From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!gatech!mailer.cc.fsu.edu!sun13!ntuix!eoahmad Thu Apr 30 15:22:49 EDT 1992
Article 5270 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!gatech!mailer.cc.fsu.edu!sun13!ntuix!eoahmad
>From: eoahmad@ntuix.ntu.ac.sg (Othman Ahmad)
Newsgroups: comp.ai.philosophy
Subject: Quantitative measure of Intelligence
Message-ID: <1992Apr23.083023.14050@ntuix.ntu.ac.sg>
Date: 23 Apr 92 08:30:23 GMT
Organization: Nanyang Technological University Singapore
Lines: 36

Alan Turing test is only a subjective test. I have developed a way of
measuring Intelligence quantitatively using Information Theory.
 Information theory just measures the non-determinism of the occurrence of a
 symbol. This symbol defines the information. The objection is that it does not
 give any semantic content to the information. However the symbol representing
 the semantic only has meaning defined by the user. For example a symbol defines
 the action of switching on a light. A meaning attached to a symbol is just 
defined by concensus. A pen is red because we all agree that it is red. There is
 no other explaination.
	Information theory had been used extensively for analysis of 
communcation channels. A few attempts had been made by Saridis and Goodman 
to use Information theory to measure Intelligence but what they are actually 
doing is to measure the knowledge or throughput of the intelligent machine.
	There lies the controversy of my theory. I separate Intelligence from
Knowledge and Solution(throughput). Current AI beliefs defines intelliigence in
combinations of knowledge and solutions. This method looks like an electrical 
engineer trying to understand an electrical black box by measuring power only.
	That is why AI people come up with various complicated equations, which
are not wrong but unnecessarity complex and does not give further understanding
for the thinking process.
	The essence of my argument is that an unpredictable machine exibits moreintelligence than a predictable one but both can come up with the same solution.Obviously the intelligent machine needs more reasoning steps than the stupid onebut the stupid one alr
eady has the knowledge(stored intelligence).
	I have difficulty in publishing this theory but I badly need to refer toit. We do not have an internal report system at NTU. If someone had known of anypublication that refer to the same idea, please inform me. If someone needs moredetail on this idea I 
am willing to send him copies of the letters that I had 
tried to be published. Hopefully someone could develop it further so thst it is
suitable for publication.
	Or I could be grossly wrong. So far all the editors had not mentioned my
errors. They only say that it is too short/undeveloped. I am not interested in
developing it further than what I need. I am basically a circuit designer. The
theory is already useful to me as it is. In fact I am able to explain the Truth
table and Chinese in a room, problems mentioned in this group.
	You could point out any error in this method so that I
could either abandon or improve it.

Othman bin Ahmad, School of EEE,
Nanyang Technological University, Singapore 2263.
Internet Email: eoahmad@ntuix.ntu.ac.sg
Bitnet Email: eoahmad@ntuvax.bitnet


