This repo contains code to reproduce the experiments in the paper

La Cava, W. and Moore, J.H. (2020). "Genetic programming approaches to learning fair classifiers." GECCO 2020. doi:10.1145/3377930.3390157, arxiv:2004.13282

reproducing the experiments

The experiments can be run by navigating to analysis/ and running

python submit_jobs.py

check out submit_jobs.py to get help on configuration.

figures

The summary figures are generated in hypervolume_comparison.ipynb, and the stats tests are in stats.ipynb.

Pareto front comparisons are generated by pareto_front_plots.ipynb.

References

Aside from the usual (sklearn etc.), we make heavy use of the GerryFair repository and its associated work (https://arxiv.org/abs/1711.05144, https://arxiv.org/abs/1808.08166)

Hypervolume calculations come from DEAP

The GP models are based on feat

Acknowledgments

The authors would like thank colleagues in the Warren Center for Data Science and the Institute for Biomedical Informatics at Penn for their discussions. This work is supported by NIH grants K99 LM012926-02, R01 LM010098 and R01 AI116794.