On Using Secure Aggregation in Differentially Private Federated Learning with Multiple Local Steps

TL;DR

This paper introduces a new analysis method in federated learning that allows multiple local optimization steps while maintaining privacy guarantees through secure aggregation, improving model utility with limited communication rounds.

Contribution

A novel analysis enabling multiple local steps in federated learning with secure aggregation, enhancing privacy and utility without increasing communication rounds.

Findings

Allows multiple local steps with secure aggregation

Improves model utility under privacy constraints

Reduces communication rounds needed for privacy guarantees

Abstract

Federated learning is a distributed learning setting where the main aim is to train machine learning models without having to share raw data but only what is required for learning. To guarantee training data privacy and high-utility models, differential privacy and secure aggregation techniques are often combined with federated learning. However, with fine-grained protection granularities, e.g., with the common sample-level protection, the currently existing techniques generally require the parties to communicate for each local optimization step, if they want to fully benefit from the secure aggregation in terms of the resulting formal privacy guarantees. In this paper, we show how a simple new analysis allows the parties to perform multiple local optimization steps while still benefiting from using secure aggregation. We show that our analysis enables higher utility models with guaranteed privacy protection under limited number of communication rounds.