Work on SCATE project presented at EAMT in Prague

On Wednesday May 31st 2017, Arda Tezcan presented his latest work on quality estimation: a Neural Network (NN) architecture for detecting grammatical errors in Statistical Machine Translation (SMT) using monolingual morpho-syntactic word representations in combination with surface and syntactic context windows. Click here to view the paper.

The day before Vincent Vandeghinste presented the progress on the SCATE project.

The SCATE project primarily aims at the improvement of translators' efficiency. The project involves various Belgian universities and research groups and is coordinated by the Centre for Computational Linguistics (University of Leuven).