Re: Kolmogorov complexity and prediction

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Thu Oct 18 2007 - 20:09:25 MDT


Nick Hay wrote:
>
> A program of a fixed complexity can write understandable systems (i.e.
> those it can predict fit some specification) of arbitrary complexity
> (the above example works here).

Note that this requires interacting with a complexity-bearing
environment or writing more than one such system. Reason: Any
"complex" system uniquely output by a simpler system acting in a
vacuum, actually has the Kolmogorov complexity of the simpler system.

Kolmogorov complexity in this sense has very little to do with
intelligence.

-- 
Eliezer S. Yudkowsky                          http://intelligence.org/
Research Fellow, Singularity Institute for Artificial Intelligence


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:58 MDT