Russell, Thanks for your comments... 
In an investment AI System any improvement would be to make more return on 
investments. Other systems are harder to identify like what you say "what 
an improvement is". Although doing the same task with less instructions 
would be one factor in identifing higher performance Intelligence. 

Any Seed AI system will primarilly be a learning system environment to 
begin with. 

That environoment would be the first, and the goal is machine learning, 
with a high learning curve, and to achieve trans-human function. 
Minimizing instructions to reach answers within a resonable time, and 
knowing the resources required. 

Those in the Military Industrial Complex can do almost anything, if you 
are willing to pay the price. 

I am sure the same annology will hold for the AI Systems.  

There are other environments such as medical, to research and find 
cures/perventatives for diseases. 

The ability to perceive, reason and act to acheive goals.

Knowledge extention and generation of new knowledge and its applications. 


 



----------------------------------------------------
>From : Russell Wallace <[EMAIL PROTECTED]>
To : [email protected]
Subject : Re: [agi] Flow charts? Source Code? .. Computing 
Intelligence? How too? ................. ping 
Date : Thu, 6 Jul 2006 01:43:06 +0100
> On Wed, 05 Jul 2006 15:58:28 -500, [EMAIL PROTECTED] <
> [EMAIL PROTECTED]> wrote:
> >
> > Eugen:
> > > If intelligence is ability to solve hard tasks,>
> >
> > I use the singinst.org web site def:
> >
> > www.singinst.org/seedAI/seedAI.html
> >
> > Self-understanding:  The ability to read and comprehend source code; 
the 
> > ability to understand the function of a code fragment or the purpose 
of a 
> > module; the ability to understand how subsystems sum together to yield
> > intelligence; plus the standard general-intelligence ability to 
observe 
> > and reflect on one's own thoughts.
> >
> > Self-modification:  The ability to optimize a code fragment, modify 
the 
> > function of a module, or redesign one's own architecture.
> >
> > Recursive self-enhancement:  The ability to make changes that 
genuinely 
> > increase intelligence, smartness, such that new possible improvements, 
or 
> > a new class of possible improvements, become visible to the AI.
> >
> > I do believe that might qualify as a very very hard task.
> 
> 
> That just gets you a circular definition: if intelligence is the ability 
to 
> self-improve, what counts as improvement? Change in the direction of 
greater 
> intelligence? But then what's intelligence? Etc.
> 
> Basically the problem with all this is that there's no such thing as
> intelligence in the sense of a mathematical property of an algorithm.
> Intelligence is an informal term for certain types of effectiveness of 
an 
> algorithm in carrying out tasks in an environment. So the first thing 
you 
> need to do is figure out what sort of environments you want your AI 
system 
> to work in, and what sort of tasks you want it to carry out.
> 
> -------
> To unsubscribe, change your address, or temporarily deactivate your 
subscription, 
> please go to http://v2.listbox.com/member/[EMAIL PROTECTED]

-------
To unsubscribe, change your address, or temporarily deactivate your 
subscription,
please go to http://v2.listbox.com/member/[EMAIL PROTECTED]

Reply via email to