US 11,902,176 B2Granted

System for detecting multi-dimensional linkage and layering of resource transfers using a distributed server network

AssigneeBank of America Corporation

InventorsRajalakshmi Arumugam, Mahesh Ganesan, Sreenivas Chintada

Life of this patent

Estimating the term…

Loading patent…

Abstract

A system is provided for detecting multi-dimensional linkage and layering of resource transfers using a distributed server network. In particular, the system may comprise a plurality of distributed server nodes that each host a copy of a distributed register, where each of the nodes may be operated by an entity. Each distributed server node may submit, to the distributed register, data records that may contain data regarding potential unauthorized users, accounts, and/or resource transfers. Based on the information within the distributed register, along with various other data inputs, the system may use a deep learning-based graph processing algorithm to identify a multi-dimensional linkage between the users, accounts, and/or resource transfers to extract hidden relationships and potentially unauthorized activity.

Claims20

Dependent claims

A system for detecting multi-dimensional linkage and layering of resource transfers using a distributed server network, the system comprising:

a memory device with computer-readable program code stored thereon;
a communication device;
and
a processing device operatively coupled to the memory device and the communication device, wherein the processing device is configured to execute the computer-readable program code to:
retrieve a set of resource transfer data from a distributed register, wherein the set of resource transfer data comprises data submitted by one or more entities regarding potentially unauthorized users, accounts, and resource transfers;
based on the set of resource transfer data, execute a deep learning-based machine learning algorithm to identify multi-dimensional linkages between the potentially unauthorized users, accounts, and resource transfers, wherein identifying the multi-dimensional linkages comprises:
detecting a resource transfer between a first potentially unauthorized user and a second potentially unauthorized user;
and
based on detecting the resource transfer, forming a multi-dimensional linkage between the first potentially unauthorized user and the second potentially unauthorized user, wherein the multi-dimensional linkage comprises information regarding a direction of the resource transfer, a resource type associated with the resource transfer, and geo-spatial information associated with the resource transfer;
execute time series based tracking of the potentially unauthorized users, accounts, and resource transfers;
identify, within the set of resource transfer data, one or more layering patterns associated with a conversion of resources from a first format to a second format;
and
identify one or more source agents, one or more layering agents, and one or more destination agents from the one or more potentially unauthorized users, accounts, and resource transfers, wherein identifying the one or more destination agents comprises uncovering, using the machine learning algorithm, a hidden recipient of an unauthorized resource transfer, wherein the hidden recipient was not included in the set of resource transfer data.

10 more elements

A computer program product for detecting multi-dimensional linkage and layering of resource transfers using a distributed server network, the computer program product comprising at least one non-transitory computer readable medium having computer-readable program code portions embodied therein, the computer-readable program code portions comprising executable portions for:

retrieving a set of resource transfer data from a distributed register, wherein the set of resource transfer data comprises data submitted by one or more entities regarding potentially unauthorized users, accounts, and resource transfers;
based on the set of resource transfer data, executing a deep learning-based machine learning algorithm to identify multi-dimensional linkages between the potentially unauthorized users, accounts, and resource transfers, wherein identifying the multi-dimensional linkages comprises:
detecting a resource transfer between a first potentially unauthorized user and a second potentially unauthorized user;
and
based on detecting the resource transfer, forming a multi-dimensional linkage between the first potentially unauthorized user and the second potentially unauthorized user, wherein the multi-dimensional linkage comprises information regarding a direction of the resource transfer, a resource type associated with the resource transfer, and geo-spatial information associated with the resource transfer;
executing time series based tracking of the potentially unauthorized users, accounts, and resource transfers;
identifying, within the set of resource transfer data, one or more layering patterns associated with a conversion of resources from a first format to a second format;
and
identifying one or more source agents, one or more layering agents, and one or more destination agents from the one or more potentially unauthorized users, accounts, and resource transfers, wherein identifying the one or more destination agents comprises uncovering, using the machine learning algorithm, a hidden recipient of an unauthorized resource transfer, wherein the hidden recipient was not included in the set of resource transfer data.

6 more elements

A computer-implemented method for detecting multi-dimensional linkage and layering of resource transfers using a distributed server network, the computer-implemented method comprising:

retrieving a set of resource transfer data from a distributed register, wherein the set of resource transfer data comprises data submitted by one or more entities regarding potentially unauthorized users, accounts, and resource transfers;
based on the set of resource transfer data, executing a deep learning-based machine learning algorithm to identify multi-dimensional linkages between the potentially unauthorized users, accounts, and resource transfers, wherein identifying the multi-dimensional linkages comprises:
detecting a resource transfer between a first potentially unauthorized user and a second potentially unauthorized user;
and
based on detecting the resource transfer, forming a multi-dimensional linkage between the first potentially unauthorized user and the second potentially unauthorized user, wherein the multi-dimensional linkage comprises information regarding a direction of the resource transfer, a resource type associated with the resource transfer, and geo-spatial information associated with the resource transfer;
executing time series based tracking of the potentially unauthorized users, accounts, and resource transfers;
identifying, within the set of resource transfer data, one or more layering patterns associated with a conversion of resources from a first format to a second format;
and
identifying one or more source agents, one or more layering agents, and one or more destination agents from the one or more potentially unauthorized users, accounts, and resource transfers, wherein identifying the one or more destination agents comprises uncovering, using the machine learning algorithm, a hidden recipient of an unauthorized resource transfer, wherein the hidden recipient was not included in the set of resource transfer data.

6 more elements

Dependencies are inferred from the claim text.