SSGD: A safe and efficient method of gradient descent

TL;DR

This paper introduces SSGD, a novel gradient descent method that enhances security against gradient leakage attacks while maintaining high accuracy and efficiency in multi-node machine learning systems.

Contribution

The paper proposes the super stochastic gradient descent approach that conceals gradient modulus, improving security without sacrificing model accuracy.

Findings

SSGD effectively defends against gradient leakage attacks.

SSGD maintains high accuracy comparable to standard gradient descent.

SSGD shows superior robustness and scalability in experiments.

Abstract

With the vigorous development of artificial intelligence technology, various engineering technology applications have been implemented one after another. The gradient descent method plays an important role in solving various optimization problems, due to its simple structure, good stability and easy implementation. In multi-node machine learning system, the gradients usually need to be shared. Shared gradients are generally unsafe. Attackers can obtain training data simply by knowing the gradient information. In this paper, to prevent gradient leakage while keeping the accuracy of model, we propose the super stochastic gradient descent approach to update parameters by concealing the modulus length of gradient vectors and converting it or them into a unit vector. Furthermore, we analyze the security of super stochastic gradient descent approach. Our algorithm can defend against attacks on the gradient. Experiment results show that our approach is obviously superior to prevalent gradient descent approaches in terms of accuracy, robustness, and adaptability to large-scale batches.