Official implementation of Improving the Transferability of Adversarial Examples by Feature Augmentation

This work has been accepted by IEEE Transactions on Neural Networks and Learning Systems, and the paper is here.

Run faug attack

python faug_attack.py [optional parameters]

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
configs		configs
models		models
transferattack		transferattack
LICENSE		LICENSE
README.md		README.md
attack_command.py		attack_command.py
baseline_adversarial_example.py		baseline_adversarial_example.py
baseline_adversarial_example_ens_faug.py		baseline_adversarial_example_ens_faug.py
baseline_transfer_analysis.py		baseline_transfer_analysis.py
config.py		config.py
contributors.md		contributors.md
data_analysis.py		data_analysis.py
data_loader.py		data_loader.py
faug_attack.py		faug_attack.py
main.py		main.py
process_stl10.py		process_stl10.py
purify_nullify_folder.py		purify_nullify_folder.py
requirements.txt		requirements.txt