Re: Optimal Prediction

From: Wei Dai <weidai.domain.name.hidden>
Date: Thu, 4 Apr 2002 13:41:48 -0800

On Tue, Mar 26, 2002 at 02:25:21PM +0100, Juergen Schmidhuber wrote:
> In particular, if p is deterministic, then the M-predictor soon
> won't make any errors any more.

That's a bit imprecise, right? There's no guarantee that the M-predictor
*soon* won't make any errors any more. It will only make a finite number
of errors, but there's no telling how many errors will be made, or when
the last error will be made.

BTW, isn't the justification for universal prediction taken in this paper
kind of opposite to the one you took? The abstract says "The problem,
however, is that in many cases one does not even have a reasonable guess
of the true distribution. In order to overcome this problem ..." Your
papers on the other hand assume that the true distribution can be known
and proposed that it must be the Speed Prior. (Later you said you believe
it is the Speed Prior with probability 1.) Is this still your position?
Received on Thu Apr 04 2002 - 13:55:54 PST