United States

Machine Translation with Human Post-Editing (MT + PE)

Machine translation (MT) is the fully automated translation of text into another language. MT has improved in recent years, owing to better software and faster computers. Well-known purchasable MT tools in the language industry include IBM Watson, memoQ and DeepL.

It is now possible for computers to learn how words relate to one another and what they mean in different languages. This is achieved using statistical models and artificial intelligence, e.g. in neural machine translation (NMT). These machines, programmed in specific fields, are able to translate speedily large quantities of usable text.

However, although text translated in this way is understandable, it often reads rather wooden. Jokes, sarcasm, alliteration, subtleties and, unfortunately, even meaning are often lost. Such a text may also give the reader the impression that your product or service is of inferior quality.

Translations intended to persuade the reader should, of course, read well and fit the right cultural context. After all, you want to evoke a positive feeling and achieve results with your text. This is why it makes sense to have machine translation edited by a native speaker of the target language (post-editing or PE).

Many traditional translation agencies are resistant to the emergence of machine translations. Our agency thinks differently. Machine translation is improving every day. Clever application of MT results in major translation-process efficiency gains and cost-savings. However, incorrect use of MT can result in an unusable and unwanted translation. This is why human post-editing remains indispensable to the translation process.

Fasttranslator’s approach to MT + PE is explained in more detail below.