Re: Fw: Intro to natural language processing
Rob Freeman wrote:
<snip/>
> I've seen enough. Professional Functionalists spend most of their time
> trying
> to squeeze a subjective reality into an objective description, just like
> the
> Generativists. The only difference is they do it for categories of
> meaning.
Yes, that's what bothers me about most of the NLP work going on today and
in the past. All our perspectives on NL have been through a disembodied,
ungrounded view of the world. All good practitioners of language (i.e.,
people)
are emodied, self-interested, and grounded in a very complex reality.
Luc Steels is the first writer I've seen where the agent is at least
slightly grounded
in a perceptive world. I still don't like the fact that the agents don't
have a
teacher, because all the examples (people) I know did in fact have teachers
from birth. His work may shed theoretical light on the origins of language,
before there were any teachers, but I would like to see some similar work
done with a teacher (using WordNet?) and with well simulated grounding in an
environment (Wierzbicka's NSM?) to support what goes on in children's heads
in
the first few years of life.
The example John showed of the 34 month old girl using modal logic, self
description, and personalization within a single sentence indicates that we
humans normally have that capability growing in the first three years of
life.
<snip/>
> Oh, and if you like Luc Steels' stuff you might like what this team is
> doing
> up in Finland (e.g.
> http://www.cis.hut.fi/~tho/publications/honkelawinter2003tr.pdf, and
> they're
> having a conference: http://www.cis.hut.fi/AKRR05/.)
Thanks, I took a look at the paper, and it seems to have some good ideas,
though
still not of the teacher-assisted variety. I've added it to my to-complete
list.
>I think consensus is a
> good model for the emergence of lexicon. But the the consensual concept of
> meaning does not explain the creation of new meaning by syntax (From
> memory
> the best LS has is genetic algorithms selecting new syntax rules by
> consensus, not quite the same thing.)
But somehow, the consensus should include the teaching sources as well as
the learners. Maybe the teachers should be agents that already know a
competent
lexicon but are unable to learn, while the consensus forms around them.
JMHO,
Rich