Technical Translation Services – What Is Machine Translation?

Machine Translation is the use of computer software to translate one language into another. It’s been around since the 1960s, with several ups and downs in terms of research and development. Fictional stories abound of the early days of MT with phrases such as “out of sight, out of mind” being translated as “blind fool,” and “the spirit is willing but the body is weak” becoming “the vodka is strong but the meat is rotten.”

Today, MT is gaining public exposure as the Web has become an essential part of global commercial communication. MT “engines” are being incorporated in browsers and search engines. Improvements in MT’s output quality have helped in this resurgence, both for general purposes and for use by professional language service providers (LSPs).

MT systems first came about as a combination of some computational linguists’ attempts to understand human language through computer models and the US Department of Defense’s need to translate Russian documents into English during the Cold War. Since then, a number of different approaches to MT have emerged: 

• Rules-based. In this approach, just like you might have learned to diagram the grammatical structure of sentences in school, the software attempts to deconstruct the grammar of the input language to build a grammatical model of each sentence. The grammatical model of the input language is then mapped to the grammatical model of the output language.
• Statistically-based. Here, the MT engine is trained based on large volumes of existing content and its translation known as “bilingual text corpora.” The MT engine uses the large volumes of data to create statistical rules. These rules determine the appropriate selection based on the probability that given a certain word, phrase, or sentence in one language, a particular word, phrase, or sentence is the correct translation in the target language. While this approach is not language specific, large volumes of electronic text of similar content are required to get the best quality output from the MT engine.
• Example-based. Similar to the “statistically-based” MT approach, a bilingual text corpus is required. However, in the example-based approach, the corpus is used as a knowledge base to derive translations directly from examples of parallel structures of source text and translation found in the corpus.

Although there are frequent, heated debates about which MT philosophy is more effective, when it comes to MT for commercial translations, it’s probably not worth getting too hung up on which approach is used. Typically, language service providers use MT as one element of a complete quality translation process.

Read the Full FAQ
To learn more about MT and how it works, read our latest Translation FAQ, “What’s Machine Translation?”