From Graph Diffusion to Graph Classification

TL;DR

This paper explores the application of score-based graph diffusion models for graph classification, introducing a novel training objective that enhances accuracy and achieves state-of-the-art results.

Contribution

It proposes a new training objective tailored for graph classification using diffusion models, addressing the challenges of complex graph topologies.

Findings

Score-based graph diffusion models can be effectively applied to graph classification.

A novel training objective improves classification accuracy.

Achieves state-of-the-art performance on benchmark datasets.

Abstract

Generative models such as diffusion models have achieved remarkable success in state-of-the-art image and text tasks. Recently, score-based diffusion models have extended their success beyond image generation, showing competitive performance with discriminative methods in image {\em classification} tasks~\cite{zimmermann2021score}. However, their application to classification in the {\em graph} domain, which presents unique challenges such as complex topologies, remains underexplored. We show how graph diffusion models can be applied for graph classification. We find that to achieve competitive classification accuracy, score-based graph diffusion models should be trained with a novel training objective that is tailored to graph classification. In experiments with a sampling-based inference method, our discriminative training objective achieves state-of-the-art graph classification accuracy.