A General Approach to Fairness with Optimal Transport

Silvia Chiappa, Ray Jiang, Tom Stepleton, Aldo Pacchiano, Heinrich Jiang, John Aslanides

AAAI 2020 pp. 3633-3640

doi:10.1609/AAAI.V34I04.5771 /aaai/2020/chiappa2020aaai-general/

Abstract

We propose a general approach to fairness based on transporting distributions corresponding to different sensitive attributes to a common distribution. We use optimal transport theory to derive target distributions and methods that allow us to achieve fairness with minimal changes to the unfair model. Our approach is applicable to both classification and regression problems, can enforce different notions of fairness, and enable us to achieve a Pareto-optimal trade-off between accuracy and fairness. We demonstrate that it outperforms previous approaches in several benchmark fairness datasets.

PDF AAAI Semantic Scholar

Cite

Text

Chiappa et al. "A General Approach to Fairness with Optimal Transport." AAAI Conference on Artificial Intelligence, 2020. doi:10.1609/AAAI.V34I04.5771

Markdown

[Chiappa et al. "A General Approach to Fairness with Optimal Transport." AAAI Conference on Artificial Intelligence, 2020.](https://mlanthology.org/aaai/2020/chiappa2020aaai-general/) doi:10.1609/AAAI.V34I04.5771

BibTeX

@inproceedings{chiappa2020aaai-general,
  title     = {{A General Approach to Fairness with Optimal Transport}},
  author    = {Chiappa, Silvia and Jiang, Ray and Stepleton, Tom and Pacchiano, Aldo and Jiang, Heinrich and Aslanides, John},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2020},
  pages     = {3633-3640},
  doi       = {10.1609/AAAI.V34I04.5771},
  url       = {https://mlanthology.org/aaai/2020/chiappa2020aaai-general/}
}