Heterogeneous Graph Embedding Malware Detection.
for train model, run:
python3 train.py --view multi
for adversarial attack, run:
python3 attack.py jsma 50 --view multi
The repository is organised as follows:
train.py
entry script for training model.attack.py
entry script for adversarial attack.setting.py
config file.layers/
define layers used in modelsmodel_zoo/
define modelsdata/
data pathadv/
adversarial attacks. jsma and fgsm