Code Repository for the paper "Fair Empirical Risk Minimization Revised" by Franco, D. and Oneto, L. and Anguita, D.
In order to make it work, you need to execute all the code contained in 'notebooks/data_cleaning.ipynb' (it will download and clean the required datasets). Then, you can either follow the steps inside 'notebooks/cv_experiments.ipynb' or launch the python script 'launchers/gurobi_bias_mitigation_validation.py'. The paper tables and images can be generated following 'notebooks/cv_experiments.ipynb'.
All the python dependencies (along with their working version tested up to the paper submission) are contained in 'requirements.txt'.