US 11,694,205 B2
Systems and methods for data breach detection using virtual card numbers
Jacob Learned, Brooklyn, NY (US); Michael Saia, New York, NY (US); Max Miracolo, Brooklyn, NY (US); and Kaylyn Gibilterra, New York, NY (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Jan. 29, 2021, as Appl. No. 17/161,732.
Prior Publication US 2022/0245640 A1, Aug. 4, 2022
Int. Cl. G06Q 20/40 (2012.01); G06N 3/08 (2023.01); G06Q 20/34 (2012.01); H04L 9/40 (2022.01)
CPC G06Q 20/4016 (2013.01) [G06N 3/08 (2013.01); G06Q 20/351 (2013.01); H04L 63/1416 (2013.01); H04L 63/1425 (2013.01)] 20 Claims
OG exemplary drawing
 
20. A computer-implemented method for training and using a recurrent neural network for data breach identification, the method comprising:
generating, by one or more processors, a plurality of virtual card numbers, wherein each one of the plurality of virtual card numbers is associated with a user device, a provider device, and security data to form a virtual card number data set;
storing, by the one or more processors, one or more of the virtual card number data set on a first database;
receiving, by the one or more processors, one or more compromised virtual card number data sets, wherein the one or more compromised virtual card number data sets is parsed from compromised data stored on a second database isolated from communication with the first database, and wherein the compromised data is obtained from a scan of unindexed websites on a network;
comparing, by the one or more processors, the one or more compromised virtual card number data sets with the one or more virtual card number data sets stored on the first database;
determining, by the one or more processors, one of the one or more of the virtual card number data sets has been compromised based on the comparison and whether a pre-determined threshold has been met;
for each of the one or more compromised virtual card number data sets, training, by the one or more processors, the recurrent neural network to associate the compromised virtual card number data set with one or more sequential patterns found within the compromised virtual card number data set, to generate a trained recurrent neural network;
receiving, by the one or more processors, a first virtual card number data set from the first database;
determining, using the trained recurrent neural network, by the one or more processors, whether the first virtual card number data set is a compromised virtual card number data set by detecting at least one of the one or more sequential patterns found within the compromised virtual card number data set;
upon determining the first virtual card number data set is a compromised virtual card number data set, automatically regenerating a virtual card number for the first virtual card number data set; and
transmitting, by the one or more processors, the regenerated virtual card number and a message to a user or provider device associated with the first virtual card number data set indicating the first virtual card number data set is compromised;
receiving, by the one or more processors, a request to authenticate a transaction; and
declining to authenticate the transaction upon determining, by the one or more processors, that the transaction involves a compromised virtual card number data set.