After explaining the difference between CAT Tools and Machine Translation (MT) in our previous blog “ Translation Technologies for the Ultimate Translation”, we now move on to what comes next after performing a machine translation and the reason why we, humans, will never be taken out of the translation process’ equation. At least not for a while!

Improve your Machine Translation

As we mentioned before, machine translation software needs to be trained, either with large amounts of data or corpora  in the case of Statistical MT, with  linguistic rules  in the case of Rule-based MT, or with a mix of both data and rules  in the case of the Hybrid MT. The better the foundation, data and or rules fed to the engine, the better the outcome. This training is done by humans (hint #1 of why we will never become obsolete) but, as we all know, almost no machine output is perfect: they can’t really think. So, no matter how good the initial training was, chances are, a human will need to intervene to fix what the machine couldn’t do perfectly on its own (hint #2). This process of fixing the results obtained after machine translation is known as Postediting.

Officially, postediting is the process of improving a machine-generated translation with required manual labor to obtain the quality desired. There are two types of postediting: Light and full postediting.

Types of Postediting

Light postediting: this level of postedition is the minimal required by the posteditor to give the end user just a general idea of the text’s meaning.  This process is mostly use in cases where the client plans on using the translation for internal purposes only , or when the translated text is needed urgently.

Full postediting: comprises a greater degree of involvement by the posteditor to achieve the required level of quality established at the beginning of the project. This modality is used  in cases where the resulting translation is expected to be both understandable and stylistically appropriate, so it can be of publishable quality, for internal and external usage.

Benefits of Postediting

Time: There is a battle between the industry and academic reports. The former says machine translation-postediting saves around 40% of the time needed while the latter says a number between 15-20% is more realistic. In any case, it is a significant improvement for this important resource in a translation project.

Money: Clients can save up to 20% of their budget if they go with machine translation-postediting combo instead of human translation, if the quality produced is an option for their project (machine translation is not suited for all subjects but, that will be another post of this translation blog).

CAT tools and machine translation are used in aiding the translation process; both are technologies that are improving and getting better day by day. At Eging, we use CAT tools every day and Machine Translation with postediting in software aided translations according to the project and content. The ability to incorporate machine translation, human translation, and a combination of both is a versatility featured in one of our stellar solutions for website translation- ATLAS.

Who wouldn’t want to save time and money without potentially risking quality? Too good to be true? Not when we use the benefits of machine translation and postediting to our advantage, in the appropriate subject matter, the appropriate project type , with the proper quality and the proper resources to do it right!