Processing apparatus , learning apparatus, processing method, and nonvolatile recording medium

A learning apparatus that learns setting values of a neural network model on the basis of predetermined learning data, the apparatus comprising:

• a processor configured to execute a process including:
• learning setting values of a learning target neural network model on the basis of the learning data using the learning target neural network model and
• a learning result storage unit configured to store the setting values of the learning target neural network model learned by the processor, wherein
• the processor is configured to execute a process including:
• when forward propagation is executed, using, as the learning target neural network model, a neural network model which includes a first processing layer that executes operations with modulo operation with respect to plain text data input from a previous stage and which is configured such that a nonlinear function is not applied to an operation result of the first processing layer in a posterior stage of the first processing layer;
• and
• when backward propagation is executed to update the setting value of the learning target neural network model, using, as the learning target neural network model, a neural network model including, instead of the first processing layer, a second processing layer that performs operations excluding a modulo operation from the operations with modulo operation in the first processing layer and an activation function layer that is disposed at a posterior stage of the second processing layer to perform a process based on a differentiable activation function, and using, as input, output data output when the forward propagation is executed.

A processing method executed by a processor to control a processing apparatus,

• the processing method comprising:
• the process executed by the processor including:
• executing a predetermined process with respect to encryption data using a neural network model, wherein
• the neural network model has a processing layer that executes a process of generating second encryption data from a plurality of pieces of first encryption data input from a previous stage and outputting the second encryption data, the second encryption data providing, by being decrypted, a processing result obtained by performing operations with modulo operation with respect to plurality of pieces of plain text data corresponding to the plurality of pieces of first encryption data, the neural network model being configured such that a nonlinear function is not applied to the second encryption data in a posterior stage of the processing layer.

A computer-readable nonvolatile recording medium having stored therein a processing program for causing a processor to execute a process, the process including:

• receiving encryption data;
• and
• executing a predetermined process with respect to the encryption data using a neural network model, wherein
• the neural network model has a processing layer that executes a process of generating second encryption data from a plurality of pieces of first encryption data input from a previous stage and outputting the second encryption data, the second encryption data providing, by being decrypted, a processing result obtained by performing operations with modulo operation with respect to plurality of pieces of plain text data corresponding to the plurality of pieces of first encryption data, the neural network model being configured such that a nonlinear function is not applied to the second encryption data in a posterior stage of the processing layer.