ACL 2023findings1 citations

VoteTRANS: Detecting Adversarial Text without Training by Voting on Hard Labels of Transformations

Hoang-Quoc Nguyen-Son, Seira Hidano, Kazuhide Fukushima, Shinsaku Kiyomoto, Isao Echizen

Abstract

Adversarial attacks reveal serious flaws in deep learning models. More dangerously, these attacks preserve the original meaning and escape human recognition. Existing methods for detecting these attacks need to be trained using original/adversarial data. In this paper, we propose detection without training by voting on hard labels from predictions of transformations, namely, VoteTRANS. Specifically, VoteTRANS detects adversarial text by comparing the hard labels of input text and its transformation. The evaluation demonstrates that VoteTRANS effectively detects adversarial text across various state-of-the-art attacks, models, and datasets.

BibTeX
@inproceedings{nguyen-son-etal-2023-votetrans,
    title = "{V}ote{TRANS}: Detecting Adversarial Text without Training by Voting on Hard Labels of Transformations",
    author = "Nguyen-Son, Hoang-Quoc  and
      Hidano, Seira  and
      Fukushima, Kazuhide  and
      Kiyomoto, Shinsaku  and
      Echizen, Isao",
    editor = "Rogers, Anna  and
      Boyd-Graber, Jordan  and
      Okazaki, Naoaki",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
    month = jul,
    year = "2023",
    address = "Toronto, Canada",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.findings-acl.315/",
    doi = "10.18653/v1/2023.findings-acl.315",
    pages = "5090--5104"
}