Build your own Machine Translation System (Ghent, 3/7)

Create your own MT-systemOn Wednesday July 3 CrossLang (Ghent) a hands-on training in Ghent for translation agencies who want to build their own machine translation system (MT-system). The technology that enables users to build an MT system is called Moses. Moses is free open source oftware and is currently seen as a valuable, inexpensive and quickly deployable alternative to expensive commercial software packages or general applications such as Google. As a translation agency, how do you deploy this MT-technology successfully in your own translation production process? That is the subject of this training.

Moses is a statistical machine translation system (SMT) that can be trained on any language pair combination or on any knowledge domain. In principle, all you need is a good parallel corpus - a collection texts with their translations. Building and training Moses is not that complicated. But subsequently incorporating Moses in an existing production environment is already less evident.

Moses Machine Translation is open source technology originally developed at the University of Edinburgh. The European Union finances the continued development of the system via various projects (e.g. EuroMatrix, MosesCore).

CrossLang is an independent consultant agency. Since 2002 CrossLang helps translation companies and large organizations with the automation of translation processes and the integration of technologies for automatic translation, for the translation management and for multilingual content management.

The training will be held in English ( 'Building Open Source MT Systems') and is aimed at employees with a technical profile.

 

For more info contact Nathalie De Sutter (Tel. +32 486 515113 ornathalie.de.sutter@crosslang.com)

More info: www.crosslang.com/en/event/training-building-open-source-mt-systems

More info: www.crosslang.com

More info: www.statmt.org/moses

(dd - 7/6/2013)


Powered by CrossLang

Author: Dries Debackere

Machine translation: SDL Machine Translation (previously SDL BeGlobal)

Post-editing: Lennart Tondeleir

Source language: Nederlands (nl)


Additional information