Hard label adversarial attack with high query efficiency against NLP models
Hard label adversarial attack with high query efficiency against NLP models
About this item
Full title
Author / Creator
Qiu, Shilin , Liu, Qihe , Zhou, Shijie , Gou, Min , Zeng, Yi , Zhang, Zhun and Wu, Zhewei
Publisher
London: Nature Publishing Group UK
Journal title
Language
English
Formats
Publication information
Publisher
London: Nature Publishing Group UK
Subjects
More information
Scope and Contents
Contents
Current black-box adversarial attacks have demonstrated significant efficacy in creating adversarial texts against natural language processing models, exposing potential robustness vulnerabilities of these models. However, present attack techniques exhibit inefficiency due to their failure to account for the query counts needed in the adversarial t...
Alternative Titles
Full title
Hard label adversarial attack with high query efficiency against NLP models
Authors, Artists and Contributors
Author / Creator
Identifiers
Primary Identifiers
Record Identifier
TN_cdi_doaj_primary_oai_doaj_org_article_245cf27b1d1f4d80bf89f62d729e1a02
Permalink
https://devfeature-collection.sl.nsw.gov.au/record/TN_cdi_doaj_primary_oai_doaj_org_article_245cf27b1d1f4d80bf89f62d729e1a02
Other Identifiers
ISSN
2045-2322
E-ISSN
2045-2322
DOI
10.1038/s41598-025-93566-5
