Enhancing GNNs with Architecture-Agnostic Graph Transformations: A Systematic Analysis

TL;DR

This paper systematically analyzes how various graph transformations as pre-processing steps affect the performance and expressivity of different GNN architectures, revealing benefits and limitations of these techniques.

Contribution

It provides a comprehensive evaluation of architecture-agnostic graph transformations on GNN expressivity, highlighting their effects and trade-offs across standard datasets.

Findings

Node centrality augmentation improves expressivity.

Graph encoding enhances expressivity but causes numerical inaccuracies.

Transformations have limited impact on complex WL indistinguishable graphs.

Abstract

In recent years, a wide variety of graph neural network (GNN) architectures have emerged, each with its own strengths, weaknesses, and complexities. Various techniques, including rewiring, lifting, and node annotation with centrality values, have been employed as pre-processing steps to enhance GNN performance. However, there are no universally accepted best practices, and the impact of architecture and pre-processing on performance often remains opaque. This study systematically explores the impact of various graph transformations as pre-processing steps on the performance of common GNN architectures across standard datasets. The models are evaluated based on their ability to distinguish non-isomorphic graphs, referred to as expressivity. Our findings reveal that certain transformations, particularly those augmenting node features with centrality measures, consistently improve expressivity. However, these gains come with trade-offs, as methods like graph encoding, while enhancing expressivity, introduce numerical inaccuracies widely-used python packages. Additionally, we observe that these pre-processing techniques are limited when addressing complex tasks involving 3-WL and 4-WL indistinguishable graphs.