Re: Let me try once more (now AI discussion)
Kat,
I wasn't really talking about the AI writing it's own random code, but
rather choosing randomly from any/all existing code modules for actions,
particularly in AI game, which is what Chris was talking about in this
regard. However, I don't know enough about programming to know if that's
possible either, just assumed (!) that it is.
Dan
----- Original Message -----
From: "Kat"
Sent: Monday, January 29, 2001 7:14 PM
Subject: Re: Let me try once more
> On 29 Jan 2001, at 18:20, Dan B Moyer wrote:
>
> > Actually, I suspect that AI *can* be written to learn beyond what the
> > programmer has thought up to allow as choices: just allow *random*
actions,
> > plus analysis of useful effect; if useful, keep it as an option; if
harmful,
> > don't do it any more (or assign "low usefulness index").
>
> Well, of course, it's been done, decades ago even, but how would you allow
for that in
> a program written for Eu, without writing your own parser and interpreter
to exec the
> random code? Write the new code, then shut down and totally restart the
program?
> Ewww. Only even partway decent way to do that would be to have two
identical
> programs running, so the old one can keep running and send var contents
and
> machine state to the newly re-started program with the new code,, before
the new
> code was executed. Not terribly elegant.
>
> Kat
>
> ____________________________________________________________
> T O P I C A -- Learn More. Surf Less.
> Newsletters, Tips and Discussions on Topics You Choose.
> http://www.topica.com/partner/tag01
>
____________________________________________________________
T O P I C A -- Learn More. Surf Less.
Newsletters, Tips and Discussions on Topics You Choose.
http://www.topica.com/partner/tag01
|
Not Categorized, Please Help
|
|
