Fairness seen as Global Sensitivity Analysis

2021/03/08 by Clément Bénesse, Fabrice Gamboa, Bénesse, Clément +5 · 4 citations
Business, Management and Accounting · Social Sciences · #FOS: Computer and information sciences #FOS: Mathematics #Global trade, sustainability, and social impact #Methodology (stat.ME) #Qualitative Comparative Analysis Research #Statistics Theory (math.ST)

paper · doi:10.48550/arxiv.2103.04613

openalex publication_date 2021/03/08 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28

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.

Cited by

• The Fair Game: Auditing & Debiasing AI Algorithms Over Time
• Robust ML Auditing using Prior Knowledge
• Marginal Fairness: Fair Decision-Making under Risk Measures
• On Measuring Intrinsic Causal Attributions in Deep Neural Networks

Related

• The Fragility of Fairness: Causal Sensitivity Analysis for Fair Machine Learning
• The impossibility of "fairness": a generalized impossibility result for decisions
• Understanding Fairness and Prediction Error through Subspace Decomposition and Influence Analysis
• Overcoming Fairness Trade-offs via Pre-processing: A Causal Perspective
• Towards Auditability for Fairness in Deep Learning