Systems and methods for private authentication with helper networks
A set of measurable encrypted feature vectors can be derived from any biometric data and/or physical or logical user behavioral data, and then using an associated deep neural network (“DNN”) on the output (i.e., biometric feature vector and/or behavioral feature vectors, etc.) an authentication system can determine matches or execute searches on encrypted data. Behavioral or biometric encrypted feature vectors can be stored and/or used in conjunction with respective classifications, or in subsequent comparisons without fear of compromising the original data. In various embodiments, the original behavioral and/or biometric data is discarded responsive to generating the encrypted vectors. In other embodiment, helper networks can be used to filter identification inputs to improve the accuracy of the models that use encrypted inputs for classification.
1. A system for privacy-enabled identification or authentication, the system comprising:
at least one processor operatively connected to a memory;
an authentication data gateway, executed by the at least one processor, configured to filter identification information, the authentication data gateway comprising at least:
a first pre-trained validation helper network associated with a first authentication modality configured to:
process identification information of a first authentication type, and
validate the identification information of the first authentication type or reject the identification information of the first authentication type preventing use of rejected data in training other neural networks for enrollment, identification, or authentication, wherein the first pre-trained validation helper network is trained to validate or reject based on evaluation criteria independent of an identity of a subject of the identification information seeking to be enrolled, identified, or authenticated; and
an authentication subsystem configured to:
generate with an embedding network associated with the first authentication type an output of encrypted feature vectors for identifying or authenticating respective users based on input of the filtered identification information to the embedding network; and
train a classification network on the encrypted feature vectors of the first authentication type and labels associated with the respective users to identify or authenticate respective users based on input of the encrypted feature vectors to the classification network.
2. The system of claim 1 , wherein the authentication data gateway is configured to filter bad authentication data from training data sets used to build embedding network models.
3. The system of claim 1 , wherein the authentication data gateway further comprises a plurality of validation helper networks each associated with a respective type of identification information, wherein each of the plurality of validation helper networks generate an evaluation of respective authentication inputs to establish a probability of validity.
4. The system of claim 3 , wherein the first pre-trained validation helper network is configured process an image input as identification information, and output a probability that the image input is invalid.
5. The system of claim 3 , wherein the first pre-trained validation helper network is configured to process an image input as identification information, and output a probability that the image input is a presentation attack.
6. The system of claim 3 , wherein the first pre-trained validation helper network is configured to:
process a video input as identification information; and
output a probability that the video input is invalid.
7. The system of claim 4 , wherein the first pre-trained validation helper network is configured to:
process a video input as identification information, and
output a probability that the video input is a presentation attack.
8. The system of claim 1 , wherein the authentication data gateway is configured to determine if an identification information input improves training set entropy as part of filtering.
9. The system of claim 1 , wherein the authentication data gateway further comprises:
a first pre-trained geometry helper network configured to:
process identification information of a first type,
accept as input plaintext identification information of the first type, and
output processed identification information of the first type.
10. The system of claim 9 , wherein the first pre-trained validation helper network is configured to accept the output of the first pre-trained geometry helper neural network.
11. A computer implemented method for privacy-enabled authentication, the method comprising:
filtering, by at least one processor, invalid identification information;
executing by the at least one processor, a first pre-trained validation helper network;
accepting, by the first pre-trained validation helper network, identification information of a first type;
validating the identification information of the first authentication type or rejecting the identification information of the first authentication type preventing use of rejected data in training other neural networks for enrollment, identification, or authentication, wherein the first pre-trained validation helper network is trained to validate or reject based on evaluation criteria independent of an identity of a subject of the identification information seeking to be enrolled, identified, or authenticated;
generating with an embedding network associated with the first authentication type encrypted feature vectors for identifying or authenticating respective users based on input of filtered identification information to the embedding network; and
training a classification network on the encrypted feature vectors of the first authentication type and labels associated with the respective users to identify or authenticate respective users based on input of the encrypted feature vectors to the classification network.
12. The method of claim 11 , wherein the method further comprises filtering bad authentication data.
13. The method of claim 11 , wherein the method further comprises executing a plurality of validation helper networks each associated with a respective type of identification information, and generating a binary evaluation of respective authentication inputs by respective ones of the plurality of validation helper networks to establish validity.
14. The method of claim 11 , wherein the method further comprises executing at least a first geometry helper network a configured to process identification information of the first type.
15. The method of claim 14 , wherein the method further comprises processing, by the first pre-trained validation helper network an image input as identification information, and output a probability that the image input is invalid.
16. The method of claim 11 , wherein the method further comprises processing an image input as identification information, and generating a probability that the image input is a presentation attack, by the first pre-trained validation helper network.
17. The method of claim 11 , wherein the method further comprises
processing, by the first pre-trained validation helper network, a video input as identification information; and
generating, by the first pre-trained validation helper network, a probability that the video input is invalid, by the first pre-trained validation helper network.
18. The method of claim 11 , wherein the method further comprises
processing, by the first pre-trained validation helper network, a video input as identification information, and
generating, the first pre-trained validation helper network, a probability that the video input is a presentation attack.
19. The method of claim 11 , wherein the method further comprises determining if an identification information input improves training set entropy.