Post-editing of machine translations
Launching into the MTPE future with comtexto
What is post-editing?
Machine translation (MT) in combination with artificial intelligence or an artificial neural network (hence NMT for Neural Machine Translation) now makes translations fast and cost-effective. Even though the quality of the translations generated has improved in recent years, they are primarily used to gain information quickly or for internal communication, for example. Since an error-free and consistent translation is not yet possible with machine translation, professional post-editing of the machine-generated texts nevertheless remains necessary.
The extent of post-editing required largely depends on the raw machine output and the subsequent purpose of the translation. Post-editing at comtexto is carried out by experienced specialist translators who have received specific training. They correct, supplement and harmonise the machine-generated translation and adapt it to the relevant specialist terminology and, if necessary, to the customer’s corporate language.
comtexto is specifically certified for post-editing and MT technology, among other things. ISO 18587 certification ensures that our services meet a verifiable standard and sets out the requirements for all those involved in the process. In addition, ISO 18587 certification guarantees that we use only qualified translators who regularly undergo further training in the field of post-editing.
What are the different types of post-editing?
There are two types of post-editing depending on the requirements the text is expected to fulfil and on the area of application.
Light Post-Editing
The aim of light post-editing is simply to ensure that the text is comprehensible. The style and flow of the text are not modified. Even punctuation that is not entirely accurate, or the occasional uncorrected spelling mistake, can be tolerated. The reader must be able to understand the information in the text.
Full Post-Editing
After full post-editing, on the other hand, you should no longer be able to tell that the text has been machine translated. Its quality is similar to that achieved with a professional human translation.