Everyone in the translation industry has an opinion about machine translation – you either hate it with passion or support it wholeheartedly. Discussions are held about it at every major translation conference, bringing together its biggest supporters and its biggest skeptics.  

Of course it’s a little funny when Google Translate tries to tell you that some machine comes with the season of spring (kevad) and nuts that grow on trees (pähklid), even though logic says that ‘nut’ in machines refers to a type of fastener (mutter) and ‘spring’ to an elastic coiled piece of metal. If a text produced like that is not properly post-edited, then you end up with an unintelligible user manual. 

This is why it’s so important to develop and use machine translation correctly and to incorporate it into workflow properly. A machine translation engine that is developed with a specific subject field and target group in mind wouldn’t make these types of mistakes because it is ‘taught’ to operate in a technical context, unlike Google Translate, which simply absorbs absolutely everything. We believe that if we were to further incorporate Grata’s experienced technical translators and editors into this type of a project, it would be possible to improve the quality and the reputation of machine translation even in the case of such a complicated and grammar-heavy language as Estonian. 

Therefore, we have decided that now is precisely the right time to jump on the machine translation bandwagon. Grata’s own machine translation system will be developed in cooperation with the University of Tartu by mid-summer of 2016. The goal of the project is to completley intergrate it into our workflow by the end of the year. As a result, we hope to increase our productivity and offer our clients new types of services made available by machine translation. 

The European Regional Development Fund finances the project with 4,000 EUR.