[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: RBFs

To: krauss@iafrica.com
Subject: Re: RBFs
From: "Randall C. O'Reilly" <oreilly@grey.colorado.edu>
Date: Tue, 28 Nov 2000 16:58:00 -0700
CC: pdp-discuss@psych.colorado.edu
In-reply-to: <00c401c0594b$54bdfb90$581b07c4@bv> (message from Bruno VOISIN onTue, 28 Nov 2000 16:52:34 +0200)
References: <00c401c0594b$54bdfb90$581b07c4@bv>
Reply-to: oreilly@psych.colorado.edu
Sender: owner-PDP-DISCUSS@psych.colorado.edu

Bruno VOISIN <krauss@iafrica.com> writes:
> My question is the following:
> PDP V1 used to cover RBFs (or wasn't it?). I see no mention to that =
> family of neural nets in V2.

RBFBpUnitSpec still exists in bp++, and is described in the manual.

Also, in so++, the SoftClUnitSpec (soft competitive learning) computes a
Gaussian function for the net input, meaning it is a self-organizing
RBF unit.

				- Randy

Follow-Ups:
- Employment opportunity for pdp++ wizard
  - From: "Norman Packard" <n@predict.com>

References:
- RBFs
  - From: Bruno VOISIN <krauss@iafrica.com>

Prev by Date: RBFs
Next by Date: Re: Kohonen networks and internals
Prev by thread: RBFs
Next by thread: Employment opportunity for pdp++ wizard
Index(es):
- Date
- Thread