Outils Neuronaux End-to-End pour la TRAduction des Communications (ON-TRAC)

Date: 01/2019 - 01/2021
Funding: ANR
Call: Generic
Partners: LIA (Avignon) (France), LIG (Grenoble) (France), Airbus (France)
URL: https://lium.univ-lemans.fr/en/on-trac/


The ON-TRAC project proposes to radically change the architectures currently used in speech translation. It is based on end-to-end neural models for machine translation and is particularly aimed at lightweight and portable speech translation applications.

Current speech translation systems consist of the sequencing of two modules: the first module performs speech recognition (speech to text in source language) and the second automatically translates the outputs (in the form of strings or word graphs in source language) from the first module to a text in target language. ON-TRAC proposes to adapt and evaluate end-to-end architectures for direct speech translation.