Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uwm.edu!ux1.cso.uiuc.edu!herodotus.cs.uiuc.edu!kadie
From: kadie@herodotus.cs.uiuc.edu (Carl M. Kadie)
Newsgroups: comp.ai.neural-nets
Subject: Reasons for NN (was back-prop NNs and `SAS' regression!)
Message-ID: <1989Dec19.201357.15543@ux1.cso.uiuc.edu>
Date: 19 Dec 89 20:13:57 GMT
References: <220700005@uxe.cso.uiuc.edu> <44929@bu-cs.BU.EDU> <15039@boulder.Colorado.EDU>
Sender: news@ux1.cso.uiuc.edu (News)
Reply-To: kadie@herodotus.cs.uiuc.edu.UUCP (Carl M. Kadie)
Organization: University of Illinois, Urbana-Champaign
Lines: 24

In article <15039@boulder.Colorado.EDU> bill@synapse.Colorado.EDU 
   (Bill Skaggs) writes:
...
>  The secondary reason (IMHO) is that neural nets are massively parallel.
>When one has reached the limits of sequential speed, one must go to
>parallelism in order to get greater power.  Neural nets are unlikely to
>ever provide especially _elegant_ solutions to very many problems:  their
>virtue is that they provide a brutal and simplistic solution that sometimes
>(surprisingly) actually works.  
...

Anyone interested in parallel machine learning systems should look at:

Omohundro, S. (1987) Efficient algorithms with neural network behavior. Complex
     Systems, 1:273-347.

He shows how to parallelize an ID3-like algorithm. Even when run on
a serial machine, ID3 is  much, much faster than most neural-inspired
algorithms. Putting it on a parallel machine makes it faster still.


Carl Kadie
University of Illinois at Urbana-Champaign
ARPA:  kadie@m.cs.uiuc.edu