Convergence Analysis of Sequential Split Learning on Heterogeneous Data

TL;DR

This paper provides a rigorous convergence analysis of Sequential Split Learning (SSL) on heterogeneous data, showing it can outperform Federated Averaging (FedAvg) in such settings, supported by empirical validation.

Contribution

It offers the first convergence guarantees for SSL on heterogeneous data and compares its performance to FedAvg, revealing its advantages.

Findings

SSL converges better than FedAvg on heterogeneous data.

Theoretical guarantees are derived for non-convex objectives.

Empirical results validate the theoretical insights.

Abstract

Federated Learning (FL) and Split Learning (SL) are two popular paradigms of distributed machine learning. By offloading the computation-intensive portions to the server, SL is promising for deep model training on resource-constrained devices, yet still lacking of rigorous convergence analysis. In this paper, we derive the convergence guarantees of Sequential SL (SSL, the vanilla case of SL that conducts the model training in sequence) for strongly/general/non-convex objectives on heterogeneous data. Notably, the derived guarantees suggest that SSL is better than Federated Averaging (FedAvg, the most popular algorithm in FL) on heterogeneous data. We validate the counterintuitive analysis result empirically on extremely heterogeneous data.