Path: utzoo!attcan!uunet!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!uflorida!winnie!zach!faee0ntt
From: faee0ntt@zach.fit.edu ( N. Tepedelenlioglu)
Newsgroups: comp.ai.neural-nets
Subject: Re: Learning parity function by backprob.
Message-ID: <1442@winnie.fit.edu>
Date: 10 Oct 90 01:05:54 GMT
References: <4803@tuminfo1.lan.informatik.tu-muenchen.dbp.de>
Sender: usenet@winnie.fit.edu
Reply-To: faee0ntt@zach.UUCP ( N. Tepedelenlioglu)
Organization: Florida Institute of Technology, ACS, Melbourne, FL
Lines: 23

In article <4803@tuminfo1.lan.informatik.tu-muenchen.dbp.de> li@kiss.informatik.tu-muenchen.de.informatik.tu-muenchen.dbp.de () writes:
>Parity functions may be realized by NN with one hidden layer (a simple
>solution was given in PDP-1). It is however a hard  problem to get such
>solution by back-propagation algorithm. I was able to train a  NN with
>backprog and some heuristcs to realize the P_4 (i.e. the parity
>function of 4 bits vectors, P_2 is the XOR function). The P_5 seems, by
>my experience, already to be too difficult to be learned by backprop,

The key is the number of nodes in the hidden layer. That number 
should be at least as big as the number of bits at the input. 
So if you try a net with say 8 nodes in the hidden layer I
am pretty sure you will have no difficulty for the P_5 problem.
>
>Thanks.
>Xinzhi Li

Nazif.


__________________________
Nazif Tepedelenlioglu            faee0ntt@zach.fit.edu
Dept. EE/CP
Florida Institute of Technology, Melbourne, FL 32901, USA