Wiener Graph Deconvolutional Network Improves Graph Self-Supervised Learning

TL;DR

This paper introduces the Wiener Graph Deconvolutional Network (WGDN), a novel graph decoder leveraging the graph Wiener filter, which enhances self-supervised learning by improving information reconstruction capabilities.

Contribution

The paper proposes WGDN, a new augmentation-adaptive decoder with a graph Wiener filter, demonstrating superior reconstruction ability and competitive performance in graph SSL tasks.

Findings

WGDNs outperform existing decoders in graph SSL tasks.

Theoretical analysis confirms the superior reconstruction ability of the graph Wiener filter.

Experimental results show effectiveness across various datasets.

Abstract

Graph self-supervised learning (SSL) has been vastly employed to learn representations from unlabeled graphs. Existing methods can be roughly divided into predictive learning and contrastive learning, where the latter one attracts more research attention with better empirical performance. We argue that, however, predictive models weaponed with powerful decoder could achieve comparable or even better representation power than contrastive models. In this work, we propose a Wiener Graph Deconvolutional Network (WGDN), an augmentation-adaptive decoder empowered by graph wiener filter to perform information reconstruction. Theoretical analysis proves the superior reconstruction ability of graph wiener filter. Extensive experimental results on various datasets demonstrate the effectiveness of our approach.