[MUD-Dev] A footnote to Procedural Storytelling

[Joe Andrieu]

> -----Original Message-----
> From: mud-dev-admin at kanga.nu
> [mailto:mud-dev-admin at kanga.nu]On Behalf Of
> Travis Nixon
> Sent: Wednesday, May 03, 2000 2:58 PM
> To: mud-dev at kanga.nu
> Subject: Re: [MUD-Dev] A footnote to Procedural Storytelling
>
>
> Brandon Rickman wrote:
>
> >Highlights of the symposium included:
> >
> >- Selmet Bringsjord's proof that "interesting stories are
> not computable,
> >therefore cannot be generated."  (I've yet to buy the
> book, but it might
> >be of academic interest to some of you: _Artificial
> Intelligence and
> >Literary Creativity : Inside the Mind of Brutus, a
> Storytelling Machine_)
>
>
> I might have to have a look at this book, because I firmly
> believe that
> statement to be false.  It may very well be that
> interesting stories are not
> computable by todays technology, but frankly, I've gone
> through some (very )
> rough and preliminal on-paper tests of some ideas for doing
> just that, and
> it certainly seems possible.  Anybody have any web-based
> reading on ideas of
> this nature?  (mostly because I'm too lazy to go to a bookstore)


The great thing about Selmer is that he is his own best Devil's
Advocate. Pretty much for his entire academic career he has
simultaneously challenged the idea that computers can think, while
doing his best to actually create programs that can tackle the hardest
AI problems, including creativity.

The book is a worthwhile read, a real mental workout in formal logic
and thinking about how computers can pass Selmer's modified version of
the Turing Test, which he calls the super-short-story (SSS) test.
Start with any first sentence, given to a number of human writers and
computer writers. Each writes a story that's a page or two long (500
words, I believe). Compare. Pick out the computer generated ones.  If
you can't, the computers pass the test.

The cool thing is that the first version of his software, Brutus.1,
has actually written stories that have passed a web-based version of a
limited SSS test.  However, as Brandon pointed out, most work in this
area is focusing on either one-player experiences with a small stage
and a few actors or actual written story generation.  Transforming
this to MUDs is, shall we say, a challenge.

For more info on Selmer's thoughts on computer creativity, check out
"Chess is too easy:"
http://www.techreview.com/articles/ma98/bringsjord.html


-j

--
Joe Andrieu
Realtime Drama

joe at andrieu.net
+1 (925) 973-0765






_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev