Horse (horse@wasted.demon.nl)
Tue, 31 Mar 1998 17:08:03 +0100
Hi Bo and LS
I thought I'd make a bit of time to reply to some of your comments in
your welcome post before the Ethics thread flares up again. Not
that I mind the Ethics thread continuing - I don't, as it is
something that has interested me for some time.
"In your introduction you hinted to a MOQ angle to Artificial Life. (is
that AI by the way?)"
Not exactly. Alife is a response to the often rigid patterns that
developed within AI. Alife covers a lot of areas of thought and
experimentation but the part that I'm interested in is to do with a
school of thought developed by Rodney Brooks at M.I.T. some years
back. It uses a system called subsumption architecture, which, in a
nutshell, is a layered system of actions wherein higher (more
complex) levels build on lower levels. Its rules are extremely
simple, (move: stop: turn: etc) but from these rules, complex
behaviour develops. It's complexity develops in response to it's
environment. It also has a great slogan:
"Fast, Cheap and Out of Control"
"Back then we had a plunge into this field. I am not sure what
conclusion we reached and cannot summarize the argumentation, but I
have a hunch that the number-crunching "Deep Blue" isn't the road
ahead. I recently read that that kind of approach is impotent when
playing the game GO."
As far as I can see this is basically correct (especially the bit
about Go - I started to learn this 'game' a few years ago but had to
drop out due to other pressures). I think the biggest mistake that AI
theorists made, quite understandably given the knowledge and paradigm
within which this field started, was to take a reductionist approach
to 'intelligence'. By making the assumption that any system is no
more than the sum of its parts, the various means of collecting and
outputting sense data were studied and an attempt made to reproduce
them, mainly in isolation from each other. This is a bit
over-simplified but a fairly reasonable description. The one thing
that they appeared to forget was that human intelligence had several
million years to evolve in the way it did with all of the various
individual physical attributes interacting with each other within
systems which involved interaction with other humans in an extremely
hostile environment.
In recent years, changes in direction have produced some more
interesting results with the introduction of connectionist systems,
evolvable programs and fuzzy logic.
Artificial Intelligence is really a 'top-down' approach to machine
intelligence, whilst Artificial Life is a 'bottom-up' approach to
life. Both seem to be learning from each other and hopefully they may
achieve some form of 'middle-out' solution.
"And yet, the human brain IS a kind of computer, so unless we
introduce "mind" as something transcending "matter", there is a
connection between the grey matter and intelligence. Give us an
input."
This is where I think that MoQ comes in. I also mentioned in my
Bio that:
"I feel that there is a strong connection between fuzzy or
multi-valued logic, Chaos and the Metaphysics of Quality."
Within SOM, which is basically dualist, Mind/matter, subject/object
etc. are the extremes whilst the bits in between seem to
be forgotten. This is where the fuzzy approach SEEMS to meet with
MoQ, or at least there appears to be some sort of conjunction.
Everything in between the extremes is of varying value w.r.t. the
extremes and is a combination of those extremes. When a person is
asked if they are male of female they can, generally, say that they
fall into one category or another. When asked if they are happy the
answer will be fuzzy. Sometimes, often, occasionally etc. They are
making a value judgement and the perception of that value is fuzzy.
This breaks away from the discrete system of A OR NOTA and becomes
A AND NOTA.
I think that this may possibly apply to the value hierarchy that
Pirsig describes in Lila and be at the root of some of the arguments
I've been reading in TLS. Are SPoV's really Fuzzy SPoV's. In the
divisions between Inorganic, Biological, Social and Intellect, where
does one start and the other finish. This may also reflect some sort
of chaotic/fractal value. As I said in my bio, I'm not sure yet.
These are just initial thoughts and need a lot more serious
contemplation. As these nascent ideas start to become more coherent I
shall throw them into the LS pot and see what sort of reaction they
get. Hopefully I shan't get too badly scorched.
A possible approach to ALife may then come out of some of these
ideas, as the degree to which something possesses life, is
autonomuos, is intelligent etc.is, IMHO, fuzzy. How does
artificial life exist in the MoQ universe. To what degree in an
artificial life-form do static and dynamic quality contribute to this
type of creature.
I can feel that I'm painting myself into a corner at the moment so I
shall rest for now and think a bit (lot) more.
Horse
"So far as the laws of mathematics refer to reality, they are not certain.
And so far as they are certain, they do not refer to reality"
Albert Einstein - Geometry and Experience
mailto:horse@wasted.demon.nl
mailto:darkstar@abduction.org
-- post message - mailto:lilasqd@hkg.com unsubscribe/queries - mailto:diana@asiantravel.com homepage - http://www.geocities.com/Athens/Forum/4670
This archive was generated by hypermail 2.0b3 on Thu May 13 1999 - 16:42:58 CEST