Evaluating Graph Generative Models with Contrastively Learned Features

TL;DR

This paper introduces a new method for evaluating graph generative models using contrastively trained GNN features, which improves reliability over traditional and random GNN metrics, and explores their complementary strengths.

Contribution

It proposes using contrastively trained GNN representations for evaluation, demonstrating their effectiveness and the complementary nature of traditional, GNN-based, and GSN approaches.

Findings

Contrastively trained GNNs provide more reliable evaluation metrics.

Traditional and GNN-based approaches have complementary strengths.

Graph Substructure Networks better distinguish between graph datasets.

Abstract

A wide range of models have been proposed for Graph Generative Models, necessitating effective methods to evaluate their quality. So far, most techniques use either traditional metrics based on subgraph counting, or the representations of randomly initialized Graph Neural Networks (GNNs). We propose using representations from contrastively trained GNNs, rather than random GNNs, and show this gives more reliable evaluation metrics. Neither traditional approaches nor GNN-based approaches dominate the other, however: we give examples of graphs that each approach is unable to distinguish. We demonstrate that Graph Substructure Networks (GSNs), which in a way combine both approaches, are better at distinguishing the distances between graph datasets.