home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!sun-barr!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!ira.uka.de!uka!i41s14!prechelt
- From: prechelt@i41s14.ira.uka.de (Lutz Prechelt)
- Newsgroups: comp.ai.neural-nets
- Subject: Comparison of SuperSAB vs. Quickprop
- Date: 14 Sep 1992 09:13:37 GMT
- Organization: University of Karlsruhe, FRG
- Lines: 54
- Distribution: world
- Message-ID: <191l41INNi23@iraul1.ira.uka.de>
- NNTP-Posting-Host: i41s14.ira.uka.de
-
-
- Has anybody compared the SuperSAB variant of Backpropagation
- (Tollenaere) with the Quickprop variant (Fahlman) ?
-
- I would like to know about any experiences.
- Or does somebody know about any paper where such a comparison is done
- or mentioned ?
-
- Here are the principal references to SuperSAB and Quickprop:
-
- @article{Tollenaere90,
- author = {Tom Tollenaere},
- title = {SuperSAB: Fast Adaptive Backpropagation with Good
- Scaling Properties},
- journal = {Neural Networks},
- year = {1990},
- Volume = {3},
- number = {},
- pages = {561-573},
- month = {},
- Class = {nn, learning},
- abstract = {},
- notes = {^^, very good paper},
- }
-
- @techreport{Fahlman88,
- author = {Scott E. Fahlman},
- title = {An Empirical Study of Learning Speed in Back-Propagation
- Networks},
- institution = {School of Computer Science, Carnegie Mellon University},
- year = {1988},
- Type = {},
- number = {CMU-CS-88-162},
- address = {Pittsburgh, PA 15213},
- month = sep,
- Class = {nn, learning},
- abstract = {Excellent paper.
- discusses "what is a good benchmark?", encoder/decoder task,
- complement encoder/decoder, XOR,
- termination of learning, how to report learning times,
- tuning of backprop learning parameters, derivation of quickprop },
- notes = {^^, neuroprose},
- }
-
-
- Please reply by email, I will summarize.
-
- Lutz
-
- --
- Lutz Prechelt (email: prechelt@ira.uka.de) | Whenever you
- Institut fuer Programmstrukturen und Datenorganisation | complicate things,
- Universitaet Karlsruhe; D-7500 Karlsruhe 1; Germany | they get
- (Voice: ++49/721/608-4317, FAX: ++49/721/694092) | less simple.
-