paper-with-me

홈 › Papers

Differential Privacy in Two-Layer Networks: How DP-SGD Harms Fairness and Robustness

2026-03-05 · Ruichen Xu, Kexin Chen arxiv

Differentially private learning is essential for training models on sensitive data, but empirical studies consistently show that it can degrade performance, introduce fairness issues like disparate impact, and reduce adversarial robustness. The theoretical underpinnings of these phenomena in modern, non-convex neural networks remain largely unexplored. This paper introduces a unified feature-centric framework to analyze the feature learning dynamics of differentially private stochastic gradient descent (DP-SGD) in two-layer ReLU convolutional neural networks. Our analysis establishes test loss bounds governed by a crucial metric: the feature-to-noise ratio (FNR). We demonstrate that the noise required for privacy leads to suboptimal feature learning, and specifically show that: 1) imbalanced FNRs across classes and subpopulations cause disparate impact; 2) even in the same class, noise has a greater negative impact on semantically long-tailed data; and 3) noise injection exacerbates vulnerability to adversarial attacks. Furthermore, our analysis reveals that the popular paradigm of public pre-training and private fine-tuning does not guarantee improvement, particularly under significant feature distribution shifts between datasets. Experiments on synthetic and real-world data corroborate our theoretical findings.

📄 PDF Abstract BibTeX arXiv:2603.04881

Code (0)

등록된 구현이 없습니다.

Tasks

Adversarial Robustness

Similar Papers 제목 키워드 기반

RESFL: An Uncertainty-Aware Framework for Responsible Federated Learning by Balancing Privacy, Fairness and Utility in Autonomous Vehicles

2025-03-20 · Dawood Wasif, Terrence J. Moore, Jin-Hee Cho

Autonomous vehicles (AVs) increasingly rely on Federated Learning (FL) to enhance perception models while preserving privacy. However, existing FL frameworks struggle to balance privacy, fairness, and robustness, leading…

Autonomous VehiclesDisentanglementFairnessFederated Learning+3

Towards Explainable Federated Learning: Understanding the Impact of Differential Privacy

2026-02-10 · Júlio Oliveira, Rodrigo Ferreira, André Riker, Glaucio H. S. Carvalho 외 arxiv

Data privacy and eXplainable Artificial Intelligence (XAI) are two important aspects for modern Machine Learning systems. To enhance data privacy, recent machine learning models have been designed as a Federated Learning…

Federated Learning

Fairness Meets Privacy: Integrating Differential Privacy and Demographic Parity in Multi-class Classification

2025-11-24 · Lilian Say, Christophe Denis, Rafael Pinot arxiv

The increasing use of machine learning in sensitive applications demands algorithms that simultaneously preserve data privacy and ensure fairness across potentially sensitive sub-populations. While privacy and fairness h…

Multi-class Classification

Towards Trustworthy Federated Learning

2025-03-05 · Alina Basharat, Yijun Bian, Ping Xu, Zhi Tian

This paper develops a comprehensive framework to address three critical trustworthy challenges in federated learning (FL): robustness against Byzantine attacks, fairness, and privacy preservation. To improve the system's…

FairnessFederated Learning

Fair Differentially Private Federated Learning Framework

2023-05-23 · Ayush K. Varshney, Sonakshi Garg, Arka Ghosh, Sargam Gupta

Federated learning (FL) is a distributed machine learning strategy that enables participants to collaborate and train a shared model without sharing their individual datasets. Privacy and fairness are crucial considerati…

FairnessFederated Learning