What is machine translation?

Machine translation (MT) is the conversion of text from one language to another by a computer, with no human involvement.

Pioneered in the 1950s, MT is one of the earliest examples of artificial intelligence. Today’s MT is continuing to innovate by leveraging neural network models to deliver unprecedented linguistic quality. However, not all MT is created equal: its quality is determined by how it is trained.

More about machine translation

How does machine translation work?

There are three different types of machine translation: rules-based, statistical and neural. Neural machine translation (NMT) is the current approach that uses machine learning and neural network models.

What are the benefits of machine translation?

Machine translation is a valuable resource in the translation process. It can be used on its own or can be combined with post-editing to increase translator productivity, allowing translators to:
  • Translate more, faster: With MT, you can translate millions of words instantly, making it great for high-volume or fast-turnaround projects.
  • Access a wide breadth of languages: MT is available in hundreds of different language combinations, making it quick and easy to obtain translations in whatever language you require.
  • Reduce translation costs: With neural engines, machine translation output is more fluent and significantly higher quality, reducing the amount of post-editing work required. With less human involvement, costs can be significantly lowered.
However, it’s important to remember that not all content lends itself to machine translation. MT tends to work better with well-structured, unambiguous content, such as technical documentation; for internal communications; and for translating huge amounts of user-generated content where you just need to understand the gist. MT is less useful for creative content, like marketing or literary projects, as it often struggles to capture the nuances of local culture.
Where MT really shines is as part of a comprehensive translation workflow, supervised by human linguists.

What capabilities do modern MT offerings support?

  • Instant on-demand translation
  • Flexible integration and deployment
  • Training and adaptation