Scalable Federated Learning and Secure Aggregation With Differential Privacy
Abstract
Federated Learning (FL) is a distributed machine learning paradigm that enables multiple clients to collaboratively train a global model without sharing raw data. Instead of centralizing datasets, model updates are computed locally and aggregated by a coordinating server, thereby preserving data privacy and regulatory compliance. This chapter examines scalable federated learning with a focus on secure aggregation and differential privacy mechanisms. Secure aggregation protocols ensure that individual model updates remain hidden while enabling accurate computation of the global model. Additionally, differential privacy techniques are integrated to prevent leakage of sensitive information from model parameters during large-scale training. The chapter discusses key challenges including communication overhead, scalability limitations, system heterogeneity, adversarial threats, and privacy–utility trade-offs. Recent developments and real-world applications across healthcare, finance, and edge computing environments are also presented.
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.