Abstract: A computing platform may be configured to (i) obtain an input dataset, (ii) construct a graph from the input dataset, (iii) for a given node within the constructed graph, generate a first type of embedding vector using a first embedding technique (e.g., a shallow embedding technique) and a second type of embedding vector using a second embedding technique that differs from the first embedding technique (e.g., a deep embedding technique), and (iv) use the first and second types of embedding vectors for the given node and a data science model to render a given prediction for the given node.

Abstract: An example client device is configured to (i) display an interface for deploying a new data science environment at a computing platform, (ii) receive, via the interface, a user selection of (a) a given data science application from a list of data science applications that is presented by the interface and (b) one or more deployment configuration parameters from a set of deployment configuration parameters that is presented by the interface, (iii) transmit, to the computing platform, a first network-based communication comprising an indication of the user selection of (a) the given data science application and (b) the one or more deployment configuration parameters, and (iv) receive, from the computing platform, a second network-based communication comprising an indication that the new data science environment has been deployed based on the user selection of the (a) the given data science application and (b) the one or more configuration parameters.

Abstract: An example computing platform is configured to (a) cause a client device associated with a user to display an interface for deploying a new data science model, where the interface presents the user with a list of deployment templates, and where each of the deployment templates includes data specifying (i) a respective executable model package and (ii) a respective set of execution instructions for the respective executable model package, (b) receive, from the client device, data indicating (i) a user selection of a given deployment template for use in deploying the new data science model and (ii) a given set of configuration parameters for use in deploying the new data science model, and (c) use the given executable model package, the given set of execution instructions, and the given set of configuration parameters to deploy the new data science model.

Abstract: An example computing platform is configured to receive a request from a first user account of a first P2P payment service that a cross-service transaction be initiated with another user account of another P2P payment service. Based on the request, the computing platform generates a transaction identifier for the transaction and sends the identifier to the first P2P payment service's platform. Thereafter, the computing platform receives, from a second P2P payment service's platform, the transaction identifier and an identifier of a second user account of the second P2P payment service and responsively determines that the cross-service transaction comprises a transfer of funds between a first financial account associated with the first user account and a second financial account associated with the second user account. The computing platform then selects and causes a given payment rail platform to transfer funds between the first and second financial accounts.

Abstract: A smart payment card is provisioned with biometric data for an authorized card member and, for at least one financial account, (i) respective financial account information to be used for authorized card-present transactions and (ii) a respective set of virtual account numbers to be used for card-not-present transactions. The smart payment card is configured using a dongle and a user interface displayed at an end-user device. Processing a financial transaction involving a given financial account provisioned on the smart payment card involves authenticating a received biometric input, and if the authentication is successful, (i) transmitting identifying information for the given financial account to a point-of-sale device if the financial transaction is a card-present transaction, or (ii) presenting, at a display of the smart payment card, a first virtual account number from a respective set of virtual account numbers for the given financial account if the transaction is a card-not-present transaction.

Abstract: A framework for interpreting machine learning models is proposed that utilizes interpretability methods to determine the contribution of groups of input variables to the output of the model. Input variables are grouped based on correlation with other input variables. The groups are identified by processing a training data set with a clustering algorithm. Once the groups of input variables are defined, partial dependent plot (PDP) tables for each group are calculated and stored in a memory, which are used for calculating scores related to each group of input variables for a given instance of the input vector processed by the model. Furthermore, Shapley Additive Explanations (SHAP) values for each group can be calculated by summing the SHAP values of the input variables for a given instance of an input vector per group. These scores can then be sorted, ranked for each interpretability method, and then combined into one hybrid ranking.

Abstract: A computing system is configured to (i) inject, into an unbounded stream of events produced by the computing system, a window open sentinel that indicates one or more reconciliation rules for the window, (ii) inject, into the unbounded stream of events, a shard open sentinel that indicates one or more reconciliation rules for a shard, wherein the shard is within the window, (iii) transmit a set of events produced within the shard, (iv) execute the reconciliation rules for the shard to thereby generate one or more reconciliation attributes for the shard, (v) inject, into the unbounded stream of events, a shard close sentinel that indicates the reconciliation attributes for the shard, (vi) execute the reconciliation rules for the window to thereby generate one or more reconciliation attributes for the window, and (vii) inject, into the unbounded stream of events, a window close sentinel that indicates the reconciliation attributes for the window.

