Improving black-box optimization in VAE latent space using decoder uncertainty

Open in new window