Intersectional Divergence: Measuring Fairness in Regression

TL;DR

This paper introduces Intersectional Divergence (ID), a novel fairness measure for regression tasks that accounts for multiple protected attributes and domain preferences, enabling better fairness assessment and optimization.

Contribution

It proposes ID as the first fairness measure for regression that considers intersectionality and domain relevance, along with an adaptable loss function IDLoss for practical optimization.

Findings

ID provides unique insights into model fairness across attributes.

Incorporating IDLoss improves fairness without sacrificing predictive accuracy.

ID enables optimization of intersectional fairness in regression models.

Abstract

Fairness in machine learning research is commonly framed in the context of classification tasks, leaving critical gaps in regression. In this paper, we propose a novel approach to measure intersectional fairness in regression tasks, going beyond the focus on single protected attributes from existing work to consider combinations of all protected attributes. Furthermore, we contend that it is insufficient to measure the average error of groups without regard for imbalanced domain preferences. Accordingly, we propose Intersectional Divergence (ID) as the first fairness measure for regression tasks that 1) describes fair model behavior across multiple protected attributes and 2) differentiates the impact of predictions in target ranges most relevant to users. We extend our proposal demonstrating how ID can be adapted into a loss function, IDLoss, that satisfies convergence guarantees and has piecewise smooth properties that enable practical optimization. Through an extensive experimental evaluation, we demonstrate how ID allows unique insights into model behavior and fairness, and how incorporating IDLoss into optimization can considerably improve single-attribute and intersectional model fairness while maintaining a competitive balance in predictive performance.