Neural Machine Translation (NMT)

Neural Machine Translation (NMT) is a technology based on artificial networks of neurons. It has made considerable advances in recent years thanks to artificial intelligence and can now serve as a basis for certain professional translations.

With NMT, millions of pieces of information can be translated in real time, with a level of accuracy and reliability that is currently similar to that of a human being. In our daily lives, we are already familiar with automatic translation software such as Google Translate, but artificial intelligence is a game changer.

Today, the machine – like the human brain – is capable of producing a reliable translation as well as learning a language, and is therefore constantly improving the quality of the translated material. To improve the performance of the machine, it is “trained” by human translators. In practical terms, this means feeding the machine a very large volume of quality data (words, segments of sentences, and texts already translated) in order to improve the reliability and sophistication of the results.

A machine can also be trained to meet the specific needs of a sector (legal translation, medical translation, etc.) or a customer’s area of activity, which will have its own vocabulary.

In just a few years, the Ubiqus Group has become a major player in the creation of neural translation engines that are specialized to the nuances of a particular industry or adapted to a client’s unique terminology.

If human translation and machine translation have been poles apart for a long time, they must now work together.

To guarantee high quality translations, machine translated content must be adapted and re-read by a professional translator.  This quality control, adjustment and correction step is called post-editing.  The aim is to make the final content understandable and easy to read.