Multi-Task Graph Autoencoders

TL;DR

This paper introduces a multi-task autoencoder for graph data that jointly learns to predict links and classify nodes, demonstrating superior performance on benchmark datasets with a simple, end-to-end training process.

Contribution

It proposes a novel autoencoder architecture for simultaneous link prediction and node classification, simplifying training compared to previous multi-step methods.

Findings

Significant improvement over baseline methods on five datasets

Efficient end-to-end training process

Versatile model applicable to various graph tasks

Abstract

We examine two fundamental tasks associated with graph representation learning: link prediction and node classification. We present a new autoencoder architecture capable of learning a joint representation of local graph structure and available node features for the simultaneous multi-task learning of unsupervised link prediction and semi-supervised node classification. Our simple, yet effective and versatile model is efficiently trained end-to-end in a single stage, whereas previous related deep graph embedding methods require multiple training steps that are difficult to optimize. We provide an empirical evaluation of our model on five benchmark relational, graph-structured datasets and demonstrate significant improvement over three strong baselines for graph representation learning. Reference code and data are available at https://github.com/vuptran/graph-representation-learning