Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!news.kei.com!eff!news.umbc.edu!haven.umd.edu!darwin.sura.net!news.Vanderbilt.Edu!NewsWatcher!user
From: goldenjb@ctrvax.vanderbilt.edu (jim golden)
Subject: Re: "Trading... - Oh... really?" Part II (last)
Message-ID: <goldenjb-261094142042@129.59.170.62>
Followup-To: comp.ai.neural-nets
Sender: news@news.vanderbilt.edu
Nntp-Posting-Host: 129.59.170.62
Organization: vanderbilt
References: <83591.arcon@dial.illinois.net>
Date: Wed, 26 Oct 1994 19:07:11 GMT
Lines: 37

In article <83591.arcon@dial.illinois.net>, "USTO"
<arcon@dial.illinois.net> wrote:



> 3.) ...As to the hart of the matter... (The conceptual issues and 
> problems with the book.)
> We would like to point to a (somewhat long) quote by Robert Hecht-Nielsen:
> 
> "... This illustrates yet another aspect of neurocomputing applications: 
> the only important expertise is that associated with writing a good 
> systems specification and devising a clever way to exploit the power of 
> the neural network(s) involved. In-depth technical knowledge of the 
> particular area of information processing being considered is typically 
> not necessary; although  a general knowledge of the area is often required. 
> Thus, in this sense, neurocomputing places a higher premium on domain 
> knowledge and network use cleverness than on deep technical understanding.
> In the future, most neural-network-based information processing systems 
> may routinely carry out operations that no human understands in detail on 
> an algorithmic basis. It may even turn out that the study of algorithms  
> will become less and less important, and eventually evolve into an obscure 
> academic pursuit...."
>                                              - R.H.N -

I missed what this original thread was all about, but in regards to the
above quote and with due respect to RHN:

 wrong song dingdong!

That sentiment is a load of hooey and accounts for the 95% bullshit factor
in most papers presented at all those terrible IJCNN/INNS conferences every
year. I was hoping we'd shaken the black box approach to ANNS which has
lead to some awful publications, failed stock market prediction programs,
and really silly grant proposals.

Jim Golden
goldenjb@ctrvax.vanderbilt.edu
