Skip to content
No description, website, or topics provided.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
README.rst
adversarial_generation.py
cifar10_LA.png
cifar10_input.py
config_layer11.json
config_perturbation.json
eval.py
feature_adv_training_layer11.py
feature_model.py
fetch_model.py
imagenet_adv_LA.png
images.py
latent_adversarial_attack.py
model.py
model_robustml.py
pgd_attack.py

README.rst

LAT adversarial_robustness

A fine tuning technique over the adversarially trained models to increase further robustness

Running the code

Dataset: CIFAR10

Fetching LAT robust model

The model can be downloaded from this link - https://drive.google.com/open?id=1um2zoVYYw5YZuuV8_IeoUy-qRWSmCVUb> .

Evaluating the LAT robust model

python eval.py

The trained model achieves test accuracy of 87.8% and adversarial robustness of 53.82% against PGD attack(epsilon = 8.0/255.0)

Fetching Adversarial Trained Model

python fetch_model.py

Training via LAT

python feature_adv_training_layer11.py

Latent Attack

python latent_adversarial_attack.py

Example original and adversarial images computed via Latent Attack on CIFAR10

cifar10_LA.png

Example original and adversarial images computed via Latent Attack on Restricted Imagenet(https://arxiv.org/pdf/1805.12152.pdf.

imagenet_adv_LA.png
You can’t perform that action at this time.