Re: Fw: Intro to natural language processing

[Party of Citizens <citizens@vcn.bc.ca>]

On Fri, 29 Oct 2004, Rich Cooper wrote:

> But generally speaking, I think you have the right idea about representing
> scenarios within which the encoding can begin.

As I consider my own life I realize that there are scenarios which are
repeated and I don't know of a single scenario which I do not have in
common with many other members of my society. If I go to the PO and say
"Bite my shiny metal ass" like Bender the naughty robot, one of those
clinicians may well say my NL is outside the norm. Otherwise, there is a
finite number of word-letter strings which I can express and stay within
the "laws" of civil society.

For a clinician to make this judgement, that clinician must at least have
an idea in mind as to what the norms are. Moreover, I expect that these
clinicians would be able to spell out the details out if called into court
etc. to do so.

Thus we could ask millions of these people to pool their understanding of
the norms, scenario-by-scenario. As long as Bender has good object
recognition and can 'recognize' the time, place and person parameters of
each scenario, he only has to use the tables of norms as the clinicians do
and make a selection from them.

Perhaps we don't have the computing power for this presently but we will
have it some day.

Z

PS-That's <ai-nlp-subscribe@smartgroups.com> (above the line -)