From: Ben Goertzel (ben@goertzel.org)
Date: Wed Mar 31 2004 - 07:14:00 MST
Hi,
You seem to be assuming Novamente is a pure Bayesian learning system,
which is not the case, though it does involve important components
carrying out Bayesian inference.
-- Ben G
> -----Original Message-----
> From: owner-sl4@sl4.org [mailto:owner-sl4@sl4.org] On Behalf
> Of Wei Dai
> Sent: Wednesday, March 31, 2004 9:06 AM
> To: sl4@sl4.org
> Subject: Re: escape from simulation
>
>
> On Tue, Mar 30, 2004 at 05:37:01AM -0500, Ben Goertzel wrote:
> > The prior will naturally "drift" due to limitations on resources
> > leading to inaccurate maximization of expected utility in
> the context
> > of self-modification. How it will drift depends on
> phenomena in the
> > dynamics of iteratively self-modifying systems, which none of us
> > understand very well yet.
>
> I don't see what relevance resource limitations have in this
> context. It's trivial to prove that you'll never increase
> expected utility by changing your prior, so assuming that
> Novamente either starts off with this knowledge or quickly
> discovers it, and that it's smart enough to apply the theorem
> when deciding whether to change its prior, it will never
> change its prior.
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:46 MDT