Amazon has released a new version of its translation machine. The Translate software has a new feature that allows users to change the translation style based on previous translations.
The new feature is called Active Custom Translation (ACT). Users can add parallel data (PD), containing examples of previous translations. This data can be in TMX, TSV or CSV formats.
The data entered are then incorporated into the machine learning model on which Amazon Translate is based. As a result, the final translations are in the same style as the examples given.
In a blog post, Amazon mentions the Spanish translation of “How are you?” as an example. Depending on the input from your PD, the friendly “¿Cómo estás?” or the formal “¿Cómo está usted?” is used.
Less manual work
With the new feature, users no longer need to manually create new machine learning models to track style changes, Amazon claims. Active Custom Translation is available now.