+44 20 80501535

What is machine translation?

Machine translation (MT) is the automatic translation of a text into another language using computer software. Human interaction is dispensed with. Various translation methods are used in machine translation.

Need a translation?

Development of machine translation

The first technologies in the field of machine translation were developed in the USA in the 1960s. The military in particular hoped to gain a war advantage by translating foreign-language documents. However, research was put on hold and only resumed in the 1980s by German and Japanese scientists. Since 2006, machine translation has been available to the masses thanks to Google Translate. Artificial intelligence gave automated translation a huge boost in quality in 2016, but the technology is not yet perfect.

How machine translation works

The aim of a professional translation, whether done by man or machine, is to translate a text in such a way that the meaning corresponds to the original. Translators must therefore interpret and evaluate all text elements in order to produce a correct translation. There are many translation programs online that convert texts into another language within a matter of seconds. They use different methods for language translation, which we would like to introduce in the following paragraphs.

Rule-based machine translation

Rule-based machine translation (MT), used mainly in the 1980s, provides a literal translation. This is supplemented by language algorithms that combine the linguistic and grammatical rules of both languages. This rule-based translation system usually delivers consistent and logical translations, but is quite prone to errors.

Load more

FAQ: More questions about machine translation

What is neural machine translation?

Neural machine translation is a technology based on networks of artificial neurons. The system detects correlations between the source and target language and can thus predict the probability of a certain word sequence.

How does translation software work?

There are various providers of translation programs that work with different methods. Some software breaks down each sentence word by word and looks them up in the dictionary. The most widespread method today, however, is neural machine translation, which can use artificial intelligence to make statements about the probability of a phrase being used. This system is used by leading translation services such as Google Translate or Bing. The provider DeepL has further developed the technology for itself.

Can machine translation replace human translation?

Machine translation systems are a good way to have many words translated cheaply in a short time. Nevertheless, you should realise that this method cannot compete with human translation in terms of quality.

Which texts are suitable for machine translation with subsequent post-editing?

Where automatic translations do not yet meet all requirements in terms of quality, post editing, i.e. subsequent revision of the text by a human, can close the gap. Texts with simple vocabulary, short sentences and few convolutions are particularly suitable for this. Post-editing delivers high-quality results and saves time and costs.

This site is registered on wpml.org as a development site.