METM24 presentation

Using neural machine translation software in CAT tools: installation, cost, performance

Jennifer Gray, Malaga, Spain

Have you ever wondered whether neural machine translation (NMT) software can really boost your translation output without sacrificing quality? In this presentation, you’ll get the chance to find out when we link four of the most popular NMT tools – Google Translate, DeepL, ModernMT and Microsoft Translator – to Trados and MemoQ. We’ll go through the basic steps involved in configuring these CAT tools to automatically receive NMT suggestions for each text segment. This can speed up your workflow and give you time to focus on nuance, disambiguation and style. Is NMT software free? Yes and no. We’ll price match, and also explore third-party solutions that connect Trados and MemoQ to dozens of different NMT tools. How accurate is NMT? We’ll see the type of texts and language pairs in which NMT is most likely to perform well, so you can decide whether it’s right for you. Finally, we’ll get down to specifics, and see how Google Translate, DeepL, ModernMT and Microsoft Translator dealt with a series of short texts in the fields of law, literature, medicine and engineering contributed by MET members working in English, Spanish, French, German and Finnish.

About the presenter

Jennifer Gray is a freelance Spanish to English translator and editor specialising in biomedical texts.