GLASU: A Communication-Efficient Algorithm for Federated Learning with Vertically Distributed Graph Data

TL;DR

This paper introduces GLASU, a communication-efficient federated learning algorithm for graph neural networks on vertically distributed data, reducing communication costs while maintaining model performance.

Contribution

It proposes a novel model splitting method and a communication-efficient algorithm with lazy aggregation and stale updates for federated GNN training.

Findings

Effective training of GNNs with performance comparable to centralized models.

Significant reduction in communication overhead during training.

Theoretical analysis supports the algorithm's efficiency and effectiveness.

Abstract

Vertical federated learning (VFL) is a distributed learning paradigm, where computing clients collectively train a model based on the partial features of the same set of samples they possess. Current research on VFL focuses on the case when samples are independent, but it rarely addresses an emerging scenario when samples are interrelated through a graph. For graph-structured data, graph neural networks (GNNs) are competitive machine learning models, but a naive implementation in the VFL setting causes a significant communication overhead. Moreover, the analysis of the training is faced with a challenge caused by the biased stochastic gradients. In this paper, we propose a model splitting method that splits a backbone GNN across the clients and the server and a communication-efficient algorithm, GLASU, to train such a model. GLASU adopts lazy aggregation and stale updates to skip aggregation when evaluating the model and skip feature exchanges during training, greatly reducing communication. We offer a theoretical analysis and conduct extensive numerical experiments on real-world datasets, showing that the proposed algorithm effectively trains a GNN model, whose performance matches that of the backbone GNN when trained in a centralized manner.