Abstract: A method, system, and computer-readable medium are disclosed for detecting and mitigating bias in a trained machine learning model. The method includes the steps of: training the model based on a training data set; detecting bias in the model relative to a protected class; identifying one or more groups of input variables that contribute to the bias; and mitigating bias in the model. Mitigating the bias is performed by constructing a post-processed score function that either (a) neutralizes or partially neutralizes one or more groups of input variables in the input vector of the model, or (b) utilizes a fair score approximation of the model to project the distributions for the protected class and/or the unprotected class to substantially match. In an embodiment, detecting bias in the trained model is performed by comparing distribution for two or more subpopulations based on a distance metric, such as a Wasserstein distance.

Abstract: A first computing system is configured to (i) decide to execute a file transfer of at least one file from the first computing system to a second computing system, where the at least one file comprises data content that is related to a distributed transaction and is stored at the first computing system, (ii) before the file transfer is executed, determine at least one global context identifier for association with the file transfer, (iii) associate the at least one file with a data element that is isolated from the data content and contains the determined global context identifier, and (iv) execute the file transfer by transmitting the at least one file and the associated data element for receipt by the second computing system.

Abstract: An example computing platform is configured to (i) cause a client device to display an interface for deploying a new data science environment, where the interface presents (a) a list of data science applications and (b) a set of user-defined configuration parameters, (ii) receive, from the client device, data indicating (a) a user selection of a given data science application from the list and (b) a user selection of one or more user-defined configuration parameters from the set, (iii) based on the user selection of the given data science application, determine a deployment template for use in deploying the new data science environment, the deployment template specifying (a) an executable environment package and (b) a set of predefined configuration parameters, and (iv) use the given executable environment package, the set of predefined configuration parameters, and the one or more user-defined configuration parameters to deploy the new data science environment.

Abstract: A computing platform may be installed with software technology for creating and executing user-defined anomaly detection rules that configures the computing platform to: (1) receive, from a client device, data defining a given anomaly detection rule that has been created by a user, wherein the given anomaly detection rule comprises at least one anomaly condition that is to be applied to at least one streaming event queue, (2) store a data representation of the given anomaly detection rule in a data store, (3) convert the data representation of the given anomaly detection rule to a streaming query statement, (4) iteratively apply the streaming query statement to the at least one streaming event queue, and (5) while iteratively applying the streaming query statement, make at least one determination that the at least one anomaly condition is satisfied and then cause at least one anomaly notification to be issued to the user.

Abstract: An example computing platform is configured to (a) cause a client device associated with a user to display an interface for deploying a new data science model, where the interface presents the user with a list of deployment templates, and where each of the deployment templates includes data specifying (i) a respective executable model package and (ii) a respective set of execution instructions for the respective executable model package, (b) receive, from the client device, data indicating (i) a user selection of a given deployment template for use in deploying the new data science model and (ii) a given set of configuration parameters for use in deploying the new data science model, and (c) use the given executable model package, the given set of execution instructions, and the given set of configuration parameters to deploy the new data science model.

Abstract: Embodiments provide a methods and systems for securely providing access to sensitive information for a plurality of entities. The method includes receiving, by a payer credential service provider (CSP), from a payer, a request for payee electronically protected information (EPI). The payer CSP transmits to a payee CSP a request for payee EPI data. The payee CSP validates, using a distributed ledger system, whether the payer has permission to request the payee EPI. The payer CSP receives the payee EPI data from the payee CSP. The payer CSP sends the payee EPI data to the payer.

Abstract: Systems, methods, and software are disclosed for provisioning tokens in a networked computing environment including merchants participating in payment networks. The method includes receiving, by a merchant processor, card account data of a purchase card from a cardholder. The method includes determining, by the merchant processor, that the card account data includes a multi-token enabled BIN. The method includes, in response to determining that the card account data includes the multi-token enabled BIN, receiving and storing, by the merchant processor and in a memory of the merchant, respectively, a primary token associated with a front-of-card brand of the purchase card. The method includes, in response to receiving the primary token, transmitting, by the merchant processor, a request for a secondary token. The method includes receiving and storing, by the merchant processor and in the merchant memory, respectively, the secondary token simultaneously with storing the primary token.

Abstract: A method of testing a payment processing network by a credit card company device comprising: (a) programming a terminal to self-initiate one or more authentication requests using test primary account numbers included in a range of test bank identification numbers; (b) determining whether the test primary account numbers included in the test bank identification numbers are received; and (c) in response to the determining that the test primary account numbers are received, determining that the terminal is performing well.