Misleading Relation Classifiers by Substituting Words in Texts

Tian Jiang, Yunqi Liu, Yan Feng, Yuqing Li, Xiaohui Cui


Abstract
Relation classification is to determine the semantic relationship between two entities in a given sentence. However, many relation classifiers are vulnerable to adversarial attacks, which is using adversarial examples to lead victim models to output wrong results. In this paper, we propose a simple but effective method for misleading relation classifiers. We first analyze the most important parts of speech (POSs) from the syntax and morphology perspectives, then we substitute words labeled with these POS tags in original samples with synonyms or hyponyms. Experimental results show that our method can generate adversarial texts of high quality, and most of the relationships between entities can be correctly identified in the process of human evaluation. Furthermore, the adversarial examples generated by our method possess promising transferability, and they are also helpful for improving the robustness of victim models.
Anthology ID:
2023.findings-acl.232
Volume:
Findings of the Association for Computational Linguistics: ACL 2023
Month:
July
Year:
2023
Address:
Toronto, Canada
Editors:
Anna Rogers, Jordan Boyd-Graber, Naoaki Okazaki
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
3770–3785
Language:
URL:
https://aclanthology.org/2023.findings-acl.232
DOI:
10.18653/v1/2023.findings-acl.232
Bibkey:
Cite (ACL):
Tian Jiang, Yunqi Liu, Yan Feng, Yuqing Li, and Xiaohui Cui. 2023. Misleading Relation Classifiers by Substituting Words in Texts. In Findings of the Association for Computational Linguistics: ACL 2023, pages 3770–3785, Toronto, Canada. Association for Computational Linguistics.
Cite (Informal):
Misleading Relation Classifiers by Substituting Words in Texts (Jiang et al., Findings 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/emnlp22-frontmatter/2023.findings-acl.232.pdf