Token-modification adversarial attacks for natural language processing: A survey
AI Communications · 18.09.2024 · 37 (4) · 655-676
DOI: 10.3233/aic-230279 ↗Kurz gefasst
Fachartikel in der Zeitschrift AI Communications, erschienen am 18.09.2024. Verfasst von Tom Roth und 4 weiteren Autor:innen. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 94 Literaturangaben.
Abstract (Originalsprache)
Many adversarial attacks target natural language processing systems, most of which succeed through modifying the individual tokens of a document. Despite the apparent uniqueness of each of these attacks, fundamentally they are simply a distinct configuration of four components: a goal function, allowable transformations, a search method, and constraints. In this survey, we systematically present the different components used throughout the literature, using an attack-independent framework which allows for easy comparison and categorisation of components. Our work aims to serve as a comprehensive guide for newcomers to the field and to spark targeted research into refining the individual attack components.
Zitieren
APA
Roth, T., Gao, Y., Abuadbba, A., Nepal, S., & Liu, W. (2024). Token-modification adversarial attacks for natural language processing: A survey. AI Communications, 37(4), 655-676. https://doi.org/10.3233/aic-230279
Vancouver
Roth T, Gao Y, Abuadbba A, Nepal S, Liu W. Token-modification adversarial attacks for natural language processing: A survey. AI Communications. 2024;37(4):655-676. doi:10.3233/aic-230279
BibTeX
@article{Roth2024,
author = {Roth, Tom and Gao, Yansong and Abuadbba, Alsharif and Nepal, Surya and Liu, Wei},
title = {Token-modification adversarial attacks for natural language processing: A survey},
journal = {AI Communications},
year = {2024},
volume = {37},
number = {4},
pages = {655--676},
doi = {10.3233/aic-230279}
}
EUVIMED