SEEP: Training Dynamics Grounds Latent Representation Search for Mitigating Backdoor Poisoning Attacks

TL;DR

This paper introduces SEEP, a training dynamics-based method for detecting and removing poisoned data in NLP models, significantly reducing backdoor attack success rates while preserving accuracy.

Contribution

SEEP leverages training dynamics and label propagation to effectively identify and eliminate poisoned samples, outperforming existing defenses against backdoor attacks.

Findings

Reduces success rates of backdoor attacks significantly.

Maintains high accuracy on clean test data.

Outperforms recent defense methods.

Abstract

Modern NLP models are often trained on public datasets drawn from diverse sources, rendering them vulnerable to data poisoning attacks. These attacks can manipulate the model's behavior in ways engineered by the attacker. One such tactic involves the implantation of backdoors, achieved by poisoning specific training instances with a textual trigger and a target class label. Several strategies have been proposed to mitigate the risks associated with backdoor attacks by identifying and removing suspected poisoned examples. However, we observe that these strategies fail to offer effective protection against several advanced backdoor attacks. To remedy this deficiency, we propose a novel defensive mechanism that first exploits training dynamics to identify poisoned samples with high precision, followed by a label propagation step to improve recall and thus remove the majority of poisoned instances. Compared with recent advanced defense methods, our method considerably reduces the success rates of several backdoor attacks while maintaining high classification accuracy on clean test sets.