La traduction automatique neuronale : technologie révolutionnaire ou poudre de perlimpinpin ? Compte rendu d’une expérience pédagogique

Abstract

This paper reports on a classroom experiment carried out with Master’s students enrolled in a specialised translation course (English-French) at the University of Liège (Belgium). We present the results from the qualitative analysis of a corpus of five source texts in English, their machine translations and their post-edited versions in French. We decided to use DeepL as MT engine. The main objective of our experiment is first and foremost to provide students with the opportunity to gain perspective on the MT advantages and disadvantages, which still exist in spite of many advances made in the field. Finally, this paper also seeks to contribute to a broader discussion on the major MT challenges faced today by (future) translators.