August 2019
Intermediate to advanced
342 pages
9h 35m
English
The black-box adversarial attack strategy can also be validly used to deceive the next-generation antimalware systems, based on NNs.
A useful library for developing black-box adversarial attacks with malware examples is MalGAN, available for download at https://github.com/yanminglai/Malware-GAN/, and released under the GPL 3.0 license (https://github.com/yanminglai/Malware-GAN/blob/master/LICENSE). The fundamental idea behind MalGAN is to use a GAN to generate adversarial malware examples, which are able to bypass black-box machine-learning-based detection models. To install the MalGAN library, you need to install the TensorFlow 1.80, Keras 2.0, and Cuckoo Sandbox 2.03 (https://cuckoo.readthedocs.io/en/2.0.3/ ...