PM: "Concepts are just constants"

What do you mean? It's hard to think of anything constant in any concept.


From: Piaget Modeler 
Sent: Tuesday, July 10, 2012 11:48 AM
To: AGI 
Subject: RE: [agi] Variables vs concepts ; analog vs mapping computers



At the end of the day, Concepts are just constants. 


And I agree that we need to do analogy (i.e., constant mapping and 
substitution) primarily.


Cheers!


~PM



--------------------------------------------------------------------------------
From: [email protected]
To: [email protected]
Subject: [agi] Variables vs concepts ; analog vs mapping computers
Date: Tue, 10 Jul 2012 11:26:25 +0100



John:  The symbology is digital, unless somehow you make the symbols analog for 
some benefit. 

There are two kinds of symbols - the symbols of logic and maths and computer 
algorithms, which are only *variables* with precise, specific referents. If p 
then q - a + b = c, a geometric square or triangle - these are all variables.

But the symbols of language are *concepts*. Concepts are the "building blobs" 
of AGI thought - the magic ingredients which enable humans and animals to 
create flexible  courses of action in the real world. Concepts like 

"*I* *must* *go* *get* *some* *food* "

Concepts like these - all the concepts of language -  have vague and general 
rather than precise & specific referents.  There is no specific "I" or form of 
"I" - no specific form of "go" or "food" or "some." But horrific as that is to 
the logicomathematical mind, it is great for real world action.

It is concepts like these that all AI, in its present incarnation, is 
completely incapable of handling. And it is concepts, not bleeding 
algebraic/logical variables, Steve, that AGI must aim for, and that represent 
one of the obvious great challenges.

Such concepts enable flexible, creative courses of action in the real world. 
They are demonstrably beyond the grasp of any algorithm or logic or maths. 
There is no algorithm for "go" or "food."  Nor are there any equivalents in 
logic or maths.  It is because an AGI starts with just a vague, general 
direction like "go" that it can engage in more or less any form of movement 
that may be required on its real world journey to the fridge - can hop, jump, 
leap, crawl or walk over furniture and any of the other *completely 
unpredictable* obstacles that may lie n its path to the fridge. "Go" embraces 
jumping, leaping, crawling, walking, running, being carried  etc etc - forms of 
movement ad infinitum.  It's a concept, not a variable.

Logical and mathematical variables and differential equations are absolutely 
useless, Steve, if you are undertaking any real world journey anywhere, 
including a journey through a Woz kitchen to make coffee (or eat from a fridge).

However I suspect - and correct me here - concepts are best understood as 
extremely fluid outline "MAPS" rather than "analogs" of objects and object 
actions. My impression is that analog has a fairly precise meaning in 
computation, which has nothing to do with fluid outline maps.

If so, that is what we ideally need - not an analog computer/robot, but a 
***mapping computer** that can fluidly, loosely map the world  and map its body 
onto the world  -  a *retinal* computer. (The retina if you think about it does 
indeed literally fluidly and distortedly map the objects of the world - 
distortions which have to be corrected by the brain).



      AGI | Archives  | Modify Your Subscription  

      AGI | Archives  | Modify Your Subscription   



-------------------------------------------
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