RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP Models

Abstract

Backdoor attacks, which maliciously control a well-trained model's outputs of the instances with specific triggers, are recently shown to be serious threats to the safety of reusing deep neural networks (DNNs). In this work, we propose an efficient online defense mechanism based on robustness-aware perturbations. Specifically, by analyzing the backdoor training process, we point out that there exists a big gap of robustness between poisoned and clean samples. Motivated by this observation, we construct a word-based robustness-aware perturbation to distinguish poisoned samples from clean samples to defend against the backdoor attacks on natural language processing (NLP) models. Moreover, we give a theoretical analysis about the feasibility of our robustness-aware perturbation-based defense method. Experimental results on sentiment analysis and toxic detection tasks show that our method achieves better defending performance and much lower computational costs than existing online defense methods. Our code is available at https://github.com/lancopku/RAP.

arXiv commentEMNLP 2021 (main conference), long paper, camera-ready version

References in corpus (9)

• Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning
• BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain
• Februus: Input Purification Defense Against Trojan Attacks on Deep Neural Network Systems
• Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks
• Rethinking the Trigger of Backdoor Attack
• Can Adversarial Weight Perturbations Inject Neural Backdoors?
• Natural Backdoor Attack on Text Data
• Design and Evaluation of a Multi-Domain Trojan Detection Method on Deep Neural Networks
• Universal Adversarial Attacks with Natural Triggers for Text Classification