Translation technology has developed rapidly over the last few years, enabling language service providers to rapidly increase the speed at which they can help customers get content ready for release.
Machine Translation was once considered quite a clunky, slow and sometimes unreliable service. However, now it is an essential cornerstone of translation management systems, thanks to major new developments in machine translation technologies.
Machine translation is the use of technology to translate one language into another. It was first developed in the 1950s due to a rapidly globalising world. However, the capabilities of these first machine translators were very limited and required a large amount of post-editing by human translators. Far more computing power than was available at the time was required to develop machine translation any further.
At the start of the new millennium, when technology began to expand the capabilities of data storage and processing power, the world of machine translation began to open up. Translators and developers worked together to build complex databases of words and phrases that computers could refer to when translating from one language to the next. The machines still required a lot of manual input to be trained and post-edited to ensure a high enough quality.
In 2016, experimental teams at Google started developing new methods of translation using an approach called Deep Learning. This newer method proved to be far quicker and more effective. The translation machines could be taught to learn, which drastically improved the quality of the translations over time.
This new model of technology was called Neural Machine Translation and was quickly adopted as the standard. Other providers also started creating their own NMT models and adding them to their translation management systems.
As machine translation has evolved over the last five years, it has become an essential part of the translation process. Most language service providers now offer a service including PEMT, post-edited machine translation, where the machine translation can greatly assist and speed up the work of the human translators.
An NMT engine can also translate at extremely high speeds to the point of almost being instantaneous. This means that when working with large-volume projects, the work can be completed much more quickly. On projects where human expertise is required in post-editing, machine translation means that the workflows can be sped up immensely. Good quality translations can be passed on to language experts, who can then focus on making the content more natural and localized.
Additionally, these translations can be fed back to the translation engine via translation memories (TMs) and recalled automatically when required. Translation Memories work best when they are well-maintained as they can then allow for better quality in the translations over time. It is a cycle of continual improvement, making it quicker and more effective each round to get your content market ready.
With the evolution of NMT technology, machine translation can be and should be part of any translation workflow. It can greatly improve the speed at which any content workflow is processed.
You need to think more about the level of post-editing you require once the content has been first translated. Many technical documents can be translated by MT to an extremely high quality. For other content that may require a higher level of localization, you may want to consider a fuller post-editing and validation service where a human expert can adjust the content to match the target culture as well as the language.
Machine translation is already a key part of our present. The technology is sure to improve even further in the future, but right now, it is already a crucial part of the whole translation process.
If you would like to learn more about LanguageWire's bespoke NMT engine and how it can help speed up and automate your translation workflows, please fill in the contact form below, and someone will reach out to you soon. We have a specialised MT team in-house who will be able to answer all of your technical questions.
Do you know which industries need the most specialized translations? Find out in our top 10 list.
Build a language selection tool that makes it easy to change your website from one language to another.