Fairness via Independence: A (Conditional) Distance Covariance Framework

TL;DR

This paper introduces a statistical framework using distance covariance to measure and improve fairness in machine learning models by promoting independence between predictions and sensitive attributes, with efficient computation and theoretical guarantees.

Contribution

It proposes a novel fairness method based on distance covariance, including a computationally efficient matrix form and convergence proofs, advancing fairness assessment and mitigation techniques.

Findings

Effective reduction of fairness gap in real-world datasets

Provides theoretical guarantees for empirical distance covariance convergence

Enhances computational efficiency with matrix form implementation

Abstract

We explore fairness from a statistical perspective by selectively utilizing either conditional distance covariance or distance covariance statistics as measures to assess the independence between predictions and sensitive attributes. We boost fairness with independence by adding a distance covariance-based penalty to the model's training. Additionally, we present the matrix form of empirical (conditional) distance covariance for parallel calculations to enhance computational efficiency. Theoretically, we provide a proof for the convergence between empirical and population (conditional) distance covariance, establishing necessary guarantees for batch computations. Through experiments conducted on a range of real-world datasets, we have demonstrated that our method effectively bridges the fairness gap in machine learning. Our code is available at \url{https://github.com/liuhaixias1/Fair_dc/}.