The four environments currently released with ML-fairness-gym (college admission, lending, attention allocation, and infectious disease) each have a set of experiments corresponding to publications. Running the main files in the folder corresponding to the publications, will replicate the experiments for the publication.
These experiments should provide examples of ways ML-fairness-gym can be used to investigate long term outcomes of ml-based interventions in a variety of toy scenarios.
The experiments for the following papers are in matching folders:
Alexander D’Amour, Yoni Halpern, Hansa Srinivasan, Pallavi Baljekar, James Atwood, D. Sculley. Fairness is not Static: Deeper Understanding of Long Term Fairness via Simulation Studies. ACM FAT* 2020.
The publications can be found in the papers folder.