Machine Translation Postediting

Machine Translation, abbreviated MT, is one of those buzzwords that has been round the language industry since the 90s though it is only recently that the technology has reached a maturity level which makes it worthwhile.


Broadly speaking, there have been two main approaches to MT: linguistic-rules based and statistical based. Modern MT engines are mostly statistical based although they might have some syntactic rules and morphological process embedded. The outcome of modern MT engines is far from perfect but postediting this outcome can improve translator productivity, ie. postediting can be faster than translating a sentence from scratch.


Most modern CAT tools allow the integration of MT engines. When the CAT tool does not find a translated match in the existing Translation Memory, it uses the MT engine to generate the translation automatically. The MT translated sentence is then edited by the human translator and saved in the Translation Memory.


For certain types of texts, industries and languages, the output of Machine Translation can help the professional translator to work faster and be more productive, and clients can finally benefit from lower translation rates.



Some public MT engines are available in the cloud, ie. Google Translate or Microsoft Translator. Obviously, when translating confidential or sensitive data, sending the information to these MT engines in the cloud is not in fact safe and might put confidential information at risk.


To overcome security issues, ASSERTIO LS uses a private MT engine so you can rest assured that your confidential and sensitive information will always be stored securely and confidentially.