Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support cooperative training of machine learning (ML) models that preserves privacy in untrusted environments using distributed executable file packages. The executable file packages may include files, libraries, scripts, and the like that enable a cloud service provider configured to provide server-side ML model training to also support cooperative ML model training with multiple clients, particularly for a fraud prediction model for financial transactions. Because the cooperative training includes the clients training respective ML models and the server aggregating the trained ML models, private client data such as financial transaction data may be used to train the fraud prediction model without exposing the client data to others. Such cooperative ML model training enables offloading of computing resource-intensive training from client devices to the server and may train a more robust fraud detection model.

Systems and methods for generating one or more comparative features for machine learning based classification are disclosed. A system may be configured to obtain time series data and forecast one or more predicted values based on the time series data. The system may also be configured, for each predicted value of the one or more predicted values, to compare an actual value of the time series data to the predicted value and generate a comparative value of a comparative feature based on the comparison. The comparative feature is to be provided to a machine learning model for a classification task associated with the time series data. The classification task may include determining whether one or more data values in the time series data is fraudulent based on the comparative feature.

A method includes receiving, by a marketplace server, a sharing request regarding an exchange item from a first computing device, where the sharing request indicates a sharing approach for sharing the exchange item with one or more other computing devices and includes dynamic exchange item information generated by the first computing device. The method further includes verifying, by the marketplace server, the dynamic exchange item information received from the first computing device. When the verification is favorable, the method includes generating shared exchange item security parameters for the one or more other computing devices, generating shared exchange item information in accordance with the sharing approach, and sending the shared exchange item security parameters and the shared exchange item information to the one or more other computing devices for facilitating utilization of the exchange item by the one or more other computing devices in accordance with the shared approach.

A method executed by a computing entity includes interpreting digital records to produce a first digital record representing a first contingent asset. The method further includes interpreting a first authenticity indicator associated with the first digital record to produce a first contingent asset risk level. When the first contingent asset risk level is greater than a contingency risk threshold level, the method further includes establishing a set of first contingent asset available terms for a corresponding set of portions of the first contingent asset, generating a set of first smart contracts to represent the set of portions to include the set of first contingent asset available terms and a contingent status. The method further includes causing generation of a non-fungible token to represent the set of first smart contracts in an object distributed ledger.

The present subject matter relates to a system and method of operating one or more self-checkout (SCO) terminals of a SCO environment. The system comprises one or more video sensors configured to capture a plurality of video frames. The video frames are processed by a processing unit to detect a primary subject of interest and a second subject of interest post detection of the primary subject of interest. Further, change in location and time of appearance of the primary subject of interest and the secondary subject of interest is determined, which generates a motion trigger. Based on the motion trigger, a transaction data is received which is compared with the detected secondary subject of interest. A non-scan event alert is generated based on a mismatch in the comparison of the transaction data and the detected one or more secondary subject of interest.

Techniques are disclosed in which a computer system receives, from a plurality of user computing devices, a plurality of device-trained models and obfuscated sets of user data stored at the plurality of user computing devices, where the device-trained models are trained at respective ones of the plurality of user computing devices using respective sets of user data prior to obfuscation. In some embodiments, the server computer system determines similarity scores for the plurality of device-trained models, wherein the similarity scores are determined based on a performance of the device-trained models. In some embodiments, the server computer system identifies, based on the similarity scores, at least one of the plurality of device-trained models as a low-performance model. In some embodiments, the server computer system transmits, to the user computing device corresponding to the low-performance model, an updated model.

Transaction ID information corresponding to proof certificate-verifying transaction information is transmitted to a block chain retention server if a request for proof certificate information is sensed, when the proof certificate-verifying transaction information generated by using the proof certificate information, to be provided to a customer, is recorded in a block chain retention server and the transaction ID information is managed. The proof certificate-verifying transaction information corresponding to the transaction ID information is acquired from the block chain retention server. A proof certificate index hash value used for comparison, acquired from the proof certificate information to be provided to a customer and corresponding to a request, is compared with a proof certificate-verifying index hash value acquired from the proof certificate verifying-transaction information. Verification information generated with reference to the comparison result of the proof certificate index hash value used for comparison and the proof certificate-verifying index hash value are provided.

Systems and methods for performing a transaction with a headless point-of-sale or automated teller machine (ATM) device are disclosed using a card having a second communications path to a financial services provider. A card having a display and radio frequency (RF) communications module may be authenticated with a headless point-of-sale device using a short-range RF communications link. Characteristics of the card may be set prior to the transaction. Transaction information may be provided to the display of the card from the headless point-of-sale device. A customer may confirm the transaction at the card using a touch-sensitive input area. During the processing, a communication may be made over the second communications path to authorize the transaction independently of the transaction processing path. A transaction may then be completed at the headless point-of-sale device.

One embodiment provides a method and system for managing risk-control commands. During operation, the system can obtain statistics associated with a plurality of risk-control commands issued by a risk-control system corresponding to a plurality of transactions on an online financial platform, and determine, based on the monitored plurality of risk-control commands, whether a subset of risk-control commands meets an anomaly condition. In response to determining that the subset of risk-control commands does not meet the anomaly condition, the system can forward the subset of risk-control commands to the online financial platform to facilitate the online financial platform in performing corresponding transactions according to the subset of risk-control command. In response to determining that the subset of risk-control commands meets the anomaly condition, the system can prevent the subset of risk-control commands from being forwarded to the online financial platform.

A device receives first transaction information associated with a first transaction, and a first transaction account utilized for the first transaction and associated with a first financial institution. The device determines, based on a fraud model, that the first transaction is to be denied due to potential fraud associated with the first transaction account and receives second transaction information associated with a second transaction, and a second transaction account utilized for the second transaction and associated with a second financial institution. The device processes the first transaction information and the second transaction information, with a matching model, to determine whether the first transaction information matches the second transaction information and determines that the first transaction was incorrectly denied when the first transaction information matches the second transaction information within a predetermined threshold. The device performs one or more actions based on determining that the first transaction was incorrectly denied.