Mitigating Gender Bias in English-Ukrainian Machine Translation Models

Pavels Ivanovs, Gina Welsh, Irini Selenica


Abstract
This study investigates the presence and mitigation of gender bias in English-Ukrainian machine translation (MT) models. We focused on the transfer of gender bias in two English-Ukrainian MT models, using sentences that contained professional occupation names. We evaluated two gender bias mitigation methods: 1) gender tagging of source sentences, and 2) gender bias correction by Lapa LLM, a Ukrainian large language model (LLM), using a dataset that we curated for our evaluation. Our results showed that both zero-shot models contained English-Ukrainian gender bias transfer, particularly for gender-stereotypical occupations. The gender tagging mitigation method demonstrated changes to the gender assignment in our data; however, these changes led to mixed results in gender bias correction. The Lapa LLM-correction method had promising results in demonstrating a considerable mitigation of bias in our evaluation set. Overall, our study contributes a framework for the evaluation of gender bias in English-Ukrainian translation that could potentially be applied to translation pairs in other languages.
Anthology ID:
2026.eamt-1.14
Volume:
Proceedings of the 26th Annual Conference of the European Association for Machine Translation (Volume 1)
Month:
June
Year:
2026
Address:
Tilburg, The Netherlands
Editors:
Dimitar Shterionov, Eva Vanmassenhove, Mirella De Sisto, Fred Blain, Javad Pourmostafa Roshan Sharami, Lisa Lepp, Chiara Manna, Argentina Anna Rescigno, Alina Karakanta, Ayla Rigouts Terryn, Manuel Lardelli, Natalia Resende, Elena Murgolo, Janiça Hackenbuchner, Anna Zaretskaya, Miquel Esplà-Gomis, Thierry Etchegoyhen, Dagmar Gromann, Rachel Bawden, Barry Haddow, Sara Szoc, Mikel Forcada, Helena Moniz
Venue:
EAMT
SIG:
Publisher:
European Association for Machine Translation
Note:
Pages:
173–189
Language:
URL:
https://aclanthology.org/2026.eamt-1.14/
DOI:
Bibkey:
Cite (ACL):
Pavels Ivanovs, Gina Welsh, and Irini Selenica. 2026. Mitigating Gender Bias in English-Ukrainian Machine Translation Models. In Proceedings of the 26th Annual Conference of the European Association for Machine Translation (Volume 1), pages 173–189, Tilburg, The Netherlands. European Association for Machine Translation.
Cite (Informal):
Mitigating Gender Bias in English-Ukrainian Machine Translation Models (Ivanovs et al., EAMT 2026)
Copy Citation:
PDF:
https://aclanthology.org/2026.eamt-1.14.pdf