Can Ensembling Pre-processing Algorithms Lead to Better Machine Learning Fairness?

TL;DR

This paper evaluates three fairness pre-processing algorithms in machine learning, exploring whether combining them into an ensemble can improve fairness outcomes while managing trade-offs with accuracy.

Contribution

It introduces an ensemble approach to fairness pre-processing algorithms and provides insights on their combined effectiveness and selection strategies.

Findings

Ensembling fairness algorithms can enhance bias mitigation.

Trade-offs between fairness and accuracy are context-dependent.

Guidelines for practitioners on selecting fairness algorithms are provided.

Abstract

As machine learning (ML) systems get adopted in more critical areas, it has become increasingly crucial to address the bias that could occur in these systems. Several fairness pre-processing algorithms are available to alleviate implicit biases during model training. These algorithms employ different concepts of fairness, often leading to conflicting strategies with consequential trade-offs between fairness and accuracy. In this work, we evaluate three popular fairness pre-processing algorithms and investigate the potential for combining all algorithms into a more robust pre-processing ensemble. We report on lessons learned that can help practitioners better select fairness algorithms for their models.