Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.mathworks.com!europa.chnt.gtegsc.com!howland.reston.ans.net!nntp.crl.com!pacbell.com!amdahl.com!amd!amd.com!txnews.amd.com!news
From: Joseph Bridgewater <bridgwtr@vanzandt.amd.com>
Subject: RFD: My Learning/Thinking Neural Network
Content-Type: text/plain; charset=us-ascii
Message-ID: <DDBu2y.C3G@txnews.amd.com>
To: smyth@zorro
Sender: news@txnews.amd.com
Nntp-Posting-Host: corgi
Content-Transfer-Encoding: 7bit
Organization: Advanced Micro Devices, Austin, TX, USA
References: <40bge2$6ib@kaleka.seanet.com> <40n0a5$j18@server2.rz.uni-leipzig.de>
Mime-Version: 1.0
Date: Tue, 15 Aug 1995 00:44:56 GMT
X-Mailer: Mozilla 1.1 (Windows; U; 32bit)
Lines: 52

Darragh Smyth <smyth@zorro> wrote:
>...could the authors of
>the thinking NN please define what they mean by the term ``thinking''?

We were only able to come to an understanding (out own understanding) of what it would mean for our NN "to 
think" after very much contemplation and study.  We studied the human brain, ourselves, human phsycology 
(um, not spelling), other NNs, etc..  There were 'mind-sets' we had to overcome; we had to decide that we 
would be able to figure out a meaning (our own conclusions anyway).  Only after very much work and effort, 
were we able to decide what our design must be able to do for it "to think".  Please understand that I 
present to you that we feel justified in describing our design (note; I say design; the implementation is 
still an on-going process) as being able to "think" or more to the point; it does "think", however; I leave 
it up to you and all others to determine on your own what it is to "think"; we overcame the mind sets of 
you-all where you-all were telling us "We Don't Know What Thinking Is".  This mind set, I put to you, 
_cannot_ develop what one may call _true_ AI.

>Everyone has there own feeling for what it means, depending on which
>side of the bed we got out of that morning! [...] I am not
>sure we know enough about the brain in humans and other animals to
>give a once-off definition on ``thinking''.

I feel, IMHO, that this is definable, but I consider my definition to be the product of very much hard 
work.

>I think it would be more
>constructive for all of us to describe the properties of this NN
>in terms of what it does, its dynamics, capabilities, limits etc,
>rather than just that it thinks! 

That is the purpose of the Request For Discussion; for us all to share our views on this (and there are 
many other pursposes along these lines).

>This term is a bit meaningless you know, because we haven't defined it really.

I do not agree, but nor am I willing, at this time, to share a definition with you-all.

> So please explain what the features are of the NN that you relate to thinking, without 
>actually using the word thinking! [...]

Please find further discussion within this newgroup.  I just read a particually interting post from Stephan 
Verbeeck, Belgium, which I will be responding to at my earliest convienince.  I will need to research our 
design some to provide answers to his wonderfull questions.

>thanks,
>Darragh

This reply is being emailed and posted.

Regards,
Joseph



