This article describes how to use a glossary and the AWS translation engine to produce custom translations.
AWS Machine Learning Language Translations
You can create custom translations using "on-the-fly" machine learning to produce higher quality translations that require less time to review and edit before publishing. Using the AWS translation engine, Pairaphrase applies dynamic machine learning to your files as they translate. It will inject your glossary terms directly into your translations (and will use your translation memory) to apply machine learning to your translations. As a result it can take as long as 30 minutes to produce a translation, conversely it will save you time and improve translation quality by replacing key terminology with terms from your Glossary.
Studies suggest that 50% of translation mistakes are actually incorrect terminology.
To produce custom translations you'll need, at a minimum, to have a glossary that you can import into Pairaphrase. Glossaries must be in a comma-delimited csv format. They should look like this:
How to import a Glossary into Pairaphrase
- Go to your user profile and select Translation Engine in the left side-bar. Then select AWS as your translation engine.
- Click on Translation Glossaries in the left side-bar and select enable AWS from the dropdown next to the Glossary you want to use. You will see a RED dot in the Sync column. This will begin the process of syncing your Glossary to the AWS translation engine. This can take as long as 5-10 minutes depending on the size of your Glossary.
- The RED dot will change to GREEN once the syncing process has completed. Now you're ready to produce custom translations.
