You Can't Ignore Either: Unifying Structure and Feature Denoising for Robust Graph Learning

TL;DR

This paper introduces a unified framework for denoising both structure and features in graphs, improving the robustness of Graph Neural Networks against noise and attacks through a self-supervised, plug-and-play approach.

Contribution

The paper proposes a novel unified graph denoising framework that simultaneously addresses structure and feature noise, using high-order proximity evaluation and graph auto-encoder refinement.

Findings

Effective noise recognition via high-order neighborhood proximity

Improved robustness of GNNs in noisy environments

Self-supervised, plug-and-play denoising module

Abstract

Recent research on the robustness of Graph Neural Networks (GNNs) under noises or attacks has attracted great attention due to its importance in real-world applications. Most previous methods explore a single noise source, recovering corrupt node embedding by reliable structures bias or developing structure learning with reliable node features. However, the noises and attacks may come from both structures and features in graphs, making the graph denoising a dilemma and challenging problem. In this paper, we develop a unified graph denoising (UGD) framework to unravel the deadlock between structure and feature denoising. Specifically, a high-order neighborhood proximity evaluation method is proposed to recognize noisy edges, considering features may be perturbed simultaneously. Moreover, we propose to refine noisy features with reconstruction based on a graph auto-encoder. An iterative updating algorithm is further designed to optimize the framework and acquire a clean graph, thus enabling robust graph learning for downstream tasks. Our UGD framework is self-supervised and can be easily implemented as a plug-and-play module. We carry out extensive experiments, which proves the effectiveness and advantages of our method. Code is avalaible at https://github.com/YoungTimmy/UGD.