US 12,481,913 B2
Automatic generation of attribute sets for counterfactual explanations
Brian Barr, McLean, VA (US); and Christopher Bruss, McLean, VA (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Sep. 1, 2021, as Appl. No. 17/464,566.
Prior Publication US 2023/0066807 A1, Mar. 2, 2023
Int. Cl. G06N 20/00 (2019.01)
CPC G06N 20/00 (2019.01) 20 Claims
OG exemplary drawing
 
1. A system for accessing a computing network, the system comprising:
one or more processors; and
one or more non-transitory computer-readable storage media storing instructions, which when executed by the one or more processors cause the one or more processors to:
receive a set of attributes associated with a user device, wherein the user device received a negative decision from an automated system based on the set of attributes in connection with a request from the user device to access the computing network,
wherein an attribute, of the set of attributes, indicates a software version of a software package used by the user device;
encode, using an encoding-decoding system, the set of attributes into a latent representation, wherein encoding the set of attributes into the latent representation comprises compressing the set of attributes into a multi-dimensional value that has less dimensions than a number of attributes in the set of attributes;
pass the latent representation from the encoding-decoding system to a data processing system;
input, using the data processing system, the latent representation into a machine learning model to obtain an associated change value for the set of attributes, wherein the machine learning model is trained, using a plurality of latent representations and a plurality of change values, to output, for a received latent representation, a corresponding change value, and wherein each change value represents a minimum change needed to be made to the set of attributes to change the negative decision to a positive decision;
apply, using the data processing system, the associated change value to one or more dimensions of the latent representation to generate an updated latent representation;
pass the updated latent representation from the data processing system to the encoding-decoding system;
decode, using the encoding-decoding system, the latent representation into an updated set of attributes,
wherein an updated attribute, of the updated set of attributes, indicates an updated software version of the software package; and
transmit, to the user device, the updated software version of the software package for installation on the user device for the user device to access the computing network.