Gender Lost In Translation: How Bridging The Gap Between Languages Affects Gender Bias in Zero-Shot Multilingual Translation
This addresses gender bias in machine translation for users of multilingual systems, but it is incremental as it builds on existing methods for bias mitigation.
The study tackled gender bias in zero-shot multilingual neural machine translation by examining how bridging languages without parallel data affects gender preservation, finding that language-agnostic representations reduce masculine bias and that pivoting with gender-inflected bridge languages improves fairness in speaker-related gender agreement.
Neural machine translation (NMT) models often suffer from gender biases that harm users and society at large. In this work, we explore how bridging the gap between languages for which parallel data is not available affects gender bias in multilingual NMT, specifically for zero-shot directions. We evaluate translation between grammatical gender languages which requires preserving the inherent gender information from the source in the target language. We study the effect of encouraging language-agnostic hidden representations on models' ability to preserve gender and compare pivot-based and zero-shot translation regarding the influence of the bridge language (participating in all language pairs during training) on gender preservation. We find that language-agnostic representations mitigate zero-shot models' masculine bias, and with increased levels of gender inflection in the bridge language, pivoting surpasses zero-shot translation regarding fairer gender preservation for speaker-related gender agreement.