Fundamental Compressionist Philosophy.

[Detlef Morgenstern]

Dear Andrew,

--- Andrew Stanworth <andrew.stanworth at bigfoot.com> wrote
in http://www.sanna.com/casc/archive/frm00161.html:
> Hope you like my first offering. I have sackloads more of this
> sort of stuff to say if anyone is interested. Some of my earlier
> ideas, when I was starting off in this direction a few years ago,
> can be found on my site www.seaca.org

I want to reply to what you write on your site - which I find
gripping. (We may return to the switch discussion another time.)

Let's assume, our minds are emulated in a UCM. Some interesting
puzzles must be solved by a programmer who wants to code such a
brain.

[You -AS- write:]
"In effect, we seem to be acting as massive filter systems, able to
pick out the essential features and relationships in the data that is
presented to us in the form of our sensory experience."

[Me -DM-:]
There must be some global filter criterion for 'essence'. How would
you describe, how such a filter can detect essence which it had never
seen before, i.e. for which it cannot have built in patterns to
compare to - with the aim of finding a matching one?

[AS:]
"All of our quests for truth and knowledge come down to nothing more
than trying to reduce the experiences we have to the simplest
underlying patterns that we can recognise."

[DM:]
We have a very efficient sense for regularity. How would you describe
the essence of regularity?

[AS:]
"It is a common mistake to think that science produces explanations
of how or why things work."

[DM:]
Science distilles regularity - the very essence we seem to be
interested in - from experience. With a firm belief in causality, we
do so to be able to forecast effects when observing causes. If we
want to abstractly model 'science as such' in a UCM, we must define
how regularity - the essence for which we still have to give a
sufficient description - can be distilled from experience. That's
all, what remains to do... 

[AS:]
"The key point to remember is that all UCMs are fundamentally
identical in terms of what they can actually do - the only difference
between the different architectures is that some may be easier to
configure for one task rather than another."

[DM:]
This means, we must describe an architecture for a UCM which can be
easily configured for the task just mentioned.

[AS:]
"My universal computing machine is now not so much a machine, rather
it is the heart of a universal descriptive framework (which although
being the essence of Turing's machine, now makes it look hideously
complex)."

[DM:]
Is it functioning in the above sense?

[AS:]
> I have sackloads more of this sort of stuff to say if anyone is
> interested.

I am!!

Regards,

Detlef Morgenstern
Dresden
mailto:detlef_morgenstern at bigfoot.de


__________________________________________________________________
Do You Yahoo!?
Gesendet von Yahoo! Mail - http://mail.yahoo.de