IBD-PSC: Input-level Backdoor Detection via Parameter-oriented Scaling Consistency

arXiv:2405.09786v1 Announce Type: new
Abstract: Deep neural networks (DNNs) are vulnerable to backdoor attacks, where adversaries can maliciously trigger model misclassifications by implanting a hidden backdoor during model training. This paper proposes a simple yet effective input-level backdoor detection (dubbed IBD-PSC) as a 'firewall' to filter out malicious testing images. Our method is motivated by an intriguing phenomenon, i.e., parameter-oriented scaling consistency (PSC), where the prediction confidences of poisoned samples are significantly more consistent than those of benign ones when …

abstract arxiv attacks backdoor cs.cr cs.lg detection filter firewall hidden images networks neural networks paper scaling simple testing training type via vulnerable