Differentially Private Federated Learning: A Systematic Review

2024-05-14

arXiv.org

AI Summary

Existing research on integrating federated learning (FL) with differential privacy (DP) lacks a systematic survey, particularly overlooking distinctions among privacy targets (e.g., clients, gradients, models) and privacy guarantee strengths (e.g., LDP, GDP, CDP) across FL architectural layers. Method: We conduct a systematic literature review (SLR), complemented by theoretical analysis of DP principles and decomposition of FL system architecture, to develop a unified taxonomy. Contribution/Results: This work introduces the first comprehensive classification framework that jointly considers privacy definitions, guarantee levels, and FL layer-specific contexts. It constructs a structured knowledge graph mapping protection boundaries and applicability conditions for each mechanism, uncovers evolutionary trends in privacy-preserving FL, and identifies fundamental bottlenecks in the privacy–utility trade-off. The framework provides reusable evaluation criteria and a practical roadmap for designing privacy-enhanced FL systems.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Computer Vision: Bias, Fairness & Privacy

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Security and Privacy: Security and privacy of machine learning and AI applications
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

In recent years, privacy and security concerns in machine learning have promoted trusted federated learning to the forefront of research. Differential privacy has emerged as the de facto standard for privacy protection in federated learning due to its rigorous mathematical foundation and provable guarantee. Despite extensive research on algorithms that incorporate differential privacy within federated learning, there remains an evident deficiency in systematic reviews that categorize and synthesize these studies. Our work presents a systematic overview of the differentially private federated learning. Existing taxonomies have not adequately considered objects and level of privacy protection provided by various differential privacy models in federated learning. To rectify this gap, we propose a new taxonomy of differentially private federated learning based on definition and guarantee of various differential privacy models and federated scenarios. Our classification allows for a clear delineation of the protected objects across various differential privacy models and their respective neighborhood levels within federated learning environments. Furthermore, we explore the applications of differential privacy in federated learning scenarios. Our work provide valuable insights into privacy-preserving federated learning and suggest practical directions for future research.

Problem

• Systematically reviews differentially private federated learning studies
• Proposes taxonomy based on privacy models and federated scenarios
• Explores differential privacy applications in federated learning

Innovation

• Systematic review of differentially private federated learning
• New taxonomy based on privacy models and scenarios
• Classification delineates protected objects and levels