home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!tcsi.com!iat.holonet.net!news.cerf.net!usc!sol.ctr.columbia.edu!hamblin.math.byu.edu!yvax.byu.edu!cunyvm!k77bc
- Newsgroups: comp.ai.neural-nets
- Subject: a beginner's quiestion
- Message-ID: <93026.180345K77BC@CUNYVM.BITNET>
- From: <K77BC@CUNYVM.BITNET>
- Date: Tuesday, 26 Jan 1993 18:03:45 EST
- Organization: City University of New York/ University Computer Center
- Lines: 15
-
- In the october's 1990 issue Mr. Larry O'brien published his
- C++ code for backprop NN.
- The was designed to work only with one hidden layer, but it was
- designed with multilayer upgrade in mind.
- I rewrote one function and NN now works ( i think ) with multiple
- hidden layers.
-
- My question is: When I run the XOR problem with one hidden layer
- it converges (abeit slowly). But if I use more than one hidden
- layer, it gets stucked (in local maximum?).
- Is this supposed to happen or It's just my code...?
-
- Thank you.
- -Yevgeny k77bc@cunyvm.cuny.edu
-
-