FACT: A Diagnostic for Group Fairness Trade-offs

This repository contains the code for the following paper:

FACT: A Diagnostic for Group Fairness Trade-offs
Joon Sik Kim, Jiahao Chen, Ameet Talwalkar
To appear in International Conference on Machine Learning (ICML), 2020.

Requirements

The code was developed and tested on Python 3.5.2. Other required libraries can be installed via running the command:

pip install -r requirements.txt

Other papers/repositories used for comparisons

Zafar et al. 2017 and its code.

Hardt et al. 2016 and its implementation here

Tan et al. 2020 and its code

The output files from these algorithms used to plot in Jupyter notebooks are available here, along with the code portions (adpated from the above repositories) that generated them. Place all csv files in this link under the notebook directory to reproduce the plots.