Machine Translation Post-Editing
What is Machine Translation Post-Editing?
With the rise of Neural Machine Translation (NMT), the service called machine translation post-editing (MTPE) has become a reality for the language industry. Even languages traditionally characterized by poor output quality, such as Hungarian, show significantly better machine translation results when NMT engines are used.
WHY IS IT SO CRUCIAL?
Human-like quality is impossible to achieve without humans, and this is where linguists come in: they correct the errors found in the raw MT output in order to produce a perfectly understandable and error-free text.
Post-editing is different from proofreading human translation, but also from translation itself – it requires good knowledge of the types of errors that NMT usually produces, and other shortcomings of machine generated sentence structures.
Types of machine translation post-editing
There are generally two types of MTPE that are distinguished by the industry experts: light post-editing and full post-editing (sometimes called hard post-editing).
Light post-editing is usually requested when the timeframe is very strict and the aim is only to understand the basic meaning of the text, or to produce ‘good enough’ quality – in this case, minor grammatical errors such as punctuation can be left in the target text.
Full post-editing, however, aims to produce the same high level of quality as human (reviewed) translation and is often performed by two separate linguists in order to achieve this.
How we work
Reception of materials
Machine translation output assessment
Translation memory update
Delivery to the client
Handling of queries after delivery
Translator proficiency is constantly monitored and maintained
Over 400 linguists ready to help at any time
Two individual quality checks at the end of every project
We take care of multilingual projects into 26 target languages on a daily basis
We are a regional provider with a strong local presence - these languages are our main focus and strength