Federated fairness-aware classification under differential privacy

2026-03-25

AI Summary

This work addresses the challenge of simultaneously achieving differential privacy, algorithmic fairness, and high classification accuracy in federated learning. The authors propose a two-stage algorithm, FDP-Fair, which enables privacy-preserving fair classification under demographic parity constraints in multi-server federated settings, along with a lightweight variant, CDP-Fair, tailored for single-server scenarios. A key contribution is the first systematic decomposition of the excess risk in private fair classification into four distinct sources: inherent classification cost, privacy-induced cost, non-private fairness cost, and private fairness cost, accompanied by rigorous theoretical guarantees. Empirical evaluations on both synthetic and real-world datasets demonstrate that the proposed methods effectively balance privacy, fairness, and accuracy, exhibiting strong practical utility.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Bias, Fairness & Privacy
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

Privacy and algorithmic fairness have become two central issues in modern machine learning. Although each has separately emerged as a rapidly growing research area, their joint effect remains comparatively under-explored. In this paper, we systematically study the joint impact of differential privacy and fairness on classification in a federated setting, where data are distributed across multiple servers. Targeting demographic disparity constrained classification under federated differential privacy, we propose a two-step algorithm, namely FDP-Fair. In the special case where there is only one server, we further propose a simple yet powerful algorithm, namely CDP-Fair, serving as a computationally-lightweight alternative. Under mild structural assumptions, theoretical guarantees on privacy, fairness and excess risk control are established. In particular, we disentangle the source of the private fairness-aware excess risk into a) intrinsic cost of classification, b) cost of private classification, c) non-private cost of fairness and d) private cost of fairness. Our theoretical findings are complemented by extensive numerical experiments on both synthetic and real datasets, highlighting the practicality of our designed algorithms.

Problem

• federated learning
• differential privacy
• algorithmic fairness
• demographic disparity
• fair classification

Innovation

• federated learning
• differential privacy
• algorithmic fairness
• fair classification
• excess risk decomposition