He, Xuanli; Xu, Qiongkai; Wang, Jun; Rubinstein, Benjamin; Cohn, Trevor; (2023) Mitigating Backdoor Poisoning Attacks through the Lens of Spurious Correlation. In: Bouamor, Houda and Pino,, Juan and Bali, Kalika, (eds.) Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. (pp. pp. 953-967). Association for Computational Linguistics: Singapore, Singapore.

Preview

PDF 2023.emnlp-main.60.pdf - Published Version Download (467kB) | Preview

Abstract

Modern NLP models are often trained over large untrusted datasets, raising the potential for a malicious adversary to compromise model behaviour. For instance, backdoors can be implanted through crafting training instances with a specific textual trigger and a target label. This paper posits that backdoor poisoning attacks exhibit a spurious correlation between simple text features and classification labels, and accordingly, proposes methods for mitigating spurious correlation as means of defence. Our empirical study reveals that the malicious triggers are highly correlated to their target labels; therefore such correlations are extremely distinguishable compared to those scores of benign features, and can be used to filter out potentially problematic instances. Compared with several existing defences, our defence method significantly reduces attack success rates across backdoor attacks, and in the case of insertion-based attacks, our method provides a near-perfect defence.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item