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.

What problem does this paper attempt to address?

The paper aims to address the lack of systematic reviews in the field of Differentially Private Federated Learning (DPFL). Specifically: 1. **Lack of Systematic Reviews**: Although there is a substantial body of literature on combining differential privacy with federated learning, there is a lack of systematic classification and comprehensive studies. The paper proposes a new classification system to categorize existing research based on the definitions and guarantees of different differential privacy models. 2. **Classification of Differential Privacy Models**: Existing classification methods primarily distinguish based on whether there is a centralized trusted server, but this approach is overly simplistic. The new classification system considers Centralized Differential Privacy (CDP), Local Differential Privacy (LDP), and the Shuffle Model, and it explores the relationships between these models and their applications in federated learning in detail. 3. **Classification of Federated Learning Scenarios**: The paper covers not only Horizontal Federated Learning (HFL), Vertical Federated Learning (VFL), and Transfer Federated Learning (TFL) but also discusses the application of differential privacy techniques in these scenarios. 4. **Future Research Directions**: Based on the above research, the paper proposes 5 future research directions to further advance the development of differentially private federated learning. In summary, by proposing a new classification system, the paper systematically reviews the research achievements in the field of differentially private federated learning and provides valuable guidance for future research.

Differentially Private Federated Learning: A Systematic Review

A Systematic Survey for Differential Privacy Techniques in Federated Learning

Federated Transfer Learning with Differential Privacy

The Impact of Differential Privacy on Model Fairness in Federated Learning.

Fair Differentially Private Federated Learning Framework

Differentially Private Federated Learning: Algorithm, Analysis and Optimization

From distributed machine learning to federated learning: In the view of data privacy and security

A Multivocal Literature Review on Privacy and Fairness in Federated Learning

Federated f-Differential Privacy

Differential Privacy Meets Federated Learning under Communication Constraints

Privacy, accuracy, and model fairness trade-offs in federated learning

Differentially private secure multi-party computation for federated learning in financial applications

Differential Privacy for Deep and Federated Learning: A Survey

Differentially Private Federated Learning: Servers Trustworthiness, Estimation, and Statistical Inference

An overview of implementing security and privacy in federated learning

Research on the Security Technology of Federated Learning Privacy Preserving

Utility Optimization of Federated Learning with Differential Privacy

Differentially Private Byzantine-Robust Federated Learning

A review of federated learning: taxonomy, privacy and future directions

A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection