I wasn't trying for a detailed model of creative thinking with explanatory
power - merely one dimension (and indeed a foundation) of it.
In contrast to rational, deterministically programmed computers and robots wh.
can only operate in closed spaces externally, (artificial environments) and
only think in closed spaces internally, human (real AGI) agents are designed
to operate in the open world externally, (real world environments) and to think
in open worlds internally.
IOW when you think about any creative problem, like "what am I going to do
tonight?" or "let me write a post in reply to MT" - you *don't* have a nice
neat space/frame of options lined up as per a computer program, which your
brain systematically checks through. You have an open world of associations -
associated with varying degrees of power - wh. you have to search, or since AI
has corrupted that word, perhaps we should say "quest" through in haphazard,
nonsystematic fashion. You have to *explore* your brain for ideas - and it is a
risky business, wh. (with more difficult problems) may "draw a blank".
(Nor BTW does your brain "set up a space" for solving creative problems - as
was vaguely mooted in a recent discussion with Ben. Closed spaces are strictly
for rational problems).
IMO though this contrast of narrow AI/rationality as "thinking in closed
spaces" vs AGI/creativity as "thinking in open worlds" is a very powerful one.
Re your examples, I don't think Koestler or Fauconnier are talking of "defined"
or "closed" spaces. The latter is v. vague about the nature of his spaces. I
think they're rather like the "formulae" for creativity that our folk culture
often talks about. V. loosely. They aren't used in the strict senses the terms
have in rationality - logic/maths/programming.
Note that Calvin's/Piaget's idea of consciousness as designed for "when you
don't know what to do" accords with my idea of creative thinking as effectively
starting from a "blank page" rather than than a ready space of options, and
going on to explore a world of associations for ideas.
P.S. I should have stressed that the "open world" of the brain is
**multidomain**, indeed **open-domain" by contrast with the spaces of programs
wh. are closed, uni-domain. When you search for "what am I going to do..?"
your brain can go through an endless world of domains - movies,call a friend,
watch TV, browse the net, meal, go for walk, play a sport, ask s.o. for novel
ideas, spend time with my kid ... and on and on.
The "space thinking" of rationality is superefficient but rigid and useless for
AGI. The "open world" of the human, creative mind is highly inefficient by
comparison but superflexible and the only way to do AGI.
From: rob levy
Sent: Monday, July 26, 2010 1:06 AM
To: agi
Subject: Re: [agi] The Math Behind Creativity
On Sun, Jul 25, 2010 at 5:05 PM, Mike Tintner <tint...@blueyonder.co.uk> wrote:
I think it's v. useful - although I was really extending his idea.
Correct me - but almost no matter what you guys do, (or anyone in AI does) ,
you think in terms of spaces, or frames. Spaces of options. Whether you're
doing logic, maths, or programs, spaces in one form or other are fundamental.
But you won't find anyone - or show me to the contrary - applying spaces to
creative problems (or AGI problems). T
I guess we may somehow be familiar with different and non-overlapping
literature, but it seems to me that most or at least many approaches to
modeling creativity involve a notion of spaces of some kind. I won't make a
case to back that up but I will list a few examples: Koestler's bisociation is
spacial, D. T. Campbell, the Fogels, Finke et al, and William Calvin's
evolutionary notion of creativity involve a behavioral or conceptual fitness
landscape, Gilles Fauconnier & Mark Turner's theory of conceptual blending on
mental space, etc. etc.
The idea of the website you posted is very lacking in any kind of explanatory
power in my opinion. To me any theory of creativity should be able to show how
a system is able to generate "novel and good" results. Creativity is more than
just outside what is known, created, or working. That is a description of
novelty, and with no suggestions for the why or how of generating novelty.
Creativity also requires the semantic potential to reflect on and direct the
focusing in on the stream of playful novelty to that which is desired or
considered good.
I would disagree that creativity is outside the established/known. A better
characterization would be that it resides on the complex boundary of the novel
and the established, which is what make it interesting instead just a copy, or
just total gobbledygook randomness.
agi | Archives | Modify Your Subscription
-------------------------------------------
agi
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
https://www.listbox.com/member/?member_id=8660244&id_secret=8660244-6e7fb59c
Powered by Listbox: http://www.listbox.com
