Fairness risk measures

TL;DR

This paper introduces a new convex fairness measure for classifiers that generalizes existing definitions, accommodating various sensitive features and linking to risk measures from finance.

Contribution

It proposes a novel fairness definition that is convex, applicable to generic sensitive features, and connects fairness with risk measures like CVaR.

Findings

The new fairness measure is convex and generalizes previous definitions.

It enables optimization of classifiers with respect to fairness constraints.

The approach relates fairness to risk measures such as CVaR.

Abstract

Ensuring that classifiers are non-discriminatory or fair with respect to a sensitive feature (e.g., race or gender) is a topical problem. Progress in this task requires fixing a definition of fairness, and there have been several proposals in this regard over the past few years. Several of these, however, assume either binary sensitive features (thus precluding categorical or real-valued sensitive groups), or result in non-convex objectives (thus adversely affecting the optimisation landscape). In this paper, we propose a new definition of fairness that generalises some existing proposals, while allowing for generic sensitive features and resulting in a convex objective. The key idea is to enforce that the expected losses (or risks) across each subgroup induced by the sensitive feature are commensurate. We show how this relates to the rich literature on risk measures from mathematical finance. As a special case, this leads to a new convex fairness-aware objective based on minimising the conditional value at risk (CVaR).