Cyberbullying classifiers are sensitive to model-agnostic perturbations

Autor: Chris Emmery, Akos Kádár, Grzegorz Chrupała, Daelemans, W. M. P.

Publikováno v: 13th International Conference on Language Resources and Evaluation, (LREC 2022), Juin 20-25, 2022, Marseille, France
ISSUE=13;TITLE=Language Resources and Evaluation Conference
Tilburg University-PURE

A limited amount of studies investigates the role of model-agnostic adversarial behavior in toxic content classification. As toxicity classifiers predominantly rely on lexical cues, (deliberately) creative and evolving language-use can be detrimental

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::64909f8c788adaf662fec30367ad4d6f
https://hdl.handle.net/10067/1944180151162165141