Adversarial Learned Fair Representations using Dampening and Stacking

TL;DR

This paper introduces a novel adversarial learning algorithm employing dampening and stacking to improve fair data representations by better censoring sensitive variables and reconstruction quality.

Contribution

The paper presents a new algorithm that enhances fair representation learning through innovative dampening and stacking techniques, advancing adversarial methods.

Findings

Improved censoring of sensitive variables.

Enhanced data reconstruction quality.

Outperforms previous adversarial fair representation methods.

Abstract

As more decisions in our daily life become automated, the need to have machine learning algorithms that make fair decisions increases. In fair representation learning we are tasked with finding a suitable representation of the data in which a sensitive variable is censored. Recent work aims to learn fair representations through adversarial learning. This paper builds upon this work by introducing a novel algorithm which uses dampening and stacking to learn adversarial fair representations. Results show that that our algorithm improves upon earlier work in both censoring and reconstruction.