FIFA: Making Fairness More Generalizable in Classifiers Trained on Imbalanced Data

This repo contains the code to reproduce the experiments in the paper FIFA: Making Fairness More Generalizable in Classifiers Trained on Imbalanced Data.

@inproceedings{deng2023obtaining,
    title={FIFA: Making Fairness More Generalizable in Classifiers Trained on Imbalanced Data},
    author={Zhun Deng and Jiayao Zhang and Linjun Zhang and Ting Ye and Yates Coley and Weijie J Su and James Zou},
    booktitle={International Conference on Learning Representations},
    year={2023},
    url={https://openreview.net/forum?id=zVrw4OH1Lch}
}

Reproducing Experiments

(Optional) Sweeps

All experiments are done using wandb sweeps. The sweep configs are stored in ./sweep_configs/ with self-explanatory filenames.

Reproducing Figures and Tables

The results of those sweeps are stored in the csv files in ./csv_data. Please see the notebook reproduction.ipynb for details.