Computer Assisted Translation (CAT) vs Machine Translation (MT)

As Dimetra Academy’s MT masterclass approaches, the buzz around machine translation and its facets seems to grow and there is one question that keeps coming up regularly: is there a difference between CAT and MT?

Although I am not an expert in machine translation, I have 20+ years’ experience as a professional translator using both CAT tools and MT, so here are my two cents.

Machine translation for a professional translator means a machine translation system built with selected and quality linguistic resources of a specific field; such resources are available data and/or the translator’s/clients’ translation memories.

A machine translation system is an application which is fed with linguistic data and trained using machine learning technologies to combine those linguistic data and provide text translation directly or thought CAT tools.

Therefore, the machine translation outcome depends on two specific elements: the quality of the linguistic resources that are fed to the system and the level of machine learning technology applied.

It is a synergy between a human and a machine.

Having worked with CAT tools for the past 15 years and having recently started to assess the possibilities offered by machine translation systems properly trained I see machine translation as the evolution of translation memory (TM) and therefore of CAT tools;

A TM matches the segments you have already translated to new text based on similarity percentage. A machine translation system combines the entries of your translation memory based on a statistical or neural model and provides results of equal quality and precision as that of fuzzy matches.

It takes time and effort to build a good translation memory and it takes even more time and effort to build a machine translation system and this is probably something a translator cannot do without the help of a translation technologist.

There are 3 prerequisites to build a machine translation system.

Large, clean linguistic resources + a machine translation system + a translation technologist to put the whole project together.

As the translation industry is turning into an ecosystem and technology is democratized it might just be the right moment for us translators to leverage the time and effort we have invested in building our linguistic resources.

By our SuperConnector Maria Sgourou


Leave a comment

Please enter comment.
Please enter your name.
Please enter your email address.
Please enter a valid email address.