From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!spool.mu.edu!uunet!mcsun!uknet!edcastle!aisb!aifh!aiss Sun Dec  1 13:06:52 EST 1991
Article 1776 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!spool.mu.edu!uunet!mcsun!uknet!edcastle!aisb!aifh!aiss
>From: aiss@aifh.ed.ac.uk (Sven Suska)
Newsgroups: comp.ai.philosophy
Subject: Re: Defs (I)
Keywords: intelligence, definition,
Message-ID: <1991Nov30.184207@aifh.ed.ac.uk>
Date: 30 Nov 91 18:42:07 GMT
Article-I.D.: aifh.1991Nov30.184207
References: <1991Nov21.153122.15464@cc.ic.ac.uk> <1991Nov27.193459.1670@aisb.ed.ac.uk> <1991Nov28.170726.27890@cc.ic.ac.uk>
Sender: news@aisb.ed.ac.uk (Network News Administrator)
Organization: Dept of AI, University of Edinburgh, Scotland
Lines: 71


In article <1991Nov28.170726.27890@cc.ic.ac.uk>, redgers@sig.ee.ic.ac.uk
(Adrian Redgers) writes:
# Sven Saska (aiss@aifh.ed.ac.uk) and I are designing (haha) a Quantative 
# Definition of Intelligence (QDI) that can be used to compare intelligent 
# machines despite them being very different.

Only machines? -I hope not.
 
1. dead and living intelligence

To me there  still is one sense of intelligence in which 
every 100% deterministic thing has zero intelligence.
I don't know if this can be justified, I will just assume it
now. (!Cut: I don't intend to go into the Chinese room debate etc.)
You can call it an 'iota of Buddha Nature', and why should it
be impossible to measure it. Or at least study it, not (only) by
looking at our navels, but by examining NNs.
A more usual term for this 'living intelligence' would be
'ability to learn'(not memorise), 'ability to adapt'.

If everything can be modelled deterministically, then this
idea obviously collapses.

On the other hand, intelligence also means something like
mental power, ability to produce good answers.
This of course depends on the knoledge(or adaption/programming)
already aquired. Thus it is domain dependent.
Mental power would comprise both living and dead intellgc.
What is dead intelligence: rules, algorithms, dogmatic ideologies
- It's of the same stuff as knowledge.

2. measuring
I didn't really understand what you want to define by your
preliminary QDI (counting functions) it looks like complexity
to me. But can we ignore the environment? I cannot imagine
how the subjects you are defining it for are showing
intelligent behavior in every (suitable) situation. 
You still have to convince me of your function approach.
What is inside your black box when the QD is applied?
(Is the look-up-table empty or patially istantiated?)
If it were completely filled you could easily compute
how good the thing is adapted, like
    weighted sum of value(response(Input)) over all Input
      where value is how good the response is in the given universe.
You seem to suggest that all cases could be handled with this
no-internal-memory model by externalising the memory (by inroducing
extra inputs and outputs) or by simulating the whole thing in a
meta table-look-up. 
This would probably not be helpful. 

I think you have to account for the structure, that's my
objection against mere counting.




# S5: This thread is rather 1-ply (or 2-ply): perhaps we ought to move to 
#     e-mail, Sven.

I don't feel a need for it as long as some people keep posting
boring personal attacks in this newsgroup. 

 
# /*       And, as with Gods and men, the sheep remain inside their pen, 
#                Though many times they've seen the way to leave...     
# Adrian Redgers : redgers@sig.ee.ic.ac.uk : Neural Systems Lab, Elec. Eng.,



Sven Suska       *          de omnibus est dubitandum


