System and method for zero-knowledge facial recognition
AssigneeRoyal Bank of Canada
InventorsArya POURTABATABAIE, Joseph CHOI
Life of this patent
Estimating the term…
Abstract
A computer implemented system for electronic verification of credentials including at least one processor and data storage is described in various embodiments. The approach describes the architecture and use of a “max gate,” which can be used as a structural component of the neural network.
Claims20
A computer system for conducting a zero-knowledge proof protocol:
- a processor coupled to computer memory and non-transitory computer readable media, the processor configured to:
- receive an input data set that is provided to one or more computing data object nodes of an input layer;
- process the input layer using one or more computing data object nodes of at least one hidden layer, each hidden layer including at least one max gate used as an activation function, the at least one max gate configured such that given a representation of a commitment to values x and y, obtain a commitment to a value c that is 1 if y>x, and 0 otherwise, such that a maximum of x and y, m, satisfies one of the following relations:
- (1) m=x if c=0 and y otherwise;
- or
- (2) m=x+(y−x)c;
- generate one or more outputs based on the processing of the input layer using the one or more computing data object nodes of the hidden layer(s), the one or more outputs utilized to validate the input data set in accordance with the zero-knowledge proof protocol.
A method for conducting a zero-knowledge proof protocol, the method comprising receiving an input data set that is provided to one or more computing data object nodes of an input layer;
- processing the input layer using one or more computing data object nodes of at least one hidden layer, each hidden layer including at least one max gate used as an activation function, the at least one max gate configured such that given a representation of a commitment to values x and y, obtain a commitment to a value c that is 1 if y>x, and 0 otherwise, such that a maximum of x and y, m, satisfies one of the following relations:
- (1) m=x if c=0 and y otherwise;
- or
- (2) m=x+(y−x)c;
- generate one or more outputs based on the processing of the input layer using the one or more computing data object nodes of the hidden layer(s), the one or more outputs utilized to validate the input data set in accordance with the zero-knowledge proof protocol.
A non-transitory computer readable medium storing machine interpretable instructions, which when executed by a processor, cause the processor to perform a method for conducting a zero-knowledge proof protocol, the method comprising receiving an input data set that is provided to one or more computing data object nodes of an input layer;
- processing the input layer using one or more computing data object nodes of at least one hidden layer, each hidden layer including at least one max gate used as an activation function, the at least one max gate configured such that given a representation of a commitment to values x and y, obtain a commitment to a value c that is 1 if y>x, and 0 otherwise, such that a maximum of x and y, m, satisfies one of the following relations:
- (1) m=x if c=0 and y otherwise;
- or
- (2) m=x+(y−x) c;
- generate one or more outputs based on the processing of the input layer using the one or more computing data object nodes of the hidden layer(s), the one or more outputs utilized to validate the input data set in accordance with the zero-knowledge proof protocol.
Dependencies are inferred from the claim text.