Machine Translation (MT) has developed from a tool used by Cold War spies into a widely used tool available to everyone.
A back-packer in the outer reaches of Vietnam can now type a phrase into an online translator and ask for a beer and a meal in Vietnamese. It may not be 100% correct but with the ability to hear the phrase spoken, the boundaries of language between people are disappearing. This can only be a good thing.
In the 50 year history of MT, different schools of thought have emerged about what the best approach would be to building an MT system. While researchers fought over the best way to achieving Fully Automatic High Quality Translation (FAHQT), market requirements have gradually changed. End-users seem to be quite satisfied with useful translation. The difficulty is how to measure the quality of translation, be it machine or human. Our post-editing services deliver two levels of quality: fit for purpose or publishable.
MT is a productive tool if used the correct way, and coupled with the increase in the amount of content that may require translation – and the relentless drive to reduce costs – MT has already proven its worth. In our experience, the following considerations typically determine the extent of savings achievable thanks to MT:
- Language pairs
Some target languages and language combinations achieve higher productivity gains while others are still less amenable to the use of MT.
- Quality of source language
Better source language control and application of suitable authoring tools will normally significantly increase the potential of MT. This includes use of established grammar and style rules and terminology.
- Availability of customer-specific and domain dictionaries
Well-prepared and maintained dictionaries can significantly improve the quality of MT output, and by extension increase the cost benefits of MT use.