Machine Translation Post-Editing (MTPE)

We are all aware of the growing importance of artificial intelligence (AI) in many areas of our lives and work.

Machine translation post-editing (MTPE) combines the efficiency of machine translation with the precision and nuance that only a human translator can provide.

MTPE is a process whereby a human translator reviews and edits text that has been translated by a machine translation system. The translator’s input improves the accuracy, readability and overall quality of the result.

This is often deployed in sectors where large volumes of content need to be translated quickly and accurately, such as e-commerce, software localisation and customer support. Whilst this is driven by the objectives of speed and cost-saving, it ultimately combines the effectiveness of machine translation with the quality and reliability of human editing.

There are two main levels involved in the post-editing of machine-translated text, light post-editing and full post-editing. The former makes only basic or minimal changes and does not ensure a high quality of output. Therefore, most clients will require full post-editing to guarantee that their translated content is as clear and appealing as possible for their customers.

Full post-editing produces a result that cannot be distinguished from traditional translation but with the advantage of considerable savings in time and costs. The text is edited thoroughly to ensure cohesion, grammatical correctness, a suitable style and cultural / market relevance. The register adopted via the editing process is vital to ensure that the final text is fluent and sounds natural.

This is carried out using the computer-assisted translation (CAT) tools SDL Trados or MemoQ to ensure consistency and accuracy.