Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucsd!ucsdhub!hp-sdd!ncr-sd!se-sd!jim
From: jim@se-sd.NCR.COM (Jim Ruehlin)
Newsgroups: comp.ai.neural-nets
Subject: Re: Meta-Nets
Message-ID: <2519@se-sd.NCR.COM>
Date: 21 Mar 90 17:11:10 GMT
References: <4670@cbnewsl.ATT.COM>
Reply-To: jim@se-sd.SanDiego.NCR.COM (Jim Ruehlin)
Organization: NCR Corporation, SE-San Diego
Lines: 12
Summary:
Expires:
Sender:
Followup-To:
Keywords:

In article <4670@cbnewsl.ATT.COM> apr@cbnewsl.ATT.COM (anthony.p.russo) writes:
>I tend to think that the meta net, at best, would learn to
>implement the standard algorithm. That is, we are training it to
>learn some known algoithm. If this is the case, why not just use 
>the standard alg and skip the meta net?

As I remember from my NN classes, the exciting thing about them is that they
begin to generalize (if designed and trained properly).  Hopefully, the
meta-net would generalize learning heuristics based on the standard
algorithm it was taught.

- Jim Ruehlin