Necessity and sufficiency for explaining text classifiers: a case study in hate speech detection

Par Conseil national de recherches du Canada

Téléchargement	Voir la version finale : Necessity and sufficiency for explaining text classifiers: a case study in hate speech detection (PDF, 407 Kio)
DOI	Trouver le DOI : https://doi.org/10.18653/v1/2022.naacl-main.192
Auteur	Rechercher : Balkir, Esma¹; Rechercher : Nejadgholi, Isar¹; Rechercher : Fraser, Kathleen¹; Rechercher : Kiritchenko, Svetlana¹
Affiliation	Conseil national de recherches du Canada. Technologies numériques
Format	Texte, Article
Conférence	Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL 2022), July 10-15, 2022, Seattle, Washington (Hybrid Conference)
Résumé	We present a novel feature attribution method for explaining text classifiers, and analyze it in the context of hate speech detection. Although feature attribution models usually provide a single importance score for each token, we instead provide two complementary and theoretically-grounded scores – necessity and sufficiency – resulting in more informative explanations. We propose a transparent method that calculates these values by generating explicit perturbations of the input text, allowing the importance scores themselves to be explainable. We employ our method to explain the predictions of different hate speech detection models on the same set of curated examples from a test suite, and show that different values of necessity and sufficiency for identity terms correspond to different kinds of false positive errors, exposing sources of classifier bias against marginalized groups.
Date de publication	2022-07-15
Maison d’édition	Association for Computational Linguistics (ACL)
Licence	Creative Commons Attribution 4.0 International (CC BY 4.0) https://creativecommons.org/licenses/by/4.0/deed.fr
Dans	Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL 2022) (15 juillet 2022) : 2672–2686. https://doi.org/10.18653/v1/2022.naacl-main.
Langue	anglais
Publications évaluées par des pairs	Oui
Exporter la notice	Exporter en format RIS
Signaler une correction	Signaler une correction (s'ouvre dans un nouvel onglet)
Identificateur de l’enregistrement	9c698f67-b648-4edb-866a-4106252f0b1b
Enregistrement créé	2022-09-09
Enregistrement modifié	2022-09-12

Date de modification :: 2024-07-04