Machine translation evaluation in the context of scholarly communication

Sponsor
OPERAS
Partners
CrossLang

About MT evaluation for scholarly communication

This study is part of the 'Translations and Open Science' project, which explores the possibility of using machine translation (MT) to disseminate research results in different languages. In 2020, the French Ministry of Higher Education and Research launched the Translations and Open Science project with the aim to explore the opportunities offered by translation technologies to foster multilingualism in scholarly communication and thus help to remove language barriers according to Open Science principles.
 
This study focuses on the evaluation of machine translation for the English-French language pair in three scientific disciplines. It combines automatic and human evaluation methods to assess the usability of the machine translation output for the three different usage personas: translators, domain experts and laypersons.
 
LT3 is acting as a subcontractor of CrossLang in this study and is responsible for the fine-grained error annotation of MT output and the assessment of MT quality for laypersons by conducting a reading experiment.
 
The study is part of a larger project funded by OPERAS (https://operas-eu.org/), on behalf of the French Ministry of Higher Education and Research.