| CPC G06F 21/32 (2013.01) | 7 Claims |
|
1. An authorization system comprising:
at least one memory storing instructions; and
at least one processor configured to execute the instructions to:
calculate N features for first data and N features for second data by using N feature functions for obtaining a feature for data, the first data representing a biological information of an authorization subject and the second data representing biological information of users;
calculate a similarity degree between the first data and the second data based on the N features for the first data and the N features for the second data; and
authorize the authorization subject to access a computing system or a location pursuant to permissions provided in an account of an identified user, wherein the calculated similarity degree between the first data of the authorization subject and the second data of the identified user is highest and more than a predetermined authorization threshold value,
wherein when data is input into the N feature functions, values of N features obtained from each feature function are different from each other,
wherein the N feature functions are constituted by N learned feature calculation models in a model network that has a first calculation unit having the N feature calculation models and a second calculation unit having a model with a calculation result of the first calculation unit as an input, and
wherein the model network is trained using a learning data set in which data is set as an input sample and a one-hot vector indicating a class, among a plurality of classes, to which the data belongs is set as an output sample, and the model network is trained to output a vector indicating a posterior probability of the class to which the data belongs based on the input sample.
|