| CPC G06Q 50/265 (2013.01) | 21 Claims |
|
1. An electronic identity security method comprising:
receiving, by a processor, a request for identity verification from a device, the request including an identifier of an individual seeking identity verification;
accessing, by the processor, data associated with the individual seeking identity verification stored in a storage device, wherein the data associated with the individual seeking identity verification stored in the storage device comprises known facts about the individual seeking identity verification, wherein a known fact is data stored about the individual seeking identity verification stored in the storage device;
inferring, by the processor, derived facts about the individual by determining associations between multiple known facts about the individual stored in the storage device using a recurrent neural network, the derived facts being different than the known facts and not stored as known facts in the storage device, the derived facts being about the individual, and the derived facts being determined by analyzing the multiple known facts about the individual and determining associations between the multiple known facts about the individual;
inferring, by the processor, geographical data, familial relationships, or time-based factors about the individual based on multiple known facts stored in the storage device using the recurrent neural network;
generating, by the processor, at least one first degree intelligence question based on the known facts;
generating, by the processor, at least one second degree intelligence question based on the known facts and the derived facts;
generating, by the processor, at least one third degree intelligence question based on the known facts, the derived facts, and the geographical data, familial relationships, or time-based factors;
evaluating, by the processor, at least one received answer to at least one identity verification question selected from the at least one first degree intelligence question, the at least one second degree intelligence question, and the at least one third degree intelligence question to determine whether the individual answered the at least one identity verification question correctly; and
verifying, by the processor, the individual's identity based on at least one received answer to the at least one identity verification question,
wherein the recurrent neural network infers the derived facts about the individual by processing the multiple known facts as inputs in an input layer of the recurrent neural network, and the multiple known facts are provided to neurons of a hidden layer of the recurrent neural network to find potential connections between the known facts, and upon finding the potential connections, the neurons provide the potential connections as outputs to an output layer of the recurrent neural network, and the output layer determines a probability whether the potential connections are the derived facts.
|