Code corresponding to the paper, "Beware the Black-Box: on the Robustness of Recent Defenses to Adversarial Examples": https://www.mdpi.com/1099-4300/23/10/1359
We provide code for the adaptive black-box attack and pure black-box attack in TensorFlow. We currently do not give pre-trained defense models. All code was originally tested with Python 3.5, TensorFlow version 1.17 and CleverHans 3.0. For those interested in a PyTorch implementation of the adaptive black-box attack please see the following repository: https://github.com/MetaMain/ViTRobust
For questions or issues please contact: kaleel.mahmood@uconn.edu