Cost Aware Untargeted Poisoning Attack against Graph Neural Networks,
This work addresses a specific vulnerability in GNNs for security applications, representing an incremental improvement in attack methods.
The paper tackles the inefficiency in poisoning attacks on Graph Neural Networks by proposing a Cost Aware Poisoning Attack (CA-attack) that dynamically allocates attack budget based on node classification margins, resulting in significantly enhanced attack strategies as demonstrated in experiments.
Graph Neural Networks (GNNs) have become widely used in the field of graph mining. However, these networks are vulnerable to structural perturbations. While many research efforts have focused on analyzing vulnerability through poisoning attacks, we have identified an inefficiency in current attack losses. These losses steer the attack strategy towards modifying edges targeting misclassified nodes or resilient nodes, resulting in a waste of structural adversarial perturbation. To address this issue, we propose a novel attack loss framework called the Cost Aware Poisoning Attack (CA-attack) to improve the allocation of the attack budget by dynamically considering the classification margins of nodes. Specifically, it prioritizes nodes with smaller positive margins while postponing nodes with negative margins. Our experiments demonstrate that the proposed CA-attack significantly enhances existing attack strategies