README.md 5.39 KB
Newer Older
Beat Buesser's avatar
Beat Buesser committed
1
# Adversarial Robustness Toolbox (ART) v1.8
Irina Nicolae's avatar
Irina Nicolae committed
2
<p align="center">
Beat Buesser's avatar
Beat Buesser committed
3
  <img src="docs/images/art_lfai.png?raw=true" width="467" title="ART logo">
Irina Nicolae's avatar
Irina Nicolae committed
4
5
6
</p>
<br />

7
8
![Continuous Integration](https://github.com/Trusted-AI/adversarial-robustness-toolbox/workflows/Continuous%20Integration/badge.svg)
![CodeQL](https://github.com/Trusted-AI/adversarial-robustness-toolbox/workflows/CodeQL/badge.svg)
Beat Buesser's avatar
Beat Buesser committed
9
[![Documentation Status](https://readthedocs.org/projects/adversarial-robustness-toolbox/badge/?version=latest)](http://adversarial-robustness-toolbox.readthedocs.io/en/latest/?badge=latest)
Beat Buesser's avatar
Beat Buesser committed
10
[![PyPI](https://badge.fury.io/py/adversarial-robustness-toolbox.svg)](https://badge.fury.io/py/adversarial-robustness-toolbox)
Beat Buesser's avatar
Beat Buesser committed
11
12
[![Language grade: Python](https://img.shields.io/lgtm/grade/python/g/Trusted-AI/adversarial-robustness-toolbox.svg?logo=lgtm&logoWidth=18)](https://lgtm.com/projects/g/Trusted-AI/adversarial-robustness-toolbox/context:python)
[![Total alerts](https://img.shields.io/lgtm/alerts/g/Trusted-AI/adversarial-robustness-toolbox.svg?logo=lgtm&logoWidth=18)](https://lgtm.com/projects/g/Trusted-AI/adversarial-robustness-toolbox/alerts/)
13
[![codecov](https://codecov.io/gh/Trusted-AI/adversarial-robustness-toolbox/branch/main/graph/badge.svg)](https://codecov.io/gh/Trusted-AI/adversarial-robustness-toolbox)
Beat Buesser's avatar
Beat Buesser committed
14
[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
Beat Buesser's avatar
Beat Buesser committed
15
[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
Beat Buesser's avatar
Beat Buesser committed
16
[![PyPI - Python Version](https://img.shields.io/pypi/pyversions/adversarial-robustness-toolbox)](https://pypi.org/project/adversarial-robustness-toolbox/)
Beat Buesser's avatar
Beat Buesser committed
17
[![slack-img](https://img.shields.io/badge/chat-on%20slack-yellow.svg)](https://ibm-art.slack.com/)
18
19
[![Downloads](https://pepy.tech/badge/adversarial-robustness-toolbox)](https://pepy.tech/project/adversarial-robustness-toolbox)
[![Downloads](https://pepy.tech/badge/adversarial-robustness-toolbox/month)](https://pepy.tech/project/adversarial-robustness-toolbox)
20
[![CII Best Practices](https://bestpractices.coreinfrastructure.org/projects/5090/badge)](https://bestpractices.coreinfrastructure.org/projects/5090)
21

Beat Buesser's avatar
Beat Buesser committed
22
23
[中文README请按此处](README-cn.md)

Beat Buesser's avatar
Beat Buesser committed
24
Adversarial Robustness Toolbox (ART) is a Python library for Machine Learning Security. ART provides tools that enable
Beat Buesser's avatar
Beat Buesser committed
25
developers and researchers to defend and evaluate Machine Learning models and applications against the
Beat Buesser's avatar
Beat Buesser committed
26
27
adversarial threats of Evasion, Poisoning, Extraction, and Inference. ART supports all popular machine learning frameworks
(TensorFlow, Keras, PyTorch, MXNet, scikit-learn, XGBoost, LightGBM, CatBoost, GPy, etc.), all data types
Beat Buesser's avatar
Beat Buesser committed
28
29
(images, tables, audio, video, etc.) and machine learning tasks (classification, object detection, speech recognition,
generation, certification, etc.).
Beat Buesser's avatar
Beat Buesser committed
30

Beat Buesser's avatar
Beat Buesser committed
31
32
## Adversarial Threats

Beat Buesser's avatar
Beat Buesser committed
33
34
35
36
37
<p align="center">
  <img src="docs/images/adversarial_threats_attacker.png?raw=true" width="400" title="ART logo">
  <img src="docs/images/adversarial_threats_art.png?raw=true" width="400" title="ART logo">
</p>
<br />
Irina Nicolae's avatar
Irina Nicolae committed
38

Beat Buesser's avatar
Beat Buesser committed
39
40
41
42
43
44
45
## ART for Red and Blue Teams (selection)

<p align="center">
  <img src="docs/images/white_hat_blue_red.png?raw=true" width="800" title="ART Red and Blue Teams">
</p>
<br />

Beat Buesser's avatar
Beat Buesser committed
46
## Learn more
Beat Buesser's avatar
Beat Buesser committed
47

Beat Buesser's avatar
Beat Buesser committed
48
| **[Get Started][get-started]**     | **[Documentation][documentation]**     | **[Contributing][contributing]**           |
Beat Buesser's avatar
Beat Buesser committed
49
|-------------------------------------|-------------------------------|-----------------------------------|
Beat Buesser's avatar
Beat Buesser committed
50
| - [Installation][installation]<br>- [Examples](examples/README.md)<br>- [Notebooks](notebooks/README.md) | - [Attacks][attacks]<br>- [Defences][defences]<br>- [Estimators][estimators]<br>- [Metrics][metrics]<br>- [Technical Documentation](https://adversarial-robustness-toolbox.readthedocs.io) | - [Slack](https://ibm-art.slack.com), [Invitation](https://join.slack.com/t/ibm-art/shared_invite/enQtMzkyOTkyODE4NzM4LTA4NGQ1OTMxMzFmY2Q1MzE1NWI2MmEzN2FjNGNjOGVlODVkZDE0MjA1NTA4OGVkMjVkNmQ4MTY1NmMyOGM5YTg)<br>- [Contributing](CONTRIBUTING.md)<br>- [Roadmap][roadmap]<br>- [Citing][citing] |
Beat Buesser's avatar
Beat Buesser committed
51

Beat Buesser's avatar
Beat Buesser committed
52
53
54
55
56
57
58
59
60
61
[get-started]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/Get-Started
[attacks]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/ART-Attacks
[defences]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/ART-Defences
[estimators]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/ART-Estimators
[metrics]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/ART-Metrics
[contributing]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/Contributing
[documentation]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/Documentation
[installation]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/Get-Started#setup
[roadmap]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/Roadmap
[citing]: https://github.com/Trusted-AI/adversarial-robustness-toolbox/wiki/Contributing#citing-art
62

Beat Buesser's avatar
Beat Buesser committed
63
The library is under continuous development. Feedback, bug reports and contributions are very welcome!
Beat Buesser's avatar
Beat Buesser committed
64
65
66
67
68

# Acknowledgment
This material is partially based upon work supported by the Defense Advanced Research Projects Agency (DARPA) under
Contract No. HR001120C0013. Any opinions, findings and conclusions or recommendations expressed in this material are
those of the author(s) and do not necessarily reflect the views of the Defense Advanced Research Projects Agency (DARPA).