April 25, 2012
by Larry Ferlazzo
The quality of online translation systems is obviously dicey, but they can be very useful to teachers, students and parents. As I’ve previously posted, Google Translate (the most popular tool) has been found best for longer pieces, while Microsoft and Yahoo appear to be best for text with less than 140 characters.
Of course, none of these tools are useful for the huge number of languages that are not supported by them.
Hmong, the primary language spoken by many students and families here in Sacramento and elsewhere, has been in that list of non-supported languages.
Until this month.
Now, with a push from the Hmong community in Fresno, California, Microsoft Translator (also known as Bing Translator), has just begun supporting the Hmong language.
Here’s how Microsoft describes the process that was used to add Hmong to the system, and the ability to use a similar system to add others:
“All these years, the language has been preserved, despite efforts to eradicate it,” said Will Lewis, a Microsoft program manager who worked on the Hmong translator. “Now, the irony is that in the United States, a country where they’re free to speak it, the thing that never happened in Hmong history is happening; some children are not learning Hmong.”
The translator uses a statistical model to find patterns and assign probability to words in context. Since November, when the project started, community members and researchers have fed a computer with hundreds of documents in the two languages, as well as with entries from an online Hmong-English dictionary. Since dictionaries offer no context, community members entered sentences for each word.
A similar concept could be applied to benefit other minority and indigenous languages, most of which aren’t covered by automatic translators such as Microsoft’s Bing Translator or Google Translate, Lewis said. Less than 100 languages — mostly the dominant, widely-used ones — are currently covered, out of the more than 7,000 languages world-wide.
I’m adding this info to:
The Best Reference Websites For English Language Learners