A clean-label graph backdoor attack method in node classification task

TL;DR

This paper introduces a stealthy, clean-label backdoor attack method for graph neural networks in node classification, which does not alter labels or graph structure and achieves high attack success rates.

Contribution

The paper proposes CGBA, a novel backdoor attack that is more covert by avoiding label and structure modifications, enhancing attack success in GNNs.

Findings

Achieves an average attack success rate of 87.8% with 4% poisoning rate.

Does not modify node labels or graph structure, increasing stealth.

Effective in various experimental settings.

Abstract

Backdoor attacks in the traditional graph neural networks (GNNs) field are easily detectable due to the dilemma of confusing labels. To explore the backdoor vulnerability of GNNs and create a more stealthy backdoor attack method, a clean-label graph backdoor attack method(CGBA) in the node classification task is proposed in this paper. Differently from existing backdoor attack methods, CGBA requires neither modification of node labels nor graph structure. Specifically, to solve the problem of inconsistency between the contents and labels of the samples, CGBA selects poisoning samples in a specific target class and uses the label of sample as the target label (i.e., clean-label) after injecting triggers into the target samples. To guarantee the similarity of neighboring nodes, the raw features of the nodes are elaborately picked as triggers to further improve the concealment of the triggers. Extensive experiments results show the effectiveness of our method. When the poisoning rate is 0.04, CGBA can achieve an average attack success rate of 87.8%, 98.9%, 89.1%, and 98.5%, respectively.