Empirical Evaluation of Counterfactual Fairness in the Context of Statistical Fairness

Keywords

Loading...
Thumbnail Image

Authors

Issue Date

2019-06-01

Language

en

Document type

Journal Title

Journal ISSN

Volume Title

Publisher

Title

ISSN

Volume

Issue

Startpage

Endpage

DOI

Abstract

Machine learning is on the rise and has been in the past decade. Only in the last few years, a distinct focus on fairness in machine learning has surfaced. As machine learning systems become more in uential and widespread, the need arises to ensure that decisions that follow from these systems are fair. A variety of fairness de nitions have already been proposed to serve that purpose. Counterfactual fairness is one such de nition and is the focal point of this study. The abundant fairness de nitions that sometimes clash can complicate the realization of fairness. In this study, counterfactual fairness is examined in conjunction with statistical fairness to investigate whether counterfactual fairness also allows for statistical fairness. A counterfactually fair predictor is constructed on a real-world data set about loan applicants, and this predictor is empirically evaluated on a set of statistical fairness de nitions. The results suggest that counterfactual fairness can promote statistical fairness depending on which de nitions are considered.

Description

Citation

Faculty

Faculteit der Sociale Wetenschappen