How sensitive are translation systems to extra contexts? Mitigating gender bias in Neural Machine Translation models through relevant contexts.

Shanya Sharma, Manan Dey, Koustuv Sinha


Abstract
Neural Machine Translation systems built on top of Transformer-based architectures are routinely improving the state-of-the-art in translation quality according to word-overlap metrics. However, a growing number of studies also highlight the inherent gender bias that these models incorporate during training, which reflects poorly in their translations. In this work, we investigate whether these models can be instructed to fix their bias during inference using targeted, guided instructions as contexts. By translating relevant contextual sentences during inference along with the input, we observe large improvements in reducing the gender bias in translations, across three popular test suites (WinoMT, BUG, SimpleGen). We further propose a novel metric to assess several large pre-trained models (OPUS-MT, M2M-100) on their sensitivity towards using contexts during translation to correct their biases. Our approach requires no fine-tuning, and thus can be used easily in production systems to de-bias translations from stereotypical gender-occupation bias. We hope our method, along with our metric, can be used to build better, bias-free translation systems.
Anthology ID:
2022.findings-emnlp.143
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2022
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1968–1984
Language:
URL:
https://aclanthology.org/2022.findings-emnlp.143
DOI:
10.18653/v1/2022.findings-emnlp.143
Bibkey:
Cite (ACL):
Shanya Sharma, Manan Dey, and Koustuv Sinha. 2022. How sensitive are translation systems to extra contexts? Mitigating gender bias in Neural Machine Translation models through relevant contexts.. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 1968–1984, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
How sensitive are translation systems to extra contexts? Mitigating gender bias in Neural Machine Translation models through relevant contexts. (Sharma et al., Findings 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingest-2024-clasp/2022.findings-emnlp.143.pdf
Video:
 https://preview.aclanthology.org/ingest-2024-clasp/2022.findings-emnlp.143.mp4