Differential Privacy and Fairness in Decisions and Learning Tasks: A Survey

TL;DR

This survey explores the intersection of differential privacy and fairness, analyzing how privacy measures can impact bias and unfairness in decision-making and learning tasks, and reviews mitigation strategies.

Contribution

It provides a comprehensive overview of the challenges and risks of integrating differential privacy with fairness in machine learning and decision systems.

Findings

DP can both mitigate and exacerbate bias

Various mitigation measures exist for fairness issues in DP systems

Understanding the interplay of privacy and fairness is crucial for deployment

Abstract

This paper surveys recent work in the intersection of differential privacy (DP) and fairness. It reviews the conditions under which privacy and fairness may have aligned or contrasting goals, analyzes how and why DP may exacerbate bias and unfairness in decision problems and learning tasks, and describes available mitigation measures for the fairness issues arising in DP systems. The survey provides a unified understanding of the main challenges and potential risks arising when deploying privacy-preserving machine-learning or decisions-making tasks under a fairness lens.