Ensemble Adversarial Training With Pytorch

This repository contains pytorch code to reproduce results from the paper:

Ensemble Adversarial Training: Attacks and Defenses
Florian Tramèr, Alexey Kurakin, Nicolas Papernot, Dan Boneh and Patrick McDaniel
ArXiv report: https://arxiv.org/abs/1705.07204

REQUIREMENTS

The code was tested with Python 3.6.7 and Pytorch 1.0.1.

EXPERIMENTS

Training a few simple MNIST models. These are described in mnist.py.

python -m train models/modelA --type=0
python -m train models/modelB --type=1
python -m train models/modelC --type=2
python -m train models/modelD --type=3

(standard) Adversarial Training:

python -m train_adv models/modelA_adv --type=0 --epochs=12

Ensemble Adversarial Training:

python -m train_adv models/modelA_ens models/modelA models/modelC models/modelD --type=0 --epochs=12

The accuracy of the models on the MNIST test set can be computed using

python -m simple_eval test [model(s)]

To evaluate robustness to various attacks

python -m simple_eval [attack] [source_model] [target_model(s)] [--parameters (opt)]

REFERENCE

1. Author's code: ftramer/ensemble-adv-training