Graph Context Encoder: Graph Feature Inpainting for Graph Generation and Self-supervised Pretraining

TL;DR

The paper introduces Graph Context Encoder (GCE), a self-supervised method for graph feature inpainting that enhances graph generation and improves classification performance through effective pretraining.

Contribution

GCE is a novel self-supervised approach that uses feature masking and reconstruction for graph representation learning and generation.

Findings

GCE can generate new graphs, including molecules.

Pretraining with GCE improves classification accuracy.

GCE outperforms baseline methods in benchmark tasks.

Abstract

We propose the Graph Context Encoder (GCE), a simple but efficient approach for graph representation learning based on graph feature masking and reconstruction. GCE models are trained to efficiently reconstruct input graphs similarly to a graph autoencoder where node and edge labels are masked. In particular, our model is also allowed to change graph structures by masking and reconstructing graphs augmented by random pseudo-edges. We show that GCE can be used for novel graph generation, with applications for molecule generation. Used as a pretraining method, we also show that GCE improves baseline performances in supervised classification tasks tested on multiple standard benchmark graph datasets.