Re: Decision theory

From: Wei Dai <weidai.domain.name.hidden>
Date: Fri, 1 Jan 1999 16:06:59 -0800

On Fri, Jan 01, 1999 at 05:41:11PM -0500, Jacques M Mallah wrote:
> I don't see the problem. You were right when you said the problem
> is a practical one: if it is impossible to calculate the effects of a
> decision because that decision is physically impossible according to a
> theory with no free parameters, that's a problem. I don't see that
> problem ever occuring. In order to have no free parameters, the price
> that a theory has to pay is precisely that all possible courses of action
> will occur.

I agre that is not the problem. The problem is a physical theory making
the same (global) predictions for every possible course of action. This is
a problem because in classical decision theory, utility functions are
defined over the global states of the universe.

> Or to put it another way, even if the universe as a whole has no
> free parameters, one is still free to consider a subsystem and has free
> parameters available to specify what subsystem one chooses to look at.

This might work for a universe that CAN be decomposed into subsystems, or
in other words if the physical theory you're considering has some notion
of locality. This is not true for the theory I proposed based on the
universal prior. And even if it is true, I still don't see how to
formalize the idea. What are utility functions to be defined over, for
example? If we're still supposed to maximize expected utility, how is that
defined?
Received on Fri Jan 01 1999 - 16:08:43 PST