“Le professeur est très intelligent / La prof est très attirante”. Recognizing and Reducing Gender Bias in Neural Machine Translation

Abstract

The encounter between Translation Studies and Gender Studies has proven extremely important for the investigation of the issues related to gender representation that are implicit in every natural language, and that processes of translation help to futher unveil. The rapid evolution of Neural Machine Translation over the last years, and the overwhelming presence of machine translation in the contemporary society, call for an investigation of how these issues of gender representation have evolved in the context of machine translation. After outlining the encounter between Translation Studies and Gender Studies, as well as the evolution of machine translation over the last decades, this article will focus on the definition of the issue of gender bias in Neural Machine Translation, on its origins and consequences, and on the attempts that have been made at eliminating or at least reducing gender bias in machine translation.