To Prefer or to Choose? Generating Agency and Power Counterfactuals Jointly for Gender Bias Mitigation

Maja Stahl, Maximilian Spliethöver, Henning Wachsmuth


Abstract
Gender bias may emerge from an unequal representation of agency and power, for example, by portraying women frequently as passive and powerless (“She accepted her future”) and men as proactive and powerful (“He chose his future”). When language models learn from respective texts, they may reproduce or even amplify the bias. An effective way to mitigate bias is to generate counterfactual sentences with opposite agency and power to the training. Recent work targeted agency-specific verbs from a lexicon to this end. We argue that this is insufficient, due to the interaction of agency and power and their dependence on context. In this paper, we thus develop a new rewriting model that identifies verbs with the desired agency and power in the context of the given sentence. The verbs’ probability is then boosted to encourage the model to rewrite both connotations jointly. According to automatic metrics, our model effectively controls for power while being competitive in agency to the state of the art. In our main evaluation, human annotators favored its counterfactuals in terms of both connotations, also deeming its meaning preservation better.
Anthology ID:
2022.nlpcss-1.6
Volume:
Proceedings of the Fifth Workshop on Natural Language Processing and Computational Social Science (NLP+CSS)
Month:
November
Year:
2022
Address:
Abu Dhabi, UAE
Editors:
David Bamman, Dirk Hovy, David Jurgens, Katherine Keith, Brendan O'Connor, Svitlana Volkova
Venue:
NLP+CSS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
39–51
Language:
URL:
https://aclanthology.org/2022.nlpcss-1.6
DOI:
10.18653/v1/2022.nlpcss-1.6
Bibkey:
Cite (ACL):
Maja Stahl, Maximilian Spliethöver, and Henning Wachsmuth. 2022. To Prefer or to Choose? Generating Agency and Power Counterfactuals Jointly for Gender Bias Mitigation. In Proceedings of the Fifth Workshop on Natural Language Processing and Computational Social Science (NLP+CSS), pages 39–51, Abu Dhabi, UAE. Association for Computational Linguistics.
Cite (Informal):
To Prefer or to Choose? Generating Agency and Power Counterfactuals Jointly for Gender Bias Mitigation (Stahl et al., NLP+CSS 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.nlpcss-1.6.pdf
Note:
 2022.nlpcss-1.6.note.txt