The process of working with machine-translated content is called post-editing of machine translation (PEMT), and tasks the translator with checking accuracy, polishing style and making sure the machine has worked its ‘magic’ properly.
So, I hear you ask, what’s the problem?
At best, MT can be mostly accurate, with some polishing necessary to bring the quality up to the required standard. At worst, it can be incoherent – essentially nonsense – meaning the translator needs to start from scratch and translate as they would on a standard job, but with less time and probably lower pay. As you can imagine, the unpredictability of current MT quality causes fear and trepidation in the translation community, leading to resistance to working with any MT tools.
From a provider’s point of view, training an MT machine is a very time-consuming and resource-heavy activity. It also requires mountains of data, meaning it’s not an activity that can be taken on lightly. That said, if the above is available, having an in-house machine certainly has its merits.
The question now stands: where do we go from here?
Surrey Translation Bureau have been delving into the range of MT offerings out there and will be presenting our findings at the ITI Conference, taking place in Sheffield from 10 to 11 May. The conference is a national gathering of those working in the industry and will offer 36 talks over four tracks including technology and interpreting. It also provides an excellent opportunity to meet peers, clients and providers face-to-face. Come and say hi! We’d love to meet you.
If you can’t make it to Sheffield but would like to know whether PEMT is suitable for you, please give us a call on +44 (0) 1252 730014 or email email@example.com to speak to one of our qualified professional linguists.
Written by Jessica Truelsen