Adversarial Robustness Tasks | State-of-the-Art