A Convex Framework for Fair Regression

TL;DR

This paper presents a convex framework for fair regression that allows for efficient optimization and explores the trade-off between accuracy and fairness across multiple datasets.

Contribution

It introduces a family of convex fairness regularizers for regression, enabling computation of the accuracy-fairness trade-off frontier and analysis of the Price of Fairness.

Findings

The framework supports various notions of fairness, from group to individual.

Efficient frontiers of accuracy and fairness can be computed for different datasets.

The study quantifies the severity of fairness trade-offs using the Price of Fairness.

Abstract

We introduce a flexible family of fairness regularizers for (linear and logistic) regression problems. These regularizers all enjoy convexity, permitting fast optimization, and they span the rang from notions of group fairness to strong individual fairness. By varying the weight on the fairness regularizer, we can compute the efficient frontier of the accuracy-fairness trade-off on any given dataset, and we measure the severity of this trade-off via a numerical quantity we call the Price of Fairness (PoF). The centerpiece of our results is an extensive comparative study of the PoF across six different datasets in which fairness is a primary consideration.