Node Copying for Protection Against Graph Neural Network Topology Attacks

TL;DR

This paper introduces a node copying algorithm to defend against graph topology attacks, improving classification robustness without significant additional computation, demonstrated on real-world datasets.

Contribution

The paper proposes a novel node copying method to mitigate the impact of topology attacks on graph neural network classification tasks.

Findings

Effective defense against topology attacks demonstrated on multiple datasets

Maintains classification accuracy with minimal additional computational cost

Scalable approach suitable for large graphs

Abstract

Adversarial attacks can affect the performance of existing deep learning models. With the increased interest in graph based machine learning techniques, there have been investigations which suggest that these models are also vulnerable to attacks. In particular, corruptions of the graph topology can degrade the performance of graph based learning algorithms severely. This is due to the fact that the prediction capability of these algorithms relies mostly on the similarity structure imposed by the graph connectivity. Therefore, detecting the location of the corruption and correcting the induced errors becomes crucial. There has been some recent work which tackles the detection problem, however these methods do not address the effect of the attack on the downstream learning task. In this work, we propose an algorithm that uses node copying to mitigate the degradation in classification that is caused by adversarial attacks. The proposed methodology is applied only after the model for the downstream task is trained and the added computation cost scales well for large graphs. Experimental results show the effectiveness of our approach for several real world datasets.