http://209.85.229.132/search?q=cache:qbimDQYAiW0J:www.itl.nist.gov/ia...
On Arabic English trial 2009 GT has not produced an entry. It has been
placed first up to last year. The top translator was from Cambridge
UK. I am an Oxford man myself and although I support Oxford in rowing,
in Arabic I support them both equally!
In the BLEU and METEOR ratings there has been a considerable
improvement since last year.
http://www.itl.nist.gov/iad/894.01/tests/mt/2008/doc/mt08_official_re...
It should be noted that the tests are now divided into categories. The
news in Arabic is approaching near human quality. The Web is not quite
so good.The difference being that the news is slightly simpler Arabic.
LSA would in fact produce the required specialist vocabulary for a Web
page. Still it is improving.
How does this affect Google technology in general? Well some
translaations are more relevant than others.
http://docs.google.com/Doc?docid=0AQIg8QuzTONQZGZxenF2NnNfNzY4ZDRxcnJ...
The translation of this page (quite straightforward Arabic) is very
relevant to Web 3.0. If you cant get Stefan Boltzmann, Blue Giant or
the surface area of a sphere Web 3.0 is pure hype. You need a
translation of certain words to be absolutely accurate for RDF tags to
be generated.
A very loose translation would read "Web 3.0 is pure hype".
- Ian Parker
On Nov 5, 2:58 pm, jooster wrote:
> Hi,
> I am interested to learn how Google is progressing with the quality of
> their translations. Is there a measure of success, a goal that has
> been determined? Can we see how Google is progressing towards that
> goal?
|
