Therefore, we propose Purifying Variational AutoEncoder (PuVAE), a method to purify adversarial examples. The proposed method eliminates an adversarial perturbation by projecting an adversarial ...

Based on this method, robust internal representation in a deep Q-network (DQN) was introduced by applying adversarial noise to disturb the DQN policy; however, it was compensated for by the ...

This repository contains the code for the publication "Mastering processing-microstructure complexity through the prediction of thin film structure zone diagrams by generative ... for a conditional ...