My group is moving towards hybrid approaches, is the thing -- Mike Gasser is pretty excited about this dependency grammar approach (XDG), and we want to be able to learn rules for XDG grammars that map between languages.
But moreover, we're trying to handle languages where there aren't large training corpora available, so a purely statistical approach won't get us so far.
2 comments:
> Second International Workshop on
> Free/Open-Source Rule-Based
> Machine Translation. And it's
> in Barcelona! What could
> possibly be better?
uh... a workshop on free/open-source STATISTICAL machine translation?! (joke!!!)
I'd be pretty excited about that one too :)
My group is moving towards hybrid approaches, is the thing -- Mike Gasser is pretty excited about this dependency grammar approach (XDG), and we want to be able to learn rules for XDG grammars that map between languages.
But moreover, we're trying to handle languages where there aren't large training corpora available, so a purely statistical approach won't get us so far.
Post a Comment