GraphRNN Revisited: An Ablation Study and Extensions for Directed Acyclic Graphs

TL;DR

This paper revisits GraphRNN, confirming its performance, analyzing the impact of BFS traversal, and extending it to generate directed acyclic graphs using topological sort, with improved results on real-world data.

Contribution

It provides an ablation study on GraphRNN's components and introduces an extension for directed acyclic graphs using topological sort.

Findings

BFS traversal significantly improves model performance.

Replacing BFS with topological sort enables DAG generation.

The extended model outperforms baseline on real-world datasets.

Abstract

GraphRNN is a deep learning-based architecture proposed by You et al. for learning generative models for graphs. We replicate the results of You et al. using a reproduced implementation of the GraphRNN architecture and evaluate this against baseline models using new metrics. Through an ablation study, we find that the BFS traversal suggested by You et al. to collapse representations of isomorphic graphs contributes significantly to model performance. Additionally, we extend GraphRNN to generate directed acyclic graphs by replacing the BFS traversal with a topological sort. We demonstrate that this method improves significantly over a directed-multiclass variant of GraphRNN on a real-world dataset.