Perturbation Augmentation for Fairer NLP

Rebecca Qian, Candace Ross, Jude Fernandes, Eric Michael Smith, Douwe Kiela, Adina Williams


Abstract
Unwanted and often harmful social biases are becoming ever more salient in NLP research, affecting both models and datasets. In this work, we ask whether training on demographically perturbed data leads to fairer language models. We collect a large dataset of human annotated text perturbations and train a neural perturbation model, which we show outperforms heuristic alternatives. We find that (i) language models (LMs) pre-trained on demographically perturbed corpora are typically more fair, and (ii) LMs finetuned on perturbed GLUE datasets exhibit less demographic bias on downstream tasks, and (iii) fairness improvements do not come at the expense of performance on downstream tasks. Lastly, we discuss outstanding questions about how best to evaluate the (un)fairness of large language models. We hope that this exploration of neural demographic perturbation will help drive more improvement towards fairer NLP.
Anthology ID:
2022.emnlp-main.646
Volume:
Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
9496–9521
Language:
URL:
https://aclanthology.org/2022.emnlp-main.646
DOI:
10.18653/v1/2022.emnlp-main.646
Bibkey:
Cite (ACL):
Rebecca Qian, Candace Ross, Jude Fernandes, Eric Michael Smith, Douwe Kiela, and Adina Williams. 2022. Perturbation Augmentation for Fairer NLP. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 9496–9521, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Perturbation Augmentation for Fairer NLP (Qian et al., EMNLP 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-1/2022.emnlp-main.646.pdf