ABCFair: an Adaptable Benchmark approach for Comparing Fairness Methods

TL;DR

ABCFair is a flexible benchmarking framework designed to compare fairness methods across diverse real-world scenarios, addressing the challenge of inconsistent problem settings in fairness research.

Contribution

It introduces an adaptable benchmark approach that enables fair comparison of different fairness methods tailored to specific problem configurations.

Findings

ABCFair successfully compares multiple fairness methods across various datasets.

The framework highlights how performance varies with different problem settings.

It demonstrates the importance of context-aware benchmarking in fairness research.

Abstract

Numerous methods have been implemented that pursue fairness with respect to sensitive features by mitigating biases in machine learning. Yet, the problem settings that each method tackles vary significantly, including the stage of intervention, the composition of sensitive features, the fairness notion, and the distribution of the output. Even in binary classification, these subtle differences make it highly complicated to benchmark fairness methods, as their performance can strongly depend on exactly how the bias mitigation problem was originally framed. Hence, we introduce ABCFair, a benchmark approach which allows adapting to the desiderata of the real-world problem setting, enabling proper comparability between methods for any use case. We apply ABCFair to a range of pre-, in-, and postprocessing methods on both large-scale, traditional datasets and on a dual label (biased and unbiased) dataset to sidestep the fairness-accuracy trade-off.