Does Privacy Always Harm Fairness? Data-Dependent Trade-offs via Chernoff Information Neural Estimation

TL;DR

This paper introduces a data-dependent information-theoretic framework using Chernoff Information to analyze the fundamental trade-offs between fairness, privacy, and accuracy in machine learning models, supported by a neural estimator for real-world data.

Contribution

It proposes the Chernoff Difference for data fairness, develops the first neural estimator for Chernoff Information, and analyzes fairness-privacy trade-offs beyond synthetic data.

Findings

Noisy Chernoff Difference shows three behaviors depending on data distribution

CINE accurately estimates Chernoff Information on real datasets

Provides a principled, data-dependent view of fairness-privacy trade-offs

Abstract

Fairness and privacy are two vital pillars of trustworthy machine learning. Despite extensive research on these individual topics, their relationship has received significantly less attention. In this paper, we utilize an information-theoretic measure Chernoff Information to characterize the fundamental trade-off between fairness, privacy, and accuracy, as induced by the input data distribution. We first propose Chernoff Difference, a notion of data fairness, along with its noisy variant, Noisy Chernoff Difference, which allows us to analyze both fairness and privacy simultaneously. Through simple Gaussian examples, we show that Noisy Chernoff Difference exhibits three qualitatively distinct behaviors depending on the underlying data distribution. To extend this analysis beyond synthetic settings, we develop the Chernoff Information Neural Estimator (CINE), the first neural network-based estimator of Chernoff Information for unknown distributions. We apply CINE to analyze the Noisy Chernoff Difference on real-world datasets. Together, this work fills a critical gap in the literature by providing a principled, data-dependent characterization of the fairness-privacy interaction.