Regularity

[Sergio Navega]

    [ The following text is in the "iso-8859-1" character set. ]
    [ Your display is set for the "US-ASCII" character set.  ]
    [ Some characters may be displayed incorrectly. ]

From: Detlef Morgenstern <detlef_morgenstern at bigfoot.de>
>(1) It is regular, if it is improbable.
>
>(2) Of all the possible sets of observable facts,
>    only a marginal number of subsets
>    can be (significantly) compressed.
>
>(3) If we can compress it, it is one of those improbable subsets.
>    Hence, it is regular.
>    It's rule born.
>
>D.M.
>

Hum... I have some comments about what you say. If it is regular,
then it is *probable*, not improbable. This means that when you're
analyzing a sequential bit string, the regular patterns will be
those which are highly probable to occur in the future, a 
definitely inductive (and probabilistic) question. 

As to your assertion (2), I would say that this depends on the
kind of data you're working with. If the data is highly regular,
then most (if not all) of the subsets of the data could be 
highly compressed:

1111111111111111111111111111.....

a) [1] repeated
b) [11] repeated
c) [111] repeated
...
n) [a)] repeated
m) [b)] repeated
...

It seems that your assertion 3) would have do say: if it is
compressable, then it is one of those *regular* subsets, which
occur often.

Just a minor comment on a related issue: although I firmly believe
that redundancy in time (and regularity) is one of the hallmarks
of the intelligent process, this cannot be taken to be the only
meaningful factor. There's a potential question regarding the
kind of learning being practiced here: if one adopts a purely
unsupervised scheme, then I would say that this would not warrant
intelligence to emerge.

Without some kind of interaction with the environment, I don't 
think intelligence can emerge in an organism. That means that
the "loop" of an intelligent mechanism must contemplate a 
purely inductive, unsupervised process but this should not
be all of it; it is necessary to use it only as a helper 
function in a larger architecture that directs the organism 
to interact with the environment and explore different effects 
of actions, perhaps to disambiguate potential conflicts.

Regards,
Sergio Navega.
http://www.intelliwise.com/snavega