Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!uakari.primate.wisc.edu!sdd.hp.com!elroy.jpl.nasa.gov!ncar!boulder!eesnyder
From: eesnyder@boulder.Colorado.EDU (Eric E. Snyder)
Newsgroups: comp.ai.neural-nets
Subject: Networks on a chip (was: back-propagation)
Message-ID: <29425@boulder.Colorado.EDU>
Date: 8 Nov 90 15:57:26 GMT
References: <39365@ut-emx.uucp>
Sender: news@boulder.Colorado.EDU
Reply-To: eesnyder@boulder.Colorado.EDU (Eric E. Snyder)
Organization: University of Colorado, Boulder
Lines: 22
Nntp-Posting-Host: beagle.colorado.edu


This letter bounced... but it might be of general interest, so:
In article <1990Nov06.152539.17957@undeed.uucp> you write:
>
>I am designing a reconfigurable neural-net chip which is arranged
>as several layers, each containing a number of processing elements
>( nodes )....

I am curious how complicated a net you can build from hardware on
a single chip, how many nodes, how much connectivity?  

Just for background, I am a biologist and have been working on the
application of neural nets for modeling structure-activity relationships.

---------------------------------------------------------------------------
TTGATTGCTAAACACTGGGCGGCGAATCAGGGTTGGGATCTGAACAAAGACGGTCAGATTCAGTTCGTACTGCTG
Eric E. Snyder                            
Department of MCD Biology            We are not suspicious enough 
University of Colorado, Boulder      of words, and calamity strikes.
Boulder, Colorado 80309-0347
LeuIleAlaLysHisTrpAlaAlaAsnGlnGlyTrpAspLeuAsnLysAspGlyGlnIleGlnPheValLeuLeu
---------------------------------------------------------------------------