Quoth Tolkin, Steve:
> statistical natural language processing (This is not a mainstream part
> of linguistics for many years, because it does not pay any attention
> to even grammar.  Rather it is IR (information retrieval) informed by
> AI, statistic, data mining, etc.

I have to disagree.  NLP (which is pretty much all statistical at this
point in time) is quite a bit more than what you say.  Also known as
"computational linguistics", it covers nearly anything you can do with
natural languages and computers.  In particular, tagging and parsing are
common tasks (and you better believe parsing pays attention to grammar! :).
Other subfields include machine translation, voice recognition, dialogue
processing, anaphor resolution, prosody analysis... the list goes on.
Language modelling is one important thing that NLP/CL does, and that's a
very linguistic thing: as our (computer) language models improve, we
gain insight into the models floating around in our brains.

Anyway, that's mostly tangential.  Apropos the real topic of discussion,
I'd just like to point out that despite all the stupid bs and hype in
the popular media, most AI researchers still call themselves AI
researchers.  If they adopted some other name, the media would just get
hold of that and do the same thing.  Meanwhile, AI *does* have the
advantage of being very recognisable, and in particular, *the thing
people would think of* if that's what they were looking for.  It's even
recognisable in acronym form.

-- 
-=-Don [EMAIL PROTECTED]=-=-<http://www.cs.brown.edu/~dpb/>-=-
"In AI, much of the I is in the beholder."      --Manning & Schütze

Reply via email to