Monday, July 15, 2002 |
|
Feature |
|
PCs can’t
translate!
TAKE
this sentence, let translation tools on an Internet search engine work
their magic to translate it into Korean and back again, and this is what
you get: "It has this elder brother and boil, if magic they in
order to translate it again at a Korean and after one, it makes the
translation tool in the Internet search engine, this is what you get.
Clearly, computers
still can't translate as accurately and artfully as people do.
Many experts doubt they
ever will. But recently some researchers have stumbled on what could be
a powerful new tool for translators: the World Wide Web.
"People say, look,
possibly everything I've ever wanted to say is on the Web and probably
already translated on the Web," Eduard Hovy, a researcher at the
University of Southern California's Information Sciences Institute,
said. The Web is flooded with translations of everything from novels to
corporate documents to personal pages. Some have been translated by
people, some by translation software, some by a combination. The
software of one company, Systran (a simple version of which is found at
Altavista.com and produced the above example), translates 6 million
pages a day. Other players in the field include IBM Corp., SDL
International and Bowne Global Solutions, which purchased intellectual
property from bankrupt language software company Lernout and Hauspie.
Pooling the lessons suggested in Web pages may someday prove more
effective at creating new translations than the current method.
For now, programmers
generally assemble dictionaries of words and phrases likely to occur in
the documents to be translated, along with rules to help figure out an
unfamiliar phrase.
It works well enough
for texts with recurring vocabularies and style - weather forecasts or
owner's manuals. But it's not a tool to be used on marketing literature
or contracts. Dimensions, dates, local currencies, laws and proper nouns
- an executive named "Mike" came out, as
"microphone" on one Website - are too complex. In short,
computers have a little common sense. Even a child can tell from the
context of a sentence whether the "bank" is a place to borrow
money or to fish, but that still largely baffles machines. To deduce
those rules, computer needs millions of examples, laid out in perfectly
aligned, translated text.
So
far, researchers have found such examples amid the proceedings of the
Canadian Parliament, which are issued in French and English. And the
University of Pennsylvania's Linguistic Data Consortium has formatted a
Hong Kong legal archive and back issues of a Taiwanese magazine in
Chinese and English. But those documents offer little help translating
"Randy Johnson is a monster with a wicked fastball," for
instance, or the lyrics to a '70s rock anthem. But those phrases - or
enough parts to be useful - may well have been translated somewhere on
the Web. The translations may be bad, but they're good enough to get the
idea across. — AP
|