kasozivincent / Secure-Deep-Learning

This repository contains code that implements secure deep learning models robust to adversarial samples

Secure-Deep-Learning

This repository contains code that implements secure deep learning models robust to adversarial samples

The models folder

This folder contains the different models that we built during the project. We tried different architectures to check which were more resilient to adversarial attacks

The defenses folder

This folder contains the defense mechanisms we used to defend the models stored in models folder against selected adversarial attacks.

The attacks folder

This folder contains the different attack strategies that we launched to the different models stored in the models folder.