Feature Transportation Improves Graph Neural Networks

TL;DR

This paper introduces ADR-GNN, a novel graph neural network architecture inspired by physical systems, which models feature transportation, smoothing, and non-linear transformations to improve learning on complex graph data.

Contribution

The paper proposes ADR-GNN, integrating advection, diffusion, and reaction processes inspired by physical systems, to enhance feature modeling in GNNs.

Findings

ADR-GNN outperforms existing GNNs on node classification tasks.

It demonstrates competitive results on spatio-temporal datasets.

The architecture effectively captures complex feature interactions.

Abstract

Graph neural networks (GNNs) have shown remarkable success in learning representations for graph-structured data. However, GNNs still face challenges in modeling complex phenomena that involve feature transportation. In this paper, we propose a novel GNN architecture inspired by Advection-Diffusion-Reaction systems, called ADR-GNN. Advection models feature transportation, while diffusion captures the local smoothing of features, and reaction represents the non-linear transformation between feature channels. We provide an analysis of the qualitative behavior of ADR-GNN, that shows the benefit of combining advection, diffusion, and reaction. To demonstrate its efficacy, we evaluate ADR-GNN on real-world node classification and spatio-temporal datasets, and show that it improves or offers competitive performance compared to state-of-the-art networks.