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

Using a variational autoencoder, assuming a normal distribution, the latent spatial distribution can be approximated. However, in this article, the goal was to test this concept with adversarial ...

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 ...

In this project, I aim to create a set of images of Kuzushiji Japanese characters via the Kuzushiji-49 dataset using 2 graphical models which are Conditional-Variational Autoencoder(C-VAE), and ...

In this study, Insilico Medicine researchers developed a new model, the Bidirectional Adversarial Autoencoder, that learns a joint distribution of molecular structures and induced transcriptional ...