Repository for the experiments of the paper Estimating the Risk of Individual Discrimination of Classifiers presented in PAKDD2023.
Interested in cloning the project and running it on your computer?
First clone the project in your machine:
git clone https://github.com/jovasque156/foresee.git
And then you must install the packages that are in requirements.txt and also make sure you have the official OpenAI Python bindings installed.
pip install -r requirements.txt
Due to privacy restriction, we do not provide the dataset for Dropout dataset. However, the results can still found in the experiments/dropout
folder.
The file Supplementary Material contains all the proofs of the theorem. We also provide detail inforamtion regarding exepriments and datasets.
If you have any questions, please submit an Issue or a pull request.
@inproceedings{vasquez2023estimating,
title={Estimating the Risk of Individual Discrimination of Classifiers},
author={Vasquez, Jonathan and Gitiaux, Xavier and Rangwala, Huzefa},
booktitle={Advances in Knowledge Discovery and Data Mining: 27th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2023, Osaka, Japan, May 25--28, 2023, Proceedings, Part I},
pages={495--506},
year={2023},
organization={Springer}
}