Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!nntp.sei.cmu.edu!cis.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!news.starnet.net!wupost!nic.smsu.edu!newsdist.tc.umn.edu!umn.edu!peschko
From: peschko@cs.umn.edu (Edward S. Peschko)
Subject: translation problem...
Message-ID: <peschko.818634125@cs.umn.edu>
Sender: news@news.cis.umn.edu (Usenet News Administration)
Nntp-Posting-Host: micro.cs.umn.edu
Organization: University of Minnesota, Twin Cities
Date: Sun, 10 Dec 1995 22:22:05 GMT
Lines: 30

hey all --

Say I have a pattern that I want a neural net to learn -- it exists on a grid:

.....X....
....XOX...
....X.X...
..........
..........
..........

Now, as far as I know, there are no net algorithms that can recognize say this
pattern shifted once to the left -- ie:

......X...
.....XOX..
.....X.X..
..........
..........
..........


Is this true? Or is there some network (variation on backpropogation perhaps)
that can recognize both patterns, after training with the first?

(same goes with rotation)

Thanks much,

Ed
