apoorva1225/adversarial-robustness-toolbox

Python library for adversarial machine learning, attacks and defences for neural networks, logistic regression, decision trees, SVM, gradient boosted trees, Gaussian processes and more with multiple framework support

Adversarial Robustness 360 Toolbox (ART v0.10.0)

This is a library dedicated to adversarial machine learning. Its purpose is to allow rapid crafting and analysis of attacks and defense methods for machine learning models. ART provides an implementation for many state-of-the-art methods for attacking and defending classifiers.

Documentation for ART: https://adversarial-robustness-toolbox.readthedocs.io

The library is still under development. Feedback, bug reports and extensions are highly appreciated. Get in touch with us on Slack (invite here)!

We use Github issues for tracking bugs and feature requests. Please check the release notes for fixed bugs in earlier versions and new features.

[Important notice] ART v0.8 and v0.9 contain for certain settings a bug in the Basic Iterative Method (BIM) and Projected Gradient Descent (PGD) attacks, which has been fixed in ART v0.10. To be sure, please update ART to v0.10, especially if using BIM or PGD attack.

Supported attacks, defences and metrics

The library contains implementations of the following evasion attacks:

• DeepFool (Moosavi-Dezfooli et al., 2015)
• Fast gradient method (Goodfellow et al., 2014)
• Basic iterative method (Kurakin et al., 2016)
• Projected gradient descent (Madry et al., 2017)
• Jacobian saliency map (Papernot et al., 2016)
• Universal perturbation (Moosavi-Dezfooli et al., 2016)
• Virtual adversarial method (Miyato et al., 2015)
• C&W L_2 and L_inf attacks (Carlini and Wagner, 2016)
• NewtonFool (Jang et al., 2017)
• Elastic net attack (Chen et al., 2017)
• Spatial transformations attack (Engstrom et al., 2017)
• Query-efficient black-box attack (Ilyas et al., 2017)
• Zeroth-order optimization attack (Chen et al., 2017)
• Decision-based attack (Brendel et al., 2018)
• Adversarial patch (Brown et al., 2017)
• HopSkipJump attack (Chen et al., 2017)

The following defence methods are also supported:

• Feature squeezing (Xu et al., 2017)
• Spatial smoothing (Xu et al., 2017)
• Label smoothing (Warde-Farley and Goodfellow, 2016)
• Adversarial training (Szegedy et al., 2013)
• Virtual adversarial training (Miyato et al., 2015)
• Gaussian data augmentation (Zantedeschi et al., 2017)
• Thermometer encoding (Buckman et al., 2018)
• Total variance minimization (Guo et al., 2018)
• JPEG compression (Dziugaite et al., 2016)
• PixelDefend (Song et al., 2017)

ART also implements detection methods of adversarial samples:

• Basic detector based on inputs
• Detector trained on the activations of a specific layer
• Detector based on Fast Generalized Subset Scan (Speakman et al., 2018)

The following detector of poisoning attacks is also supported:

• Detector based on activations analysis (Chen et al., 2018)

Robustness metrics:

• CLEVER (Weng et al., 2018)
• Empirical robustness (Moosavi-Dezfooli et al., 2015)
• Loss sensitivity (Arpit et al., 2017)

Setup

Installation with pip

The toolbox is designed and tested to run with Python 3. ART can be installed from the PyPi repository using pip:

pip install adversarial-robustness-toolbox

Manual installation

For the most recent version of the library, either download the source code or clone the repository in your directory of choice:

git clone https://github.com/IBM/adversarial-robustness-toolbox

To install ART, do the following in the project folder:

pip install .

The library comes with a basic set of unit tests. To check your install, you can run all the unit tests by calling the test script in the install folder:

bash run_tests.sh

Running ART

Some examples of how to use ART when writing your own code can be found in the examples folder. See examples/README.md for more information about what each example does. To run an example, use the following command:

python examples/<example_name>.py

The notebooks folder contains Jupyter notebooks with detailed walkthroughs of some usage scenarios.

Contributing

Adding new features, improving documentation, fixing bugs, or writing tutorials are all examples of helpful contributions. Furthermore, if you are publishing a new attack or defense, we strongly encourage you to add it to the Adversarial Robustness Toolbox so that others may evaluate it fairly in their own work.

Bug fixes can be initiated through GitHub pull requests. When making code contributions to the Adversarial Robustness Toolbox, we ask that you follow the PEP 8 coding standard and that you provide unit tests for the new features.

This project uses DCO. Be sure to sign off your commits using the -s flag or adding Signed-off-By: Name<Email> in the commit message.

Example

git commit -s -m 'Add new feature'

Citing ART

If you use ART for research, please consider citing the following reference paper:

@article{art2018,
    title = {Adversarial Robustness Toolbox v0.10.0},
    author = {Nicolae, Maria-Irina and Sinn, Mathieu and Tran, Minh~Ngoc and Buesser, Beat and Rawat, Ambrish and Wistuba, Martin and Zantedeschi, Valentina and Baracaldo, Nathalie and Chen, Bryant and Ludwig, Heiko and Molloy, Ian and Edwards, Ben},
    journal = {CoRR},
    volume = {1807.01069}
    year = {2018},
    url = {https://arxiv.org/pdf/1807.01069}
}