General code for "Fair Densities via Boosting the Sufficient Statistics of Exponential Families".
We configure using conda environment.yml provides dependencies to run the code.
Additional files need to be setup and downloaded for AIF360 datasets to work.
run_discrete.py runs FBDE for discrete datasets. An example:
ipython run_discrete.py -- --equal-rr --boosting-steps 32 --max-depth 8 --sattr 'sex' --seed 1 --dataset 'compas' --sr 0.8 --leverage 'exact'
Alternatively, one can run the continuous dataset code with run_continuous.py:
ipython run_continuous.py -- --equal-rr --boosting-steps 8 --max-depth 8 --seed 1 --dataset 'minneapolis' --sr 0.8 --leverage 'exact'
Additionally, we provide *_data.py scripts to run evaluation on the raw data.
One can use compas and adult for --dataset options, with sex or race as --sattr.
Otherwise, one can use german and dutch for --dataset options, with sex or age as --sattr.
Fair Densities via Boosting the Sufficient Statistics of Exponential Families
Alexander Soen, Hisham Husain, Richard Nock
International Conference on Machine Learning, ICML 2023
Please cite the corresponding paper when using the code