Local data debiasing for fairness based on generative adversarial training

Aivodji, Ulrich, Bidet, François, Gambs, Sébastien, Ngueveu, Rosin Claude et Tapp, Alain. 2021. « Local data debiasing for fairness based on generative adversarial training ». Algorithms, vol. 14, nº 3.
Compte des citations dans Scopus : 4.

Prévisualisation

PDF Aivodji-U-2021-23913.pdf - Version publiée Licence d'utilisation : Creative Commons CC BY. Télécharger (2MB) | Prévisualisation

Résumé

The widespread use of automated decision processes in many areas of our society raises serious ethical issues with respect to the fairness of the process and the possible resulting discrimination. To solve this issue, we propose a novel adversarial training approach called GANSan for learning a sanitizer whose objective is to prevent the possibility of any discrimination (i.e., direct and indirect) based on a sensitive attribute by removing the attribute itself as well as the existing correlations with the remaining attributes. Our method GANSan is partially inspired by the powerful framework of generative adversarial networks (in particular Cycle-GANs), which offers a flexible way to learn a distribution empirically or to translate between two different distributions. In contrast to prior work, one of the strengths of our approach is that the sanitization is performed in the same space as the original data by only modifying the other attributes as little as possible, thus preserving the interpretability of the sanitized data. Consequently, once the sanitizer is trained, it can be applied to new data locally by an individual on their profile before releasing it. Finally, experiments on real datasets demonstrate the effectiveness of the approach as well as the achievable trade-off between fairness and utility

Type de document:	Article publié dans une revue, révisé par les pairs
Professeur:	Professeur Aïvodji, Ulrich
Affiliation:	Autres
Date de dépôt:	28 janv. 2022 20:48
Dernière modification:	03 mars 2022 16:09
URI:	https://espace2.etsmtl.ca/id/eprint/23913

Actions (Authentification requise)

Dernière vérification avant le dépôt