PRIVACY PRESERVING METHOD BASED ON NEURAL NETWORK AND DATA PROCESSING APPARATUS

摘要

The information protection method using a neural network model includes the steps of: a first data processing device generating initial composite data anonymized source data; a second data processing device receiving the initial composite data; and the second data processing device generating data and inputting the initial synthesized data to a generator model for generating , and using data output from the generator model as final synthesized data. The generator model is trained in advance by receiving reverse transcription information so that a target classifier that classifies class information for the final synthetic data outputs original class information.