foolbox 0.3.3

Python toolbox to create adversarial examples that fool neural networks

Foolbox

Foolbox is a Python toolbox to create adversarial examples that fool neural networks. It requires Python 3, NumPy and SciPy.

Installation

pip install foolbox

Documentation

Documentation is available on readthedocs: http://foolbox.readthedocs.io/

Example

import foolbox
import keras
from keras.applications.resnet50 import ResNet50, preprocess_input

# instantiate model
keras.backend.set_learning_phase(0)
kmodel = ResNet50(weights='imagenet')
fmodel = foolbox.models.KerasModel(kmodel, bounds=(0, 255), preprocess_fn=preprocess_input)

# get source image and label
image, label = foolbox.utils.imagenet_example()

# apply attack on source image
attack  = foolbox.attacks.FGSM(fmodel)
adv_img = attack(image=image, label=label)

Interfaces for a range of other deeplearning packages such as TensorFlow, PyTorch and Lasagne are available, e.g.

model = foolbox.models.PyTorchModel(torchmodel)

Different adversarial criteria such as Top-k, specific target classes or target probability levels can be passed to the attack, e.g.

criterion = foolbox.criteria.TargetClass(22)
attack    = foolbox.attacks.FGSM(fmodel, criterion)

Development

Foolbox is a work in progress and any input is welcome.

Authors

• Jonas Rauber

• Wieland Brendel