On Wednesday 06 June 2007 11:11:53 am James Ratcliff wrote:
> The main thin is the restriction on domain, all of [Schank's] scripts were 
very limiting, IE if you used a restaurant script and anything out of the 
ordinary happened, it would break.  It was very fragile in this fashion, and 
such, most all scripts were hand-created and of limited use outside the test 
cases.
> 
>   One way I looked at fixing those deficiencies is having many similar 
scripts allowed, and allowing an easy way for an english user to create a 
basic script.

In real life, with real people, of course, this is one of the most basic 
capabilities (on both ends) -- it's called "telling a story". Real 
intelligence generates a stock of scripts from experience, both firsthand and 
secondhand. The key is to be able to parse your experience as a script in 
such a way you can use it later. And that gets us back to representation and 
concept formation: unsolved as yet (and that's why I said that the original 
work had a gap separating it from the implementable). But in its own terms, 
at its ontological level, I think it was generally on track.

Josh

-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?member_id=231415&user_secret=e9e40a7e

Reply via email to