| CPC G06F 21/30 (2013.01) [G06N 20/00 (2019.01); H04L 9/32 (2013.01); H04L 63/08 (2013.01); G06Q 20/40 (2013.01)] | 20 Claims |
|
1. A system for determining continuous identity in real-time comprising:
a computer network connection for receiving at a first time, a request to verify the identity of an individual including a plurality of partial identifiers provided with the request;
a first database on a first server storing a plurality of identification credentials for each of a plurality of individuals,
wherein different types of identification credentials may be stored and accessed within the first database for different individuals of the plurality of individuals,
wherein a plurality of the different types of identification credentials for any individual of the plurality of individuals include various types of information that are not overlapping,
wherein one or more of the different types of identification credentials for any individual of the plurality of individuals might be expired,
wherein one or more of the different types of identification credentials for any individual of the plurality of individuals might not be expired;
a second database on the first server for storing information regarding the strength of relationships between the plurality of identification credentials stored in the first database;
a parallel processing system for applying a machine learning algorithm to the plurality of partial identifiers, the plurality of identification credentials, and the strengths of relationships, to determine whether the identity of the individual can be verified, refuted, or neither,
wherein the machine learning algorithm establishes and applies weights to various of the relationships based upon training and feedback,
wherein the machine learning algorithm applies a tunable risk tolerance to the determination; and
a memory comprising computer executable instructions for transmitting at a second time, a response to the request to verify the identity,
wherein the system may be scaled by addition of further parallel processing resources, such that the rate at which the machine learning algorithm processes a plurality of individuals scales substantially linearly over time with addition of an equal number of parallel processing resources and requests to verify; and
the memory further comprises computer executable instructions for providing a precise explanation of objective parameters used by the machine learning algorithm to determine whether the identity of the individual can be verified, refuted, or neither.
|