Intelligence, Aesthetics and Bayesianism: Game over!

From: <marc.geddes.domain.name.hidden>
Date: Tue, 29 Jul 2008 02:20:13 -0700 (PDT)

Two issues I wish to mention, here.

Firstly, I present a few rapid-fire ideas about objective morality,
culminating in an integration of aesthetics, intelligence, and
morality, all in a few brief sentences ;)

Secondly, I give a mention to computer scientist Randy Pausch, who
recently died.

As regards the first issue:

It’s been said there are clear ways to determine physical and
mathematical facts, but nothing similar for values. But, in point (2)
below I point out what appears to be an objectively existing set of
values which underlies *all* of science. I present two brief but
profound points that I what readers to consider and ponder carefully:

Point (1) there is a clear evolution to the universe. It started from
a low-entropy-density state, and is moving towards a higher-entropy
density, which, remarkably, just happens to coincide with an increase
in physical complexity with time. In the beginning the universe was in
a state with *the lowest possible* entropy. This is expressed in the
laws of thermodynamics and big bang cosmology. So it simply isn’t true
that there is no teleology (purpose) built into the universe. The laws
of thermodynamics and modern cosmology (big bang theory) clearly
express the fact that there is.

Point (2) the whole of science relies on Occam’s razor, the idea that
the universe is in some sense ‘simple’. It must be emphasized that
Occam’s razor pervades all of science – it is not simply some sort of
‘add on’. As Popper pointed out, an infinite number of theories could
explain any given set of observations; therefore any inductive
generalization requires a principle – Occam’s razor – to get any
useful results at all.

Here is the point that most haven’t quite grasped - Occam’s razor is
*a set of aesthetic principles* - the notion of ‘simplicity’ is *a set
of aesthetic principles*; Why? Because it is simply another way of
saying that some representations are more *elegant* than others, which
is the very notion of aesthetics! I repeat: the whole of science only
works because of a set of *aesthetic principles* - a *set of values*.

If all values are only subjective preferences, it would follow that
the whole of science relies on subjective preferences. But subjective
preferences have only existed as long as sentiments – therefore how
could physical laws have functioned before sentiments? The idea that
all values are subjective leads to a direct and blatant logical
contradiction.

Both these points are related and simply inexplicable without
appealing to objective terminal values. At the beginning of time the
universe was in the simplest possible state (minimal entropy density).
Why? Occam’s razor is wide-ranging and pervades the whole of science.
The simple is favored over the complex – that is– Occam’s razor is a
set of aesthetic value judgments without which not a single Bayesian
result could be obtained.

*Every single Bayesian result rests on these implicit value judgments*
to set priors. It must be repeated that *not one single scientific
result could be obtained* without these secret (implicit) value
judgments which set priors, that our defenders of the Bayesian faith
on these forums are trying to pretend are not part of science!

The secret to intelligence is aesthetics, not Bayesian math.
Initially, this statement seems preposterous, but the argument in the
next paragraph is my whole point, so it merits careful reading (the
paragraph is marked with a * to show this is the crux of this post):

*As regards the optimization of science, the leverage obtained from
setting the priors (Occam’s razor – aesthetics – art) is far greater
that that obtained from logical manipulations to update probabilities
based on additional empirical data (math). Remember, the aesthetic
principles used to set the priors (Occam’s razor) reduce a potentially
infinite set of possible theories to a manageable (finite) number,
whereas laborious mathematical probability updates based on incoming
empirical data (standard Bayesian theory) is only guaranteed to
converge on the correct theory after an infinite time, and even then
the reason for the convergence is entirely inexplicable.

The * paragraph suggests that aesthetics is the real basis of
intelligence, not Bayesian math, and further that aesthetic terminal
values are objectively real.

For those who do initially find these claims preposterous, to help
overcome your initial disbelief, I point to a superb essay from well-
respected computer hacker, Paul Graham, who explains why aesthetics
plays a far greater role in science than many have realized:

‘Taste for Makers’: http://www.paulgraham.com/taste.html

As regards the second issue, I’d like to draw readers’ attention to
computer scientist Randy Pausch. Randy Pausch was a computer scientist
who developed the famous ‘Alice’ software to teach programming in a
virtual reality setting. He was a virtual reality expert, a professor
in Human-Computer Interaction at Carnegie Mellon University. In
August, 2007 he was diagnosed with pancreatic cancer and given 3-6
months to live. He gave a famous ‘Last Lecture’ which spread virally
(via ‘YouTube’) and inspired millions (this was followed by a book
‘The Last Lecture’). He died on 25th July, 2008 .

The Randy Pausch Memorial Footbridge connects the Gates Center for
Computer Science, with an adjacent arts building, symbolizing the
bridge between art and science.

Randy Pausch Home Page: http://download.srv.cs.cmu.edu/~pausch/


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Everything List" group.
To post to this group, send email to everything-list.domain.name.hidden
To unsubscribe from this group, send email to everything-list-unsubscribe.domain.name.hidden
For more options, visit this group at http://groups.google.com/group/everything-list?hl=en
-~----------~----~----~----~------~----~------~--~---
Received on Tue Jul 29 2008 - 05:20:26 PDT

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:15 PST