Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!ira.uka.de!fauern!NewsServ!sunmanager.lrz-muenchen.de!uh311ae!sunmanager!uh311ae
From: uh311ae@sunmanager.lrz-muenchen.de (Henrik Klagges)
Newsgroups: comp.ai.neural-nets
Subject: Convervence vs no. of hidden layers + units
Message-ID: <uh311ae.675803874@sunmanager>
Date: 3 Jun 91 14:37:07 GMT
Organization: The Internet
Lines: 21

Hello,
I would like to know wether 
a) Convergence speed generally decreases as one adds hidden layers/units
b) Final convergence generally increases as one adds hidden layers/units.

If these generalizations are untrue, I would like to know wether there is
something like an algorithm's feature determining these scale properties.

Thanks & Cheers ! Rick@vee.lrz-muenchen.de
U of M