Vladimir Nesov wrote:
On 10/5/07, Richard Loosemore <[EMAIL PROTECTED]> wrote:
Vladimir Nesov wrote:
On 10/5/07, Richard Loosemore <[EMAIL PROTECTED]> wrote:
Mike Dougherty wrote:
On 10/4/07, Richard Loosemore <[EMAIL PROTECTED]> wrote:
All understood. Remember, though, that the original reason for talking
about GoL was the question: Can there ever be a scientific theory that
predicts all the "interesting creatures" given only the rules?
The question of getting something to recognize the existence of the
patterns is a good testbed, for sure.
Given finite rules about a finite world with an en effectively
unlimited resource, it seems that every "interesting creature" exists
as the subset of all permutations minus the noise that isn't
interesting. The problem is in a provable definition of interesting
(which was earlier defined for example as 'cyclic') Also, who is
willing to invest unlimited resource to exhaustively search a "toy"
domain? Even if there were parallels that might lead to formalisms
applicable in a larger context, we would probably divert those
resources to other tasks. I'm not sure this is a bad idea. Perhaps
our human attention span is a defense measure against wasting life's
resources on searches that promise fitness without delivering useful
results.
I hear you, but let me quickly summarize the reason why I introduced GoL
as an example.
I wanted to use GoL as a nice-and-simple example of a system whose
overall behavior (in this case, the existence of certain patterns that
are "stable" or "interesting") seems impossible to predict from a
knowledge of the rules.
You do predict that behavior by simulating the model. What you
supposedly can't do is to find initial conditions that will lead to
required global behavior. But you actually can - for example by
enumerating possible initial conditions in a brute force way and
looking at what happens when you simulate it. It's just very
inefficient, and as a result you can't enumerate many initial
conditions which will lead to interesting global behavior. And
probably there are tricks to get better results, by restricting search
space. You propose a framework which will help in efficient
enumeration of low-level rules and estimation of high-level behavior,
and restrain possibilities to as close as possible to existing working
system - human mind. All along these same lines. Computational
mathematics deals with this kind of thing all the time.
Vladimir,
You keep taking this example out of context! You are making statements
that are completely oblivious to the actual purpose that the GoL example
serves in the paper:
Given that this purpose is what I'm trying to understand, being
non-oblivious to it at the same time would be strange indeed.
Okay, maybe I have been misunderstanding the level of apparent
dismissiveness in your tone.
I will try to reorient myself and address the issue as carefully as
possible.
everything you say above is COMPLETELY impractical
if it is generalized to systems more complex than GoL.
I disagree. It's not specific enough to be of practical use in itself,
but it's general enough to be a correct statement about practically
useful methods. Please don't misunderstand my intention: I find your
way of presenting technical content rather obscure, so I'm trying to
construct descriptions that apply to what you're describing, starting
from simple ones and if necessary adding details. So if they are
overly general, it's OK, but if they are wrong, please point out why.
I started to reply to this directly, but now I think it is better to
refer you to the answer I have given in a parallel post to Mike
Dougherty, which addresses the core issue.
Richard Loosemore.
-----
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=8660244&id_secret=50733913-b27ff0