Yandex: the translator that gives a voice to minorities - "TechnoTron"

Yandex: the translator that gives a voice to minorities

TRANSLATOR, INTERNET, YANDEX


Yandex in Moscú


While the Google Translator offers the possibility to translate more than a hundred languages, its competitor, Russian Yandex by now it has only 94. But unlike the us giant, the company's Russian computer gives the possibility of translating minority languages, and even includes the language elvish, sindar.

The service Yandex Translator has created a unique collection of languages that have an important cultural significance. Yandex launched its translator in 2011, back then the program only 'spoke' in three languages: Russian, Ukrainian and English.

Currently, Yandex Translator has become a polyglot. Since its 'birth' Yandex has learned many languages, including papiamento, a language spoken by about 300,000 people on the islands of Curaçao, Bonaire and Aruba, located near the coast of Venezuela.

The machine learn a language in two stages. First, it analyzes the 'core' of a language that consists of words, morphological features and syntactic that appear more often. To determine this 'core' will need a small text. For example, we use texts of the Bible or of the Koran since these two sacred books were translated to almost all the existing languages.

Second, the 'nucleus' is enriched with data from related languages that the translator 'speaks'. If two languages have vocabulary that is similar, the translator Yandex starts to use data from the same model lexicon.

If these languages have rules of inflection like, also used the same morphological model that in the related language that is already part of the list of Yandex.

According to the programmers of the company, the translation system has learned eight new languages in a year and a half: the Bashkir, mari-hills, yiddish, marathi, nepali, papiamento, the Uzbek and the sindarin.

The last language is artificial, created by the writer and a british philologist J. R. R. Tolkien for the race elf in his books about Middle Earth that has no language like it in the real world.

Tolkien in fact used models of different linguistic groups real by which Yandex in particular decided to use several models of the languages of the celtic as the irish, welsh and scottish gaelic, experts explain.

In addition, representatives of the computer company Russian state that one of the services used most frequently is the translation of the Latin.

The technology behind Yandex Translator allows you to learn any language so that it is likely that the users of the service soon to see even more languages in the list.

TRANSLATOR, INTERNET, YANDEX


Comments

Popular posts from this blog

Twitter: change your profile image by default, the end of snake eggs - "TechnoTron"

WhatsApp: discover if you are spied on when using WhatsApp Web

Windows 10: with Ubuntu Bash Console how to use it - TechnoTron