Deep Ensembles for Graphs with Higher-order Dependencies

TL;DR

This paper introduces Deep Graph Ensemble (DGE), a method that improves GNN performance on higher-order dependency graphs by training diverse ensembles on different neighborhood subspaces, outperforming existing models.

Contribution

The paper proposes DGE, a novel ensemble approach that captures neighborhood variance in higher-order graphs, addressing underfitting issues of traditional GNNs.

Findings

DGE outperforms existing GNNs on six real-world datasets.

Learning diverse classifiers is key to DGE's success.

DGE maintains performance under similar parameter budgets.

Abstract

Graph neural networks (GNNs) continue to achieve state-of-the-art performance on many graph learning tasks, but rely on the assumption that a given graph is a sufficient approximation of the true neighborhood structure. When a system contains higher-order sequential dependencies, we show that the tendency of traditional graph representations to underfit each node's neighborhood causes existing GNNs to generalize poorly. To address this, we propose a novel Deep Graph Ensemble (DGE), which captures neighborhood variance by training an ensemble of GNNs on different neighborhood subspaces of the same node within a higher-order network structure. We show that DGE consistently outperforms existing GNNs on semisupervised and supervised tasks on six real-world data sets with known higher-order dependencies, even under a similar parameter budget. We demonstrate that learning diverse and accurate base classifiers is central to DGE's success, and discuss the implications of these findings for future work on ensembles of GNNs.