Lucy Xu

What is Machine Translation?

As recent stages of innovation bring in more and more advanced localization technologies and techniques, let us take a look at specific type method of translation that is gaining popularity amongst global companies of all industries and sizes: Machine Translation (or MT, for short).

For those looking to learn more about MT for the first time, or localization professionals crushing up on their new technologies, this post breaks down what Machine Translation is, how it’s used, and how it is making localization easier for companies around the world.

How does Machine Technology work?

Machine Translation (or MT, for short) is a type of technology that automates translation. Depending on the software or platform, MT may happen with or without humans in the translation loop. At a glance, here’s how it works: the MT software will take the input content and automatically translate it based on existing translations – sourcing translations from existing platforms like Google Translate and the like.

Then, the software gets ‘smarter’ as it is ‘trained’ by your localization team through additions like style guides or even continuous real-time input from humans in the translation loop. Depending on the accuracy level and customization required, the type of MT and the ratio of machine-driven versus human-edited translations may differ.

When is it used?

Machine Translation is a popular option for companies that are translating a lot of content at scale, because it is more manageable and far more efficient than any other method for translating these hundreds of thousands to millions of words into target languages.

On a more granular level, companies experienced in localization will also choose to mix and match various method for different types of content. For example, the website content may need extra care and attention from translators to make sure that the front image of the company in its new market is both on brand and also culturally conscious. On the other hand, help docs and other pages on the back end may be simpler and can therefore be mass translated with some lighter human-in-the-loop touchpoints for quality control.

How does it compare to other localization methods?

MT is still an evolving technology that is being refined day by day. Because of this, the quality of the translation output from MT stills falls on a spectrum depending on which tools you utilize, how well you train the software, and how much human-in-the-loop you integrate into your overall localization workflow.

However, the future is bright with more and more tech giants and innovative localization technologies continuing to invest time and research into elevating the accuracy and efficiency levels of Machine Translation. For example, Amazon Translate recently teamed up with Transifex to bring machine translation to enterprise content management — moving the world another step forward towards making localization more accessible for all.

 


Ready to give Machine Translation a try or test out localization for your business? Sign up for a 15-day trial of Transifex start using localization to go global.

Want to learn more about Transifex?

Give Transifex a try with our free 15 day trial, or connect with one of our team members for a personal demo.