Fairness seen as Global Sensitivity Analysis

TL;DR

This paper bridges fairness in machine learning with global sensitivity analysis, showing their conceptual links, introducing new indices as fairness proxies, and providing convergence rates to enhance fairness assessment methods.

Contribution

It demonstrates how fairness can be framed within global sensitivity analysis, introduces new indices for fairness evaluation, and establishes convergence rates for these measures.

Findings

Fairness can be interpreted as a form of global sensitivity analysis.

New sensitivity indices serve as proxies for fairness.

Convergence rates for these indices are established.

Abstract

Ensuring that a predictor is not biased against a sensible feature is the key of Fairness learning. Conversely, Global Sensitivity Analysis is used in numerous contexts to monitor the influence of any feature on an output variable. We reconcile these two domains by showing how Fairness can be seen as a special framework of Global Sensitivity Analysis and how various usual indicators are common between these two fields. We also present new Global Sensitivity Analysis indices, as well as rates of convergence, that are useful as fairness proxies.