Adversarial robustness for latent models: Revisiting the robust-standard accuracies tradeoff

Open in new window