Exploring NMT Explainability for Translators Using NMT Visualising Tools

Gabriela Gonzalez-Saez (1, 2, 3) , Mariam Nakhlé (1, 2, 3, 4) , James Robert Turner (5) , Fabien Lopez (1, 2, 3) , Nicolas Ballier (6, 7, 8) , Marco Dinarelli (1, 2, 3) , Emmanuelle Esperança-Rodier (1, 2, 3) , Sui He (1, 2, 3) , Raheel Qader (4) , Caroline Rossi (1, 3, 9) , Didier Schwab (1, 2, 3) , Jun Yang (5)

1 UGA - Université Grenoble Alpes
2 LIG - Laboratoire d'Informatique de Grenoble
3 GETALP - Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole
4 Lingua Custodia
5 Swansea University
6 UPCité - Université Paris Cité
7 LLF - UMR7110 - Laboratoire de Linguistique Formelle
8 CLILLAC-ARP (URP_3967) - Centre de Linguistique Inter-langues, de Lexicologie, de Linguistique Anglaise et de Corpus
9 ILCEA4 - Institut des Langues et Cultures d'Europe, Amérique, Afrique, Asie et Australie

This paper describes work in progress on Visualisation tools to foster collaborations between translators and computational scientists. We aim to describe how visualisation features can be used to explain translation and NMT outputs. We tested several visualisation functionalities with three NMT models based on Chinese-English, Spanish-English and French-English language pairs. We created three demos containing different visualisation tools and analysed them within the framework of performance-explainability, focusing on the translator’s perspective.

Domaines: Informatique [cs]

Gabriela Gonzalez-Saez, Mariam Nakhlé, James Robert Turner, Fabien Lopez, Nicolas Ballier, et al.. Exploring NMT Explainability for Translators Using NMT Visualising Tools. EAMT : European Association for Machine Translation, Jun 2024, Sheffield, United Kingdom. ⟨hal-04581586⟩