Transaction cards, systems and methods for performing fraud detection at POS devices based on analysis of feature sets are disclosed. In one embodiment, an exemplary transaction card may comprise one or more sensors configured to collect pre-card-use sensor data regarding a user of the card, circuitry that assembles such data into feature sets and performs fraud detection, and data storage. According to some aspects, the fraud detection may include comparing user specific sensor data, collected for a current transaction, to a user-specific risk profile validation model to determine a risk score for the current transaction, and transmitting the risk score to a card transacting device when a card is presented during a transaction. In some implementations, the risk score may enable the card transacting device to evaluate a risk associated with accepting the transaction card to complete the attempted transaction.

Logo candidates for a specific ISO data can be identified from external resources based on the enriched merchant data. Low quality images of the logo candidates are filtered out with image analysis including entropy ratio evaluations of the logo candidates. Also, the logo candidates are processed with high quality filtering including classification of the logo candidates with a deep learning classifier for distinguishing logos from non-logos. A logo from the logo candidates is selected to associate with the ISO data packets. A display having the selected logo associated with a transaction of the ISO data packets can be generated for display to users.

Techniques for displaying a stage-specific pipeline view with a prediction engine are disclosed. A system displays a plurality of regions representing various stages of completion for a plurality of transactions. The system determines a stage of completion for each of the plurality of transactions at a first point-in-time, and generates and displays visualizations representing each of the plurality of transactions in one of the plurality of regions based on the respective current stage of completion. Generating a visualization representing a first transaction includes determining a likelihood of the first transaction completing a stage associated with the first transaction. The likelihood may be determined by selecting attributes associated with the first transaction and identifying prior transactions with similar attributes. The system may compute the likelihood of the first transaction completing the stage based on completion rates associated with the prior transactions, and select the visualization based on the computed likelihood.

A payment HSM hosted in a data center and comprising a host interface accessible by a remote end-user entity running a payment application using critical resources protected in the payment HSM, a second interface for main, operational management of the payment HSM by the end-user entity, and an Out-Of-Band, OOB, management interface being distinct and physically isolated from the communication channel of the second interface, and configured to allow secure access to the payment HSM by a third-party entity, distinct from the end-user entity. A resident, remotely configurable provisioning state-machine is implemented in the HSM for the management of the provisioning of the payment HSM for service to one or more end-user entities, under the control of the third-party entity over the OOB management interface.

A method of approving expenditures in real-time, comprising receiving an expenditure request initiated by a user associated with an organization for transferring funds of the organization in exchange for one or more products and/or services, identifying request attribute(s) relating to the user, the value, the funds, the product, the service, a time of reception of the expenditure request and/or a geographical location of the user, analyzing scheduling data obtained from one or more online data sources which is indicative of one or more activity attributes of activity(s) scheduled for the user, automatically determining compliance between the request attribute(s) and one or more expenditure rules predefined for the product(s) and/or service(s) with respect to the activity attribute(s) and transmitting a response to the expenditure request according to the determination which includes approval of the expenditure request in case of compliance and rejection in case of incompliance.

Machine learning models, semantic networks, adaptive systems, artificial neural networks, convolutional neural networks, and other forms of knowledge processing systems are disclosed. An ensemble machine learning system is coupled to a graph module storing a graph structure, wherein a collection of entities and the relationships between those entities forms nodes and connection arcs between the various nodes. A hotfile module and hotfile propagation engine coordinate with the graph module or may be subsumed within the graph module, and implement the various hot file functionality generated by the machine learning systems.

Embodiments provide a method of integrating a loyalty program with a payment card of a customer. In some implementations, the method includes sending, by a merchant terminal, a payment transaction request to a server system associated with a payment network. The payment transaction request includes a payment transaction amount to be paid to a merchant account from an issuer account of the customer and a consumer identifier linked to the loyalty program associated with the merchant loyalty card. The method includes receiving a notification comprising a payment transaction approval message and a machine-readable script through the payment network. The machine-readable script comprises the consumer identifier and is executable by the merchant terminal. The method further includes storing the consumer identifier associated with the merchant loyalty card with existing customer data in the payment card of the customer by executing the machine-readable script thereby integrating the loyalty program with the payment card.

An account holder's portfolio of transactions may be represented as a network of interconnected transaction nodes where each node represents a credit card transaction. This network may then be analyzed using artificial intelligence and machine learning techniques coupled with visual representations of the interrelated nodes to draw conclusions. An account holder or other system entity may report a fraudulent transaction that employs the holder's account information. A backend system may organize transaction information as a network of data nodes that includes a variety of interrelated information. The backend system may then identify all financial transaction nodes within the network that are related or connected by common data. For example, multiple transactions may include a common merchant as the reported fraudulent transaction. The backend may then perform an analysis of the nodes to identify likely fraudulent transactions based on one or more of the data elements for each node.

Systems and methods for generating customer transaction test data that simulates real world customer transaction data are disclosed. In one embodiment, a method may include (1) generating an account type reference file defining a plurality of account types; (2) generating a transaction type reference file defining a plurality of transaction types; (3) generating a plurality of customer test data records based on at least one of the plurality of parameters, wherein each customer test data record comprises a plurality of data fields, the data in each data field having a format that simulates real-world data; (4) verifying a uniqueness of each of the plurality of customer records; (5) generating a random number of accounts for each customer test data record based on the account type reference file; and (6) generating a random number of transactions for each account based on the transaction type reference file.

Systems and methods are provided for blocking charges from a merchant to a payment account of a user. An exemplary system may include one or more memory devices storing instructions and one or more processors configured to execute the instructions to perform various operations. The operations may include receiving, from the user, a dispute request to dispute a charge to the payment account applied by the merchant. In response to the dispute request, the operations may include determining whether to block subsequent charges applied by the merchant to the payment account, based on a history of charging activities of the merchant. Responsive to a determination to block subsequent charges, the operations may include adding the merchant to a block-charge list associated with the payment account.