Getting the most out of machine translation as a linguist

Written on 6 September, 2022

Improvements in machine translation, including the substantial developments we have seen in neural machine translation, have opened up a whole world of content for translation that may otherwise have never been translated. However, this technology also has numerous well-known pitfalls, including mistranslations, spelling and grammar issues, and context-inappropriate translations. The good news is that there are steps that we linguists can take to help improve the profitability of post-editing as a service.

Post editing

Use your expertise

As linguists, your experience and knowledge are invaluable when it comes to providing post-editing guidance to your clients. We know that certain genres are far better suited to machine translation than others. With this in mind, you should be steering your clients in the right direction from the start and helping them to make post-editing decisions. It’s up to you to explain to your client why using machine translation for their next marketing campaign may not be the wisest idea. Conversely, you can also use your knowledge to advise clients – having conducted a thorough assessment – that MT could be deployed on certain files to help them save money and reduce turnaround times.

You should use your expertise to help your client see you as a language expert, someone who is able to help them make the right decisions when it comes to machine translation. You are also well placed to explain how, when used correctly, MT could help them cut expenditure, make your life as a linguist and post-editor easier, and improve the profitability of post-editing as a service.

File preparation

One factor that can seriously affect the quality of your MT output is poorly prepared source material. If your file has not been adequately prepared, the engine will struggle to re-assemble the elements of your text during machine processing. This can, in turn, result in poor quality translations. This makes everyone’s life more difficult. As a linguist, you can offer file preparation services to prepare your client’s documents prior to being processed by the machine, a step that can often reduce total project costs despite a higher initial outlay in terms of the costs for file preparation. Proper preparation is key in all aspects of our lives, and translation is no exception.

As a linguist, you can also check through your client’s files to see if there are any areas of the text that do not even need to be fed into their engine; these can include long lists of names, numbers or filler text, or even large sections of the text which are context matches from your TM. As linguists, you know you won’t have to touch these, so you can simply pre-approve this content and continue. Taking these steps will help you cut down the amount of time you spend post-editing, increasing your profitability in the process.

Getting the most out of MT

So, if you are looking to harness the benefits of machine translation but are concerned with the potential downsides, the good news is that there are lots of steps you can take to help make your life as a post-editor easier, as well as various opportunities to sell different linguistic services to your clients, and thereby open up new revenue streams.

Finally, to PEMT or not to PEMT?

Learn how to approach post-editing requests from your clients in this webinar put together by our Senior Project Manager, Greg Hyne for the Institute of Translation and Interpreting.

Written by Greg Hyne