How AdaptiveMT is Banishing Machine Translation Myths

Update October 9th 2019: This blog refers to adaptive/statistical machine translation, which has now been replaced by SDL’s powerful neural machine translation (NMT) solution. To learn more about NMT please click here.

Over the years a lot of myths have been spread about machine translation (MT) and its true benefit for translators. In this blog, we will demystify some of the most common machine translation myths and explain why adaptive machine translation will banish them once and for all…

Myth 1: Machine translation quality is poor

The quality of machine translation is not acceptable - I have to do too much post-editing

Before the innovation of adaptive machine translation, you would need to post-edit an MT suggestion each time it was offered to you. With AdaptiveMT in SDL Trados Studio 2017, you have your very own personal and secure MT engine that will learn from your post-edits. That’s right, AdaptiveMT is a self-learning MT engine tailored to your own unique style, content and terminology.

Myth 2: Not enough time to train an engine

I don't have the time to train a machine translation engine

As previously mentioned, with AdaptiveMT your engine will automatically learn in real time as you are translating, pretty impressive. This means you do not need to pre-train your engine and it will not disrupt your workflow, as your AdaptiveMT engine will be learning from your own post-edits. This will improve your overall MT output.

Quick tip: the more you use it, the more it learns!

Myth 3: Why use machine translation if you have a translation memory?

What is the point? I already have a translation memory (TM)

When you have no translation memory (TM) match, machine translation will provide you with a suggestion. This means you never have to translate from scratch as you will always have a translation suggestion. What’s even better is with AdaptiveMT, your own MT engine learns from any post-edits you make, meaning you never have to correct the same suggestion again.

Myth 4: Machine translation is not very secure or private

I am worried about machine translation security

There is no need to be concerned. SDL protects your data and we do not retain any copies of your translated materials or documents. We also do not share your data with any third parties. See our terms and conditions here.

If you would like to find out more about AdaptiveMT and how it is dispelling MT myths, please visit our dedicated AdaptiveMT webpage.