Neural Translation powered by Translation Memories & Fuzzy Matching

The use of Fuzzy Matching to improve Machine Translation quality

Fuzzy matching browses through translation memories for a certain sentence, to find approaching matches, a technique used by human translators to reduce the work of post-edition. A recent development in our technology called Neural Fuzzy Adaptation, is now enabling the Neural Machine translation model to consider fuzzy matches in the same way that a human would. This allows for faster and more accurate translations.

Exploit your TMs to boost NMT performance - SYSTRAN
SVG

A feature directly available in your TMS

Launched in partnership with XTM, this new feature named AI enhanced TM is now available to all users in the latest release of XTM Cloud v12.8. The innovative technology provided by SYSTRAN is using TM & Fuzzy Matching to increase the quality of Machine Translation output in real time. This is the very first time such feature is offered in a TMS.

Learn more about XTM Cloud v12.8

A significant improvement in MT quality

This new feature brings instant benefits to users with the customization of Machine Translation output in real time, while offering consistency with previous translations. The impact in terms of quality and accuracy is comparable to machine translation engine specialization, a service that requires time, significant data volume and technical expertise. Engine specialization combined with  Neural Fuzzy Adaptation remains a must to achieve the highest levels of translation quality.

Improvement in MT quality
Supported languages

Supported languages

The feature is today supported for the language pairs listed below. More pairs will be available in the Fall.

English <> Arabic, Chinese, Dutch, French, German, Italian, Japanese, Polish, Portuguese (Brazil), Russian, Spanish, Swedish, Turkish

French <> German, Spanish, Portuguese (Brazil)

Translation Memories and Post-edition become the fuel for incremental Neural Machine Translation engine training

Translation Memories and Post-edition for incremental Neural Machine Translation engine training

Some fields are required

We respect your privacy, and will use this information only for contact purposes.