Machine Translation

Neural Machine Translation (NMT)

Until recently, machine translation was not promising in Turkish language as proven by its performance so far. Reverse structure of Turkish language, as one of the agglutinative languages, when compared to English and most of other languages also makes translation process more complicated. As a result of Turkey's location and cultural heritage in between Europe, Russia and Middle East; stretching out on the lands where many ancient civilizations passed, Turkish language has some challenges in terms of translation. Even though there are some factors not helping machine translation, new algorithm behind Neural Machine Translation is revolutionary and brought much better results. 

NMT powered by TMs 

NMT can be run in a specialized way, in other words when NMT is fed by translation memories of your previous translations, results will be in compatible with your industry interms of terminology and style.  

Post-editing

Regarding NMT, everything is developing in a rocket speed. But, today human post-editing is necessary. Post-editing of MT is something between translation and proofreading. For this reason, post-editors need to be trained and well instructed.

Extra-large content

NMT is able to learn from previous TMs. This makes it possible to deal with large amount of content by running NMT engines. Needless to say that outcome must be post-edited. 

Project management

If machine translation is to be applied to any project, this means that the project should be large enough and there will be several project members. So, a project manager must take care of glosarries, post-editing, quality assurance, monitoring processes.

Thanks to our project managers, ideally specialized post editors will be matched to your projects, understanding the requirements of your purpose.