End-to-End Learning for Fair Ranking Systems

Learning to rank subject to constraints on group exposure, using differentiable optimization modules in predict-and-optimize fashion.

See the accompanying paper in WWW '22: https://dl.acm.org/doi/10.1145/3485447.3512247, https://arxiv.org/abs/2111.10723

Required datasets for this project can be found by following the instructions on https://github.com/him229/fultr

Credit to the original authors Himank Yadav et. al. for the LTR data and several code segments

For an example command to run the main program, see main.sh

Citing

Please cite this worky using the following:

@inproceedings{Kotary:WWW22,
  author = {James Kotary and Ferdinando Fioretto and Pascal {Van Hentenryck} and Ziwei Zhu},
  title = {End-to-end Learning for Fair Ranking Systems},
  booktitle = {In Proceedings of the ACM Web Conference ({WWW})},
  year = {2022},
  pages = {3520--3530},
  url = {https://doi.org/10.1145/3485447.3512247}
}