home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai.neural-nets
- Path: sparky!uunet!decwrl!access.usask.ca!kakwa.ucs.ualberta.ca!alberta!arms
- From: arms@cs.UAlberta.CA (Bill Armstrong)
- Subject: Re: Reducing Training time vs Generalisation
- Message-ID: <arms.714514327@spedden>
- Sender: news@cs.UAlberta.CA (News Administrator)
- Nntp-Posting-Host: spedden.cs.ualberta.ca
- Organization: University of Alberta, Edmonton, Canada
- References: <Bt9GIx.9In.1@cs.cmu.edu> <?.714340347@tazdevil>
- Date: Sat, 22 Aug 1992 20:12:07 GMT
- Lines: 37
-
- henrik@mpci.llnl.gov (Henrik Klagges) writes:
-
- >sef@sef-pmax.slisp.cs.cmu.edu writes:
-
- >>Well, since you keep pounding on this, I will point out that in most
- >>backprop-style nets after training, almost all of the hidden units are
- >>saturated almost all of the time. So you can replace them with sharp
-
- >Same in our experiments. The decision trees being built do benefit a lot
- >from the remaining nonlinearities, though (smoother decision surfaces-
- >really 8-).
-
- Isn't it a question of smooth vs discontinuous? If you don't need
- continuity, then you don't need the sigmoids at all.
-
- >>Myself, I prefer to think in terms of parallel hardware, so lazy evaluation
- >>isn't an issue. Yes, sigmoid unit hardware is a bit more expensive to
- >>implement than simple gates, but I don't need nearly as many of them.
-
- >It is not terribly expensive - a 256 entry table is usually enough. Pipe
- >lined access to such a lookup table can be made at one lookup/cycle at a
- >pipe stall of less than 5 (if not much better, hihi). Moreover, weight
- >accumulation/update are matrix operations, while lookup is only a vector
- >operation. It is no bottleneck at all.
-
- OK, but looking up things in RAM is just using combinational logic.
- Your ALN competitor will be finished the entire NN computation in
- the time it takes to do two or three of these lookups.
-
- Another problem is you seem to assume that numerical accuracy is not
- any issue at all.
-
- --
- ***************************************************
- 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
-