home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Monster Media 1993 #2
/
Image.iso
/
text
/
9305nni.zip
/
930526.ANC
< prev
next >
Wrap
Internet Message Format
|
1993-05-26
|
11KB
From ml-connectionists-request@q.cs.cmu.edu Tue May 25 22:26:49 1993
Received: by cse.uta.edu (5.57/Ultrix2.4-C)
id AA14664; Wed, 26 May 93 00:26:30 -0500
Received: from Q.CS.CMU.EDU by Q.CS.CMU.EDU id aa07350; 25 May 93 17:00:40 EDT
Received: from DST.BOLTZ.CS.CMU.EDU by Q.CS.CMU.EDU id aa07234;
25 May 93 16:19:06 EDT
Received: from DST.BOLTZ.CS.CMU.EDU by DST.BOLTZ.CS.CMU.EDU id aa26898;
25 May 93 16:17:50 EDT
Received: from RI.CMU.EDU by B.GP.CS.CMU.EDU id aa25961; 25 May 93 10:37:57 EDT
Received: from [131.246.136.50] by RI.CMU.EDU id aa06990; 25 May 93 9:39:37 EDT
Received: from uklirb.informatik.uni-kl.de by stepsun.uni-kl.de id aa14972;
25 May 93 15:33 MET DST
Received: from ag_vp_file_server.informatik.uni-kl.de
by uklirb.uklirb.informatik.uni-kl.de id aa27340;
25 May 93 15:29 MET DST
Subject: Neural net & fuzzy software (& reports) available
To: connectionists@cs.cmu.edu
From: "Uwe R. Zimmer, AG vP" <uzimmer@informatik.uni-kl.de>
Date: Tue, 25 May 93 15:29:10 +0100
Message-Id: <930525.152910.255@informatik.uni-kl.de>
Status: RO
--- Neural net & fuzzy software (& reports) available! ---
We have installed an experimental FTP-Server, which will keep parts of our
actual work in the form of software and technical reports.
As the first step we have published two of our projects: a neural fuzzy
decision system, and an unsupervised clustering system.
For detailed information please see below.
please send all comments, remarks, etc. pp. to:
uzimmer@informatik.uni-kl.de
--- What is published here? ---
Our group produces as a result of the actual projects (MOBOT, SPIN,
ALBATROSS) a huge amount of code and programs, so we would like to share
some of the sources. We put only these projects on our FTP-server, which
might be from some general interest, beyond the scientific results,
published in the associated papers.
--- How to obtain the associated papers? ---
There is a common FTP-server for the whole university at our campus, which
holds parts of our actual work.
The FTP-information is:
University of Kaiserslautern FTP-Server is : ftp.uni-kl.de
Mode is : binary
Directory is : reports_uni-kl/computer_science/mobile_robots/...
Subdirectory is : 1993/papers
File name is : Zimmer.learning_surfaces.ps.Z
Subdirectory is : 1992/papers
File name is : Zimmer.rt_communication.ps.Z
Subdirectory is : 1991/papers
File names are : Edlinger.Pos_Estimation.ps.Z
Edlinger.Eff_Navigation.ps.Z
Knieriemen.euromicro_91.ps.Z
Zimmer.albatross.ps.Z
Submitted papers and technical reports may be found on:
FTP-Server is: ag_vp_file_server.informatik.uni-kl.de
Mode is : binary
Directory is : Neural_Networks/Reports
--- What are the dedicated machines? ---
Most of our projects are written for the Apple-Macintosh world, so the
ready-to-run programs will require a Macintosh! On the other hand, the
sources are written in Pascal and if you like pascal-code you may have
benefits from the source code without a complete and running application.
--- What sources are published? ---
We make only the "kernel"-sources from the specific projects available,
because they might be of public interest. If you are interested in the
sources for user-interface handling, etc., please let me know.
--- Is the project documentation in english? ---
Sorry - not at all the projects. Some of the technical documentation is
in german!
-------------------------------------------
--- List of actually available projects ---
-------------------------------------------
---------------------------------------------------------------------------
--- Neural Fuzzy Decision System (Joerg Bruske):
---------------------------------------------------------------------------
--- Associated report is (english):
FTP-Server is: ag_vp_file_server.informatik.uni-kl.de
Mode is : binary
Directory is : Neural_Networks/Reports
File name is : Zimmer.NFDS.ps.Z
SPIN-NFDS
Learning and Preset Knowledge for Surface Fusion
- A Neural Fuzzy Decision System -
Jorg Bruske, Ewald von Puttkamer & Uwe R. Zimmer
The problem to be discussed in this paper may be characterized in short by
the question: "Are these two surface fragments belonging together (i.e.
belonging to the same surface)?". The presented techniques try to benefit
from some predefined knowledge as well as from the possibility to refine
and adapt this knowledge according to a (changing) real environment,
resulting in a combination of fuzzy-decision systems and neural networks.
