Variational Graph Normalized Auto-Encoders

TL;DR

This paper introduces VGNAE, a novel variational graph autoencoder that employs L2-normalization to improve link prediction involving isolated nodes, outperforming existing models.

Contribution

Proposes VGNAE, a new variational graph autoencoder with normalization to enhance embeddings of isolated nodes in link prediction.

Findings

VGNAE outperforms state-of-the-art models in link prediction.

Normalization improves embeddings of isolated nodes.

The method is effective for graphs with many isolated nodes.

Abstract

Link prediction is one of the key problems for graph-structured data. With the advancement of graph neural networks, graph autoencoders (GAEs) and variational graph autoencoders (VGAEs) have been proposed to learn graph embeddings in an unsupervised way. It has been shown that these methods are effective for link prediction tasks. However, they do not work well in link predictions when a node whose degree is zero (i.g., isolated node) is involved. We have found that GAEs/VGAEs make embeddings of isolated nodes close to zero regardless of their content features. In this paper, we propose a novel Variational Graph Normalized AutoEncoder (VGNAE) that utilize L2-normalization to derive better embeddings for isolated nodes. We show that our VGNAEs outperform the existing state-of-the-art models for link prediction tasks. The code is available at https://github.com/SeongJinAhn/VGNAE.