The free online translation is powered by Google Translate, Microsoft Translator, Babylon Translator and other machine translation systems.
Google Translate is a statistical multilingual machine-translation service provided by Google Inc. to translate written text from one language into another. Google translate supports 66 languages which makes 4290 language combinations.
Microsoft Translator implements machine translation platform to translate between 38 languages developed by Microsoft Research, as its backend translation software. Microsoft developed a translation portal as part of its Bing services to translate texts or entire web pages into different languages. When translating an entire web page, users are allowed to browse the original web page text and translation in parallel, supported by synchronized highlights, scrolling, navigation and language detection.
Babylon Translator integrates Language Weaver’s Enterprise Translation model that delivers high-speed automated translation technology. The Babylon translator supports 32 languages and offers translation of words, phrases, and texts. Users have the possibility of translating full sentences, and translate from virtually any language to any language.
On a basic level, machine translation performs simple substitution of words in one natural language for words in another, but that alone usually cannot produce a good translation of a text, because recognition of whole phrases and their closest counterparts in the target language is needed. Solving this problem with corpus and statistical techniques is a rapidly growing field that is leading to better translations, handling differences in linguistic typology, translation of idioms, and the isolation of anomalies.
There are several approaches to the machine translation technology.
Statistical machine translation
The statistical machine translation approach generates translations using statistical methods based on bilingual texts. It uses existing source and target language translations (done by human translators) to find patterns it then uses to build rules for translating between those languages. The statistical approach allows improving the accuracy of the translation with more bilingual texts utilized. Where such corpora are available, impressive results can be achieved translating texts of a similar kind, but such corpora are still very rare.
Rule-based machine translation
Machine translation systems work with natural language, a data set that is infinitely varying, ambiguous and structurally complex. To translate adequately, the rule-based machine translation system must encode the knowledge of hundreds of syntactic patterns, variations, and exceptions, as well as the relationship among these patterns. MT system must include the dictionary and specific semantic knowledge about the usage of tens of thousands of words.