The results are encouraging (fast convergence speed, high accuracy), and
the model might be used for a wide range of applications. The general frame
surrounding the work in this paper is the SPIN-project, where emphasis is
on sub-symbolic abstractions, based on a 3-d scanned environment.
--- Source code and technical documentation (english):
FTP-Server is: ag_vp_file_server.informatik.uni-kl.de
Mode is : binary
Directory is : Neural_Networks/Software/Neural_Fuzzy_Decision
This documentation consists of five chapters:
In Chapter 1, the author presents his approach towards implementing fuzzy
decision systems (FDS) by means of neural nets, leading to his NFDS. In
order to train (optimize) the NFDS, a slightly modified version of the
backpropagation algorithm is introduced.
In Chapter 2, the FuzNet project and its modules are described in detail.
FuzNet implements the NFDS described in Chapter 1 on Apple-Macintosh
computers and has been developed as an easy-integrable SW-component for
larger SW-projects.
In Chapter 3, we will be concerned with the details of the integration of
FuzNet in other SW-projects, taking SurfaceExtractor as an example.
However, the reader need not know the SurfaceExtractor project (which
currently is not supplied via ftp) in order to understand the details of
integrating FuzNet in their projects.
In Chapter 4, the FuzTest application is described. FuzTest is a very
primitive application intended to familiarize the user with FuzNet.
In Chapter 5, the reader will find the syntax diagram for fuzzy data- and
rule- bases as accepted by FuzNet. The file "brakingFDS" contains such a
fuzzy data- and rule- base.
A references list concerning literature about neural nets, fuzzy logic and
neural fuzzy decision systems is appended to this documentation. In
particular, [Bruske93] is recommended for a detailed discussion of neural
fuzzy decision systems and [BruPuttZi93] as a short introduction to NFDS
and one of its applications in the Research Group v. Puttkamer.
---------------------------------------------------------------------------
--- Dynamic unsupervised feature maps (Herman Keuchel):
---------------------------------------------------------------------------
--- Associated report is (english):
FTP-Server is: ftp.uni-kl.de
Mode is : binary
Directory is : reports_uni-kl/computer_science/mobile_robots/1993/papers
File name is : Zimmer.learning_surfaces.ps.Z
SPIN - Learning and Forgetting Surface Classifications
with Dynamic Neural Networks
Herman Keuchel, Ewald von Puttkamer & Uwe R. Zimmer
This paper refers to the problem of adaptability over an infinite
period of time, regarding dynamic networks. A never ending flow of
examples have to be clustered, based on a distance-measure. The
developed model is based on the self-organizing feature maps of
Kohonen [6], [7] and some adaptations by Fritzke [3]. The problem
of dynamic surface classification is embedded in the SPIN project,
where sub-symbolic abstractions, based on a 3-d scanned environment
is being done.
--- Source code and technical documentation (german):
FTP-Server is: ag_vp_file_server.informatik.uni-kl.de
Mode is : binary
Directory is : Neural_Networks/Software/Dynamic_Unsup_Feature_Map
NetSim
Ein Netzwerkmodell fuer die Klassifikation von Flaechen durch Neuronale
Netzwerke mit problemadaptiver Zellstruktur
von Herman Keuchel
Der Netzwerksimulator NetSim und diese Dokumentation (d.h. das hier ist nur
ein Teil der gesamten Dokumentation, allerdings der groessere) entstand im
Rahmen einer Arbeit deren Ziel war, festzustellen, in wie weit Neuronale
Netzwerke nach dem Modell von Bernd Fritzke zur Klassifizierung von
Flaechen
geeignet sind. Fuer eine genaue Beschreibung des Modells nach Bernd
Fritzke
siehe [Fritzke91a & Fritzke91b]. Das Netzwerkmodell ist eine
Weiterentwicklung eines der erfolgreichsten Modelle im Bereich des
unbeaufsichtigten Lernens, der self-organising feature maps von T.
Kohonen [Kohonen84]. Kohonen benutzt eine starre Zellstruktur mit einer
festen Anzahl von Zellen. Beim Modell von Fritzke passt sich die
Zellstruktur und -anzahl dynamisch den zu lernenden Vektoren an. Die
Anzahl der Zellen ist dabei von der gewuenschten
Klassifizierungsgenauigkeit abhaengig.
Die zu lernenden Flaechen werden in Form von m-dimensionalen Vektoren
reeller Zahlen dargestellt. Die Begriffe `Flaeche` und `Vektor` werden
daher im Folgenden oft synonym verwendet. Fuer detaillierte Informationen
ueber die Vektordarstellung von Flaechen siehe [Zimmer91].
In Kapitel 1 werden einige Erweiterungen des Modells erlaeutert, die noetig
wurden, um das Modell den Bedingungen im SPIN-Projekt [Zimmer91]
anzupassen. In diesem Projekt gibt es keine Aufteilung in Lernphase