Triggerless Backdoor Attack for NLP Tasks with Clean Labels

Leilei Gan, Jiwei Li, Tianwei Zhang, Xiaoya Li, Yuxian Meng, Fei Wu, Yi Yang, Shangwei Guo, Chun Fan


Abstract
Backdoor attacks pose a new threat to NLP models. A standard strategy to construct poisoned data in backdoor attacks is to insert triggers (e.g., rare words) into selected sentences and alter the original label to a target label. This strategy comes with a severe flaw of being easily detected from both the trigger and the label perspectives: the trigger injected, which is usually a rare word, leads to an abnormal natural language expression, and thus can be easily detected by a defense model; the changed target label leads the example to be mistakenly labeled, and thus can be easily detected by manual inspections. To deal with this issue, in this paper, we propose a new strategy to perform textual backdoor attack which does not require an external trigger and the poisoned samples are correctly labeled. The core idea of the proposed strategy is to construct clean-labeled examples, whose labels are correct but can lead to test label changes when fused with the training set. To generate poisoned clean-labeled examples, we propose a sentence generation model based on the genetic algorithm to cater to the non-differentiable characteristic of text data. Extensive experiments demonstrate that the proposed attacking strategy is not only effective, but more importantly, hard to defend due to its triggerless and clean-labeled nature. Our work marks the first step towards developing triggerless attacking strategies in NLP.
Anthology ID:
2022.naacl-main.214
Volume:
Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Month:
July
Year:
2022
Address:
Seattle, United States
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2942–2952
Language:
URL:
https://aclanthology.org/2022.naacl-main.214
DOI:
10.18653/v1/2022.naacl-main.214
Bibkey:
Cite (ACL):
Leilei Gan, Jiwei Li, Tianwei Zhang, Xiaoya Li, Yuxian Meng, Fei Wu, Yi Yang, Shangwei Guo, and Chun Fan. 2022. Triggerless Backdoor Attack for NLP Tasks with Clean Labels. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2942–2952, Seattle, United States. Association for Computational Linguistics.
Cite (Informal):
Triggerless Backdoor Attack for NLP Tasks with Clean Labels (Gan et al., NAACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.naacl-main.214.pdf
Code
 leileigan/clean_label_textual_backdoor_attack
Data
OLIDSST