The growth of global trade is rapidly accelerating and giving rise to the need to translate increasing volumes of information. That’s why companies are looking for ways to translate their expanding content without their costs going through the roof. The best translation agencies know that machine translation can be a good solution for achieving this goal. As a result of the unstoppable progress of technology and use of artificial intelligence, machine translation can achieve very good results in certain circumstances.
WHAT'S MACHINE TRANSLATION EXACTLY?
According to Wikipedia, machine translation is a computational linguistics field that researches using software to translate text or speech from one language to another. Machine translation shouldn’t be confused with computer-assisted translation (CAT), which refers to translators using software as a tool when they produce their translations.
The early days of machine translation involved directly substituting the words of the source language for equivalent words in the target language. But, as you may expect, translating involves much more than just replacing some words with others. As a result, that basic model has been successively refined with models able to understand the meaning of phrases or syntagmas and put them to statistical use through text corpuses.
To improve accuracy in machine translation, the software is prepared to translate texts in a specific field. That way, it’s easier to remove ambiguities, and the margin of error is much smaller. Machine translation also generates much better results for content with a formal style, formulaic language and repetitive phrases.
To get an error-free result, machine translations must pass through a review process, known in the sector as post editing. Put simply, post editing means a reviewer checking the result of the machine translation. Obviously, the better the raw machine translation output, the less time will be needed to polish the translation. In the sector, we say that the “edit distance” is higher or lower depending on how much time has to be spent reviewing the translation.
The latest revolution in the field of machine translation uses neural networks, where a computing system is able to process millions of examples and use them as the basis for making decisions.
CAN YOU USE MACHINE TRANSLATION FOR EVERYTHING?
Unfortunately, no. As we mentioned earlier, machine translation can provide fairly acceptable results in defined fields, where the likelihood of ambiguity is lower. What’s more, the production team and translation agency must be properly aligned to get the highest possible performance out of machine translation. Texts destined to be translated through machine translation must have a certain style to obtain the desired translation result.
A text that’s too creative, like in marketing translations and literary translations, or one that’s particularly complex, wouldn’t be a good candidate for machine translation.
This is just a short introduction to machine translation. If you’d like to know if it could be used with your texts, then contact your translation agency to enquire about your specific case.