VertAttack: Taking Advantage of Text Classifiers’ Horizontal Vision

Jonathan Rusert


Abstract
Text classification systems have continuouslyimproved in performance over the years. How-ever, nearly all current SOTA classifiers have asimilar shortcoming, they process text in a hor-izontal manner. Vertically written words willnot be recognized by a classifier. In contrast,humans are easily able to recognize and readwords written both horizontally and vertically.Hence, a human adversary could write problem-atic words vertically and the meaning wouldstill be preserved to other humans. We simulatesuch an attack, VertAttack. VertAttack identifieswhich words a classifier is reliant on and thenrewrites those words vertically. We find thatVertAttack is able to greatly drop the accuracyof 4 different transformer models on 5 datasets.For example, on the SST2 dataset, VertAttackis able to drop RoBERTa’s accuracy from 94 to13%. Furthermore, since VertAttack does notreplace the word, meaning is easily preserved.We verify this via a human study and find thatcrowdworkers are able to correctly label 77%perturbed texts perturbed, compared to 81% ofthe original texts. We believe VertAttack offersa look into how humans might circumvent clas-sifiers in the future and thus inspire a look intomore robust algorithms.
Anthology ID:
2024.naacl-long.41
Original:
2024.naacl-long.41v1
Version 2:
2024.naacl-long.41v2
Volume:
Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)
Month:
June
Year:
2024
Address:
Mexico City, Mexico
Editors:
Kevin Duh, Helena Gomez, Steven Bethard
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
719–732
Language:
URL:
https://aclanthology.org/2024.naacl-long.41
DOI:
Bibkey:
Cite (ACL):
Jonathan Rusert. 2024. VertAttack: Taking Advantage of Text Classifiers’ Horizontal Vision. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 719–732, Mexico City, Mexico. Association for Computational Linguistics.
Cite (Informal):
VertAttack: Taking Advantage of Text Classifiers’ Horizontal Vision (Rusert, NAACL 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.naacl-long.41.pdf
Copyright:
 2024.naacl-long.41.copyright.pdf