An information processing device comprising: one or more processors; and a memory including one or more programs, wherein the one or more programs causes, when executed by the one or more processors, the information processing device to execute: training a generator neural network, which is a generative model that generates data belonging to a designated class on the basis of a noise vector and the class and a discriminator neural network, which discriminates whether the data generated by the generator is real data or not, by executing a learning algorithm of a condi-tional generative adversarial network and determining weight parameters of the trained generator neural net-work; training a classification model that classifies input data based on whether the input data is in the designated class; verifying whether a property is satisfied when the classi-fication model classifies an output of the generative model by executing a formal verification algorithm wherein the formal verification algorithm guarantees that all possible data within a range of a certain norm of a noise vector belong to the designated class; and generating training data belonging to the designated class by inputting a randomly generated noise vector within the verified range of the certain norm and the desig-nated class into the generative model using the weight parameters of the trained generator neural network, and wherein the property is that, within the range of the certain norm of the noise vector input to the generative model that uses the weight parameters of the trained generator neural network, the generative model does not generate data classified into a class different from a first class designated for the generative model.

An information processing device comprising: one or more processors; and a memory including one or more programs, wherein the one or more programs causes, when executed by the one or more processors, the information processing device to execute: training a generator neural network, which is a generative model that generates data belonging to a designated class on the basis of a noise vector and the class and a discriminator neural network, which discriminates whether the data generated by the generator is real data or not, by executing a learning algorithm of a condi-tional generative adversarial network and determining weight parameters of the trained generator neural net-work; training a classification model that classifies input data based on whether the input data is in the designated class; verifying whether a property is satisfied when the classi-fication model classifies an output of the generative model by executing a formal verification algorithm wherein the formal verification algorithm guarantees that all possible data within a range of a certain norm of a noise vector belong to the designated class; and generating training data belonging to the designated class by inputting a randomly generated noise vector within the verified range of the certain norm and the desig-nated class into the generative model using the weight parameters of the trained generator neural network, and wherein the property is that, within the range of the certain norm of the noise vector input to the generative model that uses the weight parameters of the trained generator neural network, the generative model does not generate data classified into a class different from a first class designated for the generative model.

An information processing device comprising: one or more processors; and a memory including one or more programs, wherein the one or more programs causes, when executed by the one or more processors, the information processing device to execute: training a generator neural network, which is a generative model that generates data belonging to a designated class on the basis of a noise vector and the class and a discriminator neural network, which discriminates whether the data generated by the generator is real data or not, by executing a learning algorithm of a condi-tional generative adversarial network and determining weight parameters of the trained generator neural net-work; training a classification model that classifies input data based on whether the input data is in the designated class; verifying whether a property is satisfied when the classi-fication model classifies an output of the generative model by executing a formal verification algorithm wherein the formal verification algorithm guarantees that all possible data within a range of a certain norm of a noise vector belong to the designated class; and generating training data belonging to the designated class by inputting a randomly generated noise vector within the verified range of the certain norm and the desig-nated class into the generative model using the weight parameters of the trained generator neural network, and wherein the property is that, within the range of the certain norm of the noise vector input to the generative model that uses the weight parameters of the trained generator neural network, the generative model does not generate data classified into a class different from a first class designated for the generative model.

An information processing device comprising: one or more processors; and a memory including one or more programs, wherein the one or more programs causes, when executed by the one or more processors, the information processing device to execute: training a generator neural network, which is a generative model that generates data belonging to a designated class on the basis of a noise vector and the class and a discriminator neural network, which discriminates whether the data generated by the generator is real data or not, by executing a learning algorithm of a condi-tional generative adversarial network and determining weight parameters of the trained generator neural net-work; training a classification model that classifies input data based on whether the input data is in the designated class; verifying whether a property is satisfied when the classi-fication model classifies an output of the generative model by executing a formal verification algorithm wherein the formal verification algorithm guarantees that all possible data within a range of a certain norm of a noise vector belong to the designated class; and generating training data belonging to the designated class by inputting a randomly generated noise vector within the verified range of the certain norm and the desig-nated class into the generative model using the weight parameters of the trained generator neural network, and wherein the property is that, within the range of the certain norm of the noise vector input to the generative model that uses the weight parameters of the trained generator neural network, the generative model does not generate data classified into a class different from a first class designated for the generative model.