Communication Efficient Private Federated Learning Using Dithering

TL;DR

This paper introduces a dithering-based quantization scheme for federated learning that reduces communication costs while maintaining differential privacy and accuracy comparable to full-precision methods.

Contribution

It proposes a novel subtractive dithering quantization method that preserves privacy and communication efficiency in federated learning.

Findings

Achieves differential privacy comparable to traditional noise addition.

Reduces communication overhead significantly.

Maintains accuracy similar to full-precision gradient methods.

Abstract

The task of preserving privacy while ensuring efficient communication is a fundamental challenge in federated learning. In this work, we tackle this challenge in the trusted aggregator model, and propose a solution that achieves both objectives simultaneously. We show that employing a quantization scheme based on subtractive dithering at the clients can effectively replicate the normal noise addition process at the aggregator. This implies that we can guarantee the same level of differential privacy against other clients while substantially reducing the amount of communication required, as opposed to transmitting full precision gradients and using central noise addition. We also experimentally demonstrate that the accuracy of our proposed approach matches that of the full precision gradient method.