Re: universal pattern detector

From: Perry E.Metzger (perry@piermont.com)
Date: Fri Nov 07 2003 - 08:17:00 MST


James Rogers <jamesr@best.com> writes:
> This depends on the assumptions to a certain extent. If we are talking
> about general non-axiomatic induction, a function solution (not THE
> solution, just a GOOD approximation) can always be found for these types of
> cases, and with a known error bound.

I don't think that you can always have an absolute error bound --
indeed, you can only get an average error bound for a subset of
problems. I can think of plenty of interesting problems for which
there is not even an approximate solution that will give you an
average bound on the error.

BTW, this is reflected in the way real humans end up working. We get
into spectacular failure modes of various kinds in our perceptual
systems.

Perry



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:43 MDT