Reevaluating Adversarial Examples in Natural Language

John Morris, Eli Lifland, Jack Lanchantin, Yangfeng Ji, Yanjun Qi


Abstract
State-of-the-art attacks on NLP models lack a shared definition of a what constitutes a successful attack. We distill ideas from past work into a unified framework: a successful natural language adversarial example is a perturbation that fools the model and follows some linguistic constraints. We then analyze the outputs of two state-of-the-art synonym substitution attacks. We find that their perturbations often do not preserve semantics, and 38% introduce grammatical errors. Human surveys reveal that to successfully preserve semantics, we need to significantly increase the minimum cosine similarities between the embeddings of swapped words and between the sentence encodings of original and perturbed sentences. With constraints adjusted to better preserve semantics and grammaticality, the attack success rate drops by over 70 percentage points.
Anthology ID:
2020.findings-emnlp.341
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2020
Month:
November
Year:
2020
Address:
Online
Editors:
Trevor Cohn, Yulan He, Yang Liu
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
3829–3839
Language:
URL:
https://aclanthology.org/2020.findings-emnlp.341
DOI:
10.18653/v1/2020.findings-emnlp.341
Bibkey:
Cite (ACL):
John Morris, Eli Lifland, Jack Lanchantin, Yangfeng Ji, and Yanjun Qi. 2020. Reevaluating Adversarial Examples in Natural Language. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3829–3839, Online. Association for Computational Linguistics.
Cite (Informal):
Reevaluating Adversarial Examples in Natural Language (Morris et al., Findings 2020)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingest-2024-clasp/2020.findings-emnlp.341.pdf
Optional supplementary material:
 2020.findings-emnlp.341.OptionalSupplementaryMaterial.pdf
Code
 QData/TextAttack +  additional community code
Data
MultiNLISNLI