CRAICLFeb 19, 2024

Defending Against Weight-Poisoning Backdoor Attacks for Parameter-Efficient Fine-Tuning

MIT
arXiv:2402.12168v344 citationsh-index: 34NAACL-HLT
Originality Incremental advance
AI Analysis

This addresses security vulnerabilities in fine-tuning for language models, which is crucial for practitioners deploying AI systems, though it is incremental as it builds on existing attack and defense research.

The study found that parameter-efficient fine-tuning (PEFT) is more vulnerable to weight-poisoning backdoor attacks than full-parameter fine-tuning, with attacks achieving near 100% success rates, and proposed a defense method using a Poisoned Sample Identification Module (PSIM) that shows competitive performance in mitigating these attacks.

Recently, various parameter-efficient fine-tuning (PEFT) strategies for application to language models have been proposed and successfully implemented. However, this raises the question of whether PEFT, which only updates a limited set of model parameters, constitutes security vulnerabilities when confronted with weight-poisoning backdoor attacks. In this study, we show that PEFT is more susceptible to weight-poisoning backdoor attacks compared to the full-parameter fine-tuning method, with pre-defined triggers remaining exploitable and pre-defined targets maintaining high confidence, even after fine-tuning. Motivated by this insight, we developed a Poisoned Sample Identification Module (PSIM) leveraging PEFT, which identifies poisoned samples through confidence, providing robust defense against weight-poisoning backdoor attacks. Specifically, we leverage PEFT to train the PSIM with randomly reset sample labels. During the inference process, extreme confidence serves as an indicator for poisoned samples, while others are clean. We conduct experiments on text classification tasks, five fine-tuning strategies, and three weight-poisoning backdoor attack methods. Experiments show near 100% success rates for weight-poisoning backdoor attacks when utilizing PEFT. Furthermore, our defensive approach exhibits overall competitive performance in mitigating weight-poisoning backdoor attacks.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes