Fair Text Classification with Wasserstein Independence - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Fair Text Classification with Wasserstein Independence

Thibaud Leteno
Antoine Gourru
Rémi Emonet
Christophe Gravier

Résumé

Group fairness is a central research topic in text classification, where reaching fair treatment between sensitive groups (e.g. women vs. men) remains an open challenge. This paper presents a novel method for mitigating biases in neural text classification, agnostic to the model architecture. Considering the difficulty to distinguish fair from unfair information in a text encoder, we take inspiration from adversarial training to induce Wasserstein independence between representations learned to predict our target label and the ones learned to predict some sensitive attribute. Our approach provides two significant advantages. Firstly, it does not require annotations of sensitive attributes in both testing and training data. This is more suitable for real-life scenarios compared to existing methods that require annotations of sensitive attributes at train time. Secondly, our approach exhibits a comparable or better fairness-accuracy trade-off compared to existing methods.
Fichier principal
Vignette du fichier
2023.emnlp-main.978.pdf (618.97 Ko) Télécharger le fichier
Origine Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-04390768 , version 1 (12-01-2024)

Identifiants

Citer

Thibaud Leteno, Antoine Gourru, Charlotte Laclau, Rémi Emonet, Christophe Gravier. Fair Text Classification with Wasserstein Independence. 2023 Conference on Empirical Methods in Natural Language Processing, Dec 2023, Singapore, Singapore. pp.15790-15803, ⟨10.18653/v1/2023.emnlp-main.978⟩. ⟨hal-04390768⟩
39 Consultations
20 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More