Jim, 
Last question.   It sounds to me like you're attempting to write one program / 
application rather than a collection of applications that run in concert and 
act upon a knowledge base? 
Am I mistaken? 
~PM

Date: Fri, 14 Sep 2012 22:10:30 -0400
Subject: Re: [agi] Simplistic Test of Reason-Based Reasoning
From: [email protected]
To: [email protected]

I agree with this.  My opinion is that deliberation is a method where you can 
bring different kinds of knowledge to bear while trying to better understand a 
situation.  Simpler reactions can be used for habit like responses which are 
made in reaction to familiar situations.  However, I do not think that quick 
human level (or animal level) responses to complex situations are that feasible 
for me.  I think deliberation will be easier for me.
 However, my idea of structured reasoning may allow the program to create 
simple algorithms (pasting bugless scriptlets together) in such a way that 
other concepts could have some kind of structured influence on the algorithm 
(that the program created) as it was running without having it mash up whatever 
reasoning it was doing.  One of my criticisms of being too reliant on funnel 
reasoning is that useful information will be turned to mash and over cooked 
when the resultants of different kinds of concepts are combined.  This is why I 
don't see differentiation as a significant problem.  I am not going to be 
mashing analyses together.  On the other hand this is why I see complexity as a 
major problem.  Reasoning is going to take place on the basis of a lot of 
little pieces of information and not the extrusion of a big hopper at the end 
of a funnel-down decision process.
Jim Bromer

On Fri, Sep 14, 2012 at 9:51 PM, Piaget Modeler <[email protected]> 
wrote:






Jim,
I'd submit to you that your system have two loops: a reaction loop to handle 
familiar recurring situations, and a deliberation loop to be invoked when 
habitual or reflex reactions become insufficient (i.e., fail to 
achieve their desired outcomes beyond some threshold of the time--e.g., failing 
30% of the time)). 

One should seriously take a look at Apple's SIRI since a system like that may 
evolve into an AGI if it is 

equipped with sufficient back end services (i.e., actions).  It has a reasoner 
which I presume respondsto user requests in a rule based or case based manner 
and is tied into various service providers.

Cheers!
~PM

                                          


-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968
Powered by Listbox: http://www.listbox.com

Reply via email to