Counterfactually Augmented Data and Unintended Bias: The Case of Sexism and Hate Speech Detection
Indira Sen, Mattia Samory, Claudia Wagner, Isabelle Augenstein
Abstract
Counterfactually Augmented Data (CAD) aims to improve out-of-domain generalizability, an indicator of model robustness. The improvement is credited to promoting core features of the construct over spurious artifacts that happen to correlate with it. Yet, over-relying on core features may lead to unintended model bias. Especially, construct-driven CAD—perturbations of core features—may induce models to ignore the context in which core features are used. Here, we test models for sexism and hate speech detection on challenging data: non-hate and non-sexist usage of identity and gendered terms. On these hard cases, models trained on CAD, especially construct-driven CAD, show higher false positive rates than models trained on the original, unperturbed data. Using a diverse set of CAD—construct-driven and construct-agnostic—reduces such unintended bias.
BibTeX
@inproceedings{sen-etal-2022-counterfactually,
title = "Counterfactually Augmented Data and Unintended Bias: The Case of Sexism and Hate Speech Detection",
author = "Sen, Indira and
Samory, Mattia and
Wagner, Claudia and
Augenstein, Isabelle",
editor = "Carpuat, Marine and
de Marneffe, Marie-Catherine and
Meza Ruiz, Ivan Vladimir",
booktitle = "Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
month = jul,
year = "2022",
address = "Seattle, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.naacl-main.347/",
doi = "10.18653/v1/2022.naacl-main.347",
pages = "4716--4726"
}