code-soup

code-soup is the python code for the book "Adversarial Deep Learning" and its tutorials. You can use this in conjunction with a course on Adversarial Deep Learning, or for study on your own. We're looking for solid contributors to help.

Despite the great success of deep neural networks in a wide range of applications, they have been repeatedly shown to be vulnerable to adversarial attacks. Adversarial Deep Learning is a book being written by Dr. Di Jin, Dr. Yifang Yin, Yaman Kumar, and Dr. Rajiv Ratn Shah, which gives the reader an introduction to the progress made in this field. At code-soup we are building the codebase of these algorithms in a clean, simple and minimal manner . We strive to give the reader a smooth experience while reading the book and understanding the code in parallel with a minimal set of dependencies and library. Contact of the core developers can be seen in AUTHORS.

Hacktoberfest2021

We will be participating in Hacktoberfest 2021! For instructions join our Slack Channel here! Look at the contribution guidelines for starters!

Structure of the project

When complete, this project will have Python implementations for all the pseudocode algorithms in the book, as well as tests and examples of use. You can check the exact repository structure here Repository Structure Docs. The overall idea is to let the user read the algorithm and understand the attack in the code-soup/ch{ch_num}/models/{topic}.py and the demonstration in the tutorial.

Requirements

The requirements are stored in requirements.txt you can install them using

pip install -r requirements.txt

We recommend to use a virtual environment, the exported yaml is available at environment.yml.

Tutorials

The tutorial to each algorithm is available in the Tutorials folder.

Index

Index for tutorials and test suite for each algorithm.

Contribution

Please take a look the CONTRIBUTING.md for details, :star: us if you liked the work.