Dynamic Learning

With Dynamic Learning, MT engines can continuously improve through use, raising translation quality over time.

This powerful feature works by immediately feeding a user’s post-edits in the CAT tool environment back to the MT engine, where they are immediately used to improve further translations.

machine translation feature Dynamic Learning

How can you benefit from Dynamic Learning?

Dynamic Learning means that a system "learns' or adapts through use, eliminating the need for periodic system retraining or additional data uploads by automatically improving engines with post-edits.

The feature works best for translators who frequently translate domain-specific texts, as the MT engines can correct frequent mistakes. Best of all, the MT engines can "learn" the precise usage of items such as:

  • domain-specific terminology

  • technical jargon

  • specific vocabulary

Dynamic Learning can currently be accessed in the Tilde MT plugins for MateCat and SDL Trados Studio. To see how you can leverage Dynamic Learning to improve your engines, contact us for an online demonstration: mt@tilde.com