Den 5.7.2011 kl. 17.09 skrev Keld Jørn Simonsen:
(...)
>>>> complicated. There are a handful of people on this list who use CG,
>>>> maybe you should talk to one of them. It might do what you want.
>>> OK, who are you thinking of?
>> Francis usually relentlessly promotes it. I'm surprised he hasn't
>> chimed in by now.
I was a bit put off by the naming "disabling new homonyms", but I had already
started to suspect that this actually meant "disambiguating" or "discarding
unwanted readings".
Yes, Constraint grammar fits the bill.
Newer versions of it actually not only discards contextually improper readings
("walk" is Inf in "I want to X", V Prs in "We X fast" and N in "We went for a
X"), it also add syntactic functions, dependency relations and even changes
such relations. It is good for contextually determined lexical disambiguation
(river vs. finance bank etc.). And the compiler is Danish, open source, and
reasonably well documented.
A good illustration of how CG can contribute to MT (apart from Apertium's own
achievements) is
http://www.mt-archive.info/MTS-2007-Bick.pdf
Trond.
------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security
threats, fraudulent activity, and more. Splunk takes this data and makes
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
