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

TL;DR

This paper introduces IBD-PSC, a novel input-level backdoor detection method leveraging parameter scaling consistency, which effectively identifies poisoned samples by analyzing prediction confidence stability under parameter amplification.

Contribution

It presents a new detection approach based on the PSC phenomenon, supported by theoretical analysis and an adaptive layer selection strategy, demonstrating robustness against adaptive attacks.

Findings

High detection accuracy on benchmark datasets

Effective resistance to adaptive backdoor attacks

Theoretical validation of PSC phenomenon

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 amplifying model parameters. In particular, we provide theoretical analysis to safeguard the foundations of the PSC phenomenon. We also design an adaptive method to select BN layers to scale up for effective detection. Extensive experiments are conducted on benchmark datasets, verifying the effectiveness and efficiency of our IBD-PSC method and its resistance to adaptive attacks. Codes are available at \href{https://github.com/THUYimingLi/BackdoorBox}{BackdoorBox}.