Self-Supervised Graph Representation Learning via Topology Transformations

TL;DR

This paper introduces a self-supervised graph representation learning method that leverages topology transformations, maximizing mutual information between original and transformed graphs to improve various graph-related tasks.

Contribution

It formalizes a novel topology transformation equivariant learning paradigm using an information-theoretic approach for self-supervised graph representations.

Findings

Outperforms state-of-the-art unsupervised methods in node classification.

Effective in graph classification and link prediction tasks.

Demonstrates the benefit of topology transformations in representation learning.

Abstract

We present the Topology Transformation Equivariant Representation learning, a general paradigm of self-supervised learning for node representations of graph data to enable the wide applicability of Graph Convolutional Neural Networks (GCNNs). We formalize the proposed model from an information-theoretic perspective, by maximizing the mutual information between topology transformations and node representations before and after the transformations. We derive that maximizing such mutual information can be relaxed to minimizing the cross entropy between the applied topology transformation and its estimation from node representations. In particular, we seek to sample a subset of node pairs from the original graph and flip the edge connectivity between each pair to transform the graph topology. Then, we self-train a representation encoder to learn node representations by reconstructing the topology transformations from the feature representations of the original and transformed graphs. In experiments, we apply the proposed model to the downstream node classification, graph classification and link prediction tasks, and results show that the proposed method outperforms the state-of-the-art unsupervised approaches.