home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai.neural-nets
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!rpi!usenet.coe.montana.edu!news.u.washington.edu!uw-beaver!ubc-cs!unixg.ubc.ca!kakwa.ucs.ualberta.ca!alberta!arms
- From: arms@cs.UAlberta.CA (Bill Armstrong)
- Subject: Re: function approximation with nets.
- Message-ID: <arms.716504403@spedden>
- Sender: news@cs.UAlberta.CA (News Administrator)
- Nntp-Posting-Host: spedden.cs.ualberta.ca
- Organization: University of Alberta, Edmonton, Canada
- References: <1992Sep14.053408.20681@usage.csd.unsw.OZ.AU>
- Date: Mon, 14 Sep 1992 21:00:03 GMT
- Lines: 30
-
- s8818037@spectrum.cs.unsw.oz.au (Matthew James Leayr) writes:
-
- > I was wondering how one would go about training a network to perform
- >an operation such as conversion from x-y to polar coordinates. Do you simply
- >present examples of the conversion to the net in the training data?
-
- Sure.
-
- If so,
- >how well suited are neural networks to this sort of task? Could a NN be used
- >to perform a log-polar coordinate transform?
-
- Sure. However, the result would be approximate and complicated to compute.
- Compared to the simple exact results you can get with a calculator,
- what's the point?
-
- Actually, if you supposed that you needed to do some complicated
- coordinate conversion like I believe one finds in quantum physics or
- relativity, and you needed to do it very fast, but the computation of
- the transformation is too slow, then you might use a neural net to
- speed it up inside your supercomputer. (My choice of course would be
- to use an ALN to get the speed.)
-
- Bill
-
- --
- ***************************************************
- Prof. William W. Armstrong, Computing Science Dept.
- University of Alberta; Edmonton, Alberta, Canada T6G 2H1
- arms@cs.ualberta.ca Tel(403)492 2374 FAX 492 1071
-