When MiniBatch SGD Meets SplitFed Learning:Convergence Analysis and Performance Evaluation

TL;DR

This paper introduces MiniBatch-SFL, a novel split federated learning algorithm that combines MiniBatch SGD to improve convergence and mitigate client drift, especially with non-IID data, leading to significant accuracy gains.

Contribution

The paper proposes MiniBatch-SFL, integrating MiniBatch SGD into split federated learning to enhance convergence analysis and performance with non-IID data.

Findings

MiniBatch-SFL achieves up to 24.1% higher accuracy with non-IID data.

Server-side updates are unaffected by data non-IIDness, aiding convergence.

A later cut layer position improves gradient divergence and overall performance.

Abstract

Federated learning (FL) enables collaborative model training across distributed clients (e.g., edge devices) without sharing raw data. Yet, FL can be computationally expensive as the clients need to train the entire model multiple times. SplitFed learning (SFL) is a recent distributed approach that alleviates computation workload at the client device by splitting the model at a cut layer into two parts, where clients only need to train part of the model. However, SFL still suffers from the \textit{client drift} problem when clients' data are highly non-IID. To address this issue, we propose MiniBatch-SFL. This algorithm incorporates MiniBatch SGD into SFL, where the clients train the client-side model in an FL fashion while the server trains the server-side model similar to MiniBatch SGD. We analyze the convergence of MiniBatch-SFL and show that the bound of the expected loss can be obtained by analyzing the expected server-side and client-side model updates, respectively. The server-side updates do not depend on the non-IID degree of the clients' datasets and can potentially mitigate client drift. However, the client-side model relies on the non-IID degree and can be optimized by properly choosing the cut layer. Perhaps counter-intuitive, our empirical result shows that a latter position of the cut layer leads to a smaller average gradient divergence and a better algorithm performance. Moreover, numerical results show that MiniBatch-SFL achieves higher accuracy than conventional SFL and FL. The accuracy improvement can be up to 24.1\% and 17.1\% with highly non-IID data, respectively.