessment system determines a risk score for the merchant based on the set of features using a trained machine learning model, determines a Shapely additive explanation (SHAP) score for each feature of the set of features, and then divides the set of features into multiple groups of features based on a mapping between the features and their respective indicated financial attributes.

Abstract: A user device associated with a user may receive a document associated with the user. The user device may encrypt the received document. The user device may perform patch-based document segmentation on the received document to form a plurality of patches on the received document. The user device may extract text from each patch of the plurality of patches. The user device may analyze the extracted text from each patch to detect a field title and a field value. The user device may encrypt the extracted text and its associated field value for each patch of the plurality of patches. The user device may send the encrypted extracted text and its associated field value to the user device and instructions to display the extracted text and its associated field value on a user interface.

Abstract: A method may include obtaining, for a task of a pipeline of an application: task execution metadata including a set of previous results, and a task image including executable code and an execution environment. The method may further include executing the executable code in the execution environment to generate a set of new results, and controlling execution of the pipeline using the set of new results and the set of previous results.

Abstract: A method for classifying organizations involves obtaining, for an unknown organization, transactional data representing a multitude of transactions. The transactional data comprises a descriptive text for each of the multitude of transactions. The method further involves processing the descriptive text for each of the multitude of transactions to obtain one vector representing the unknown organization, categorizing the unknown organization using a classifier applied to the vector, and identifying a software service for the unknown organization, according to the categorization.

Abstract: Dynamic state-space modeling within a special purpose hardware platform to determine non-conversion risks for each trial user and churn risks for each active subscriber having exhibited a sequence of behaviors. The state-space model may be operable to determine a loss risk for each of a provider's active trial users and/or subscribers.

Abstract: A method implements an embedded remote desktop in integrated module. An integrated application is received, by a client device, from an integrated server. The integrated application includes an access link to a remote server and integrated application content that references live information from a live application. The live application is executing concurrently with the integrated application and is configured to produce the live information responsive to the integrated application content. A connection is established to the remote application. The remote application establishes a connection to the live application. The live information is received at the client device. A user input is received responsive to the live information. The user input is transmitted to the integrated server.

Abstract: Predictive modeling within a special purpose hardware platform to determine scenarios that are most likely to increase conversion potential for each trial user and retention potential for each active subscriber of a service, collectively referred to as a propensity score. The predictive models are integrated with a contextual marketing system that uses a loss risk assessment to learn user behavior and optimize content messaging designed to improve actual conversion or retention behavior for the user.

Abstract: A method may include receiving an unstructured question from a user having structured contextual features. The unstructured question may include tokens. The method may further include converting, using a sentence embedding model, the tokens to a question vector, assigning the question vector to a question cluster, assigning, by applying a user clustering model to the question cluster and the structured contextual features, the user to a user cluster, and assigning, using a trained machine learning model, a channel to the user cluster. The channel may be used to communicate with a customer service agent for a management application. The trained machine learning model may assign, using metrics, a channel to each user cluster. The method may further include recommending, based on assigning the channel to the user cluster, the channel to the user for the question.

Abstract: Systems and methods may be used to generate and use a structured form representation and structured metadata. The structured form representation and structured metadata may include information relevant to a particular context and may be used to update document templates, import new documents and update document versions into software, automate data entry for document completion, update records to include new and or updated information, and provide other functionality of an information service.

Abstract: A method including training predictor machine learning models (MLMs) using a first data set. The trained predictor MLMs are trained to predict classifications of data items in the first data set. The method also includes training confidence MLMs using second classifications, output by the trained predictor MLMs. The method also includes generating an aggregated ranked list of classes based on third classifications output by the trained predictor MLMs and second confidences output by the trained confidence MLMs. The method also includes training an ensemble confidence MLM using the aggregated ranked list of classes to generate a trained ensemble confidence MLM. The trained ensemble confidence MLM is trained to predict a corresponding selected classification for each corresponding data item in a training data set containing second data items similar to the first data items.

Abstract: A method for mitigating cold starts in recommendations includes receiving a request that identifies a requested page and identifying a content vector of the requested page. The content vector is generated based on providing text of the requested page to a neural network text encoder. The method further includes selecting, based on a rank engine and the content vector, a link to a cold start page that does not satisfy a threshold level of interaction data. The rank engine ranks the selected link above a second link to a warm page that does satisfy the threshold level of the interaction data. The method further includes presenting the requested page with the selected link.

Abstract: Systems and methods that may be used to provide personalized financial nudges to users of a financial service that may be used to further the users' savings intentions (e.g., a savings goal, an emergency fund, etc.). The disclosed systems and methods may increase user interactivity with the financial service and the services it offers by providing personalized nudges that are based on, among other things, an evaluation of various behavioral economics principles. A machine learning recommendation system may be used to fit and output different nudges to users in a personalized way to maximize their savings' intentions.

Abstract: Displaying a template of a document in a GUI. The template includes displayed sections. The sections includes at least one configuration defined by conditions, which are defined by rules. The sections are defined by a distinct set of conditions defined by a distinct set of rules. The configuration, conditions, and rules are stored separately from the template and from each other. When modifying the template, at least one of a condition change and a rule change required to implement the change to the template is identified. At least one of a modified condition and rule is retrieved. The first configuration is updated by modifying at least one of the conditions and the rules using at least one of the revised condition and the revised rule. The second configuration for the GUI is rendered to form a modified template. The modified template is displayed on the graphical user interface.

Abstract: Systems, methods and articles of manufacture for analyzing past year tax data for a taxpayer and determining an estimated current year tax liability in order to compute estimated tax payments. The tax preparation system comprises a tax estimate engine configured to access a plurality of calculated past year tax calculation graphs. The tax estimate engine analyzes and compares the calculated past year tax calculation graphs to determine similar node(s) and/or dissimilar node(s) and also determines a similar node estimating value for the similar nodes and/or a dissimilar node estimating value for the dissimilar nodes. A tax calculation engine calculates a current year tax calculation graph using the similar node estimating values and/or dissimilar node estimating values. The calculated current year tax calculation graph includes an estimated tax liability for the current year, which the system uses to compute estimated tax payments for the taxpayer for the current tax year.

Abstract: Systems and methods here may be used for pre-processing images, including using a computer for receiving a pixelated image of a paper document of an original size, downscaling the received pixelated image, employing a neural network algorithm to the downscaled image to identify four corners of the paper document in the received pixelated image, re-enlarging the downscaled image to the original size, identifying each of four corners of the paper document in the pixelated image, determining a quadrilateral composed of lines that intersect at four angles at the four corners of the paper document in the pixelated image, defining a projective plane of the pixelated image, and determining an inverse transformation of the pixelated image to transform the projective plane quadrilateral into a right angled rectangle.

Abstract: Certain aspects of the present disclosure provide techniques for generating a user experience for a software program product based on a knowledge engine. Techniques for generating the user experience include a UI builder tool providing a set of tabular UI views and receiving in each tabular UI view corresponding input data for generating a calculation graph, a completeness graph, and a client UI view. Based on the input data, the UI builder tool and knowledge engine can generate a set of artifact files. The knowledge engine can generate and/or execute the calculation and completeness graphs as defined in the corresponding artifact files. The UI builder tool can generate an instance of the client UI view. With the generated calculation graph(s), completeness graph(s), and an instance of the client UI view, the user experience can be provided to a computing device.

Abstract: A method may include receiving, from a software application, pre-signup data of a user, embedding the pre-signup data to obtain a pre-signup data vector, generating, by applying a topic model to the pre-signup data vector, a pre-signup topic vector, receiving action descriptions of actions of the software application, embedding the action descriptions to obtain action description vectors, generating, by applying the topic model to the action description vectors, action description topic vectors, generating scores by multiplying the pre-signup topic vector and the action description topic vectors, and recommending, to the user and using the scores, an action.

Abstract: An encrypted database system includes a memory storing a database comprising a plurality of logical structural elements each respectively including an unencrypted fuzzed value and encrypted sensitive data formed by encrypting a sensitive data value. The system also includes a processor in communication with the memory and configured to form the plurality of logical structural elements and store the plurality of logical structural elements in the memory. Forming a logical structural element comprises generating the unencrypted fuzzed value for the sensitive data value, encrypting the sensitive data value, and storing the encrypted sensitive data value and the unencrypted fuzzed value in the same logical structural element in the database. The unencrypted fuzzed value is within a predetermined value range and is different from the sensitive data value.

Abstract: A system may include transaction storage devices. Each transaction storage device may include a data store configured to receive, from a first entity, a request to push a detailed transaction corresponding to a secure identifier. The secure identifier may be generated, using an encoding function, from a user identifier of a user. The data store may be further configured to store the detailed transaction based on a first determination to trust the first entity. The system may further include an access controller configured to perform the first determination by applying a first security rule corresponding to a type of the secure identifier to the request to push the detailed transaction, and a registry configured to store at least the first security rule.

Abstract: Certain aspects of the present disclosure provide techniques for mapping natural language to stored information. The method generally includes receiving a long-tail query comprising a natural language utterance from a user of an application associated with a set of topics and providing the natural language utterance to a natural language model configured to identify nodes of a knowledge graph. The method further includes, based on output of the natural language model, identifying a node of a knowledge graph associated with the natural language utterance, wherein the output of the natural language model includes a node identifier for the node of the knowledge graph and providing the node identifier to the knowledge engine. The method further includes receiving a response associated with the node of the knowledge graph from the knowledge engine and transmitting the response to the user in response to the long-tail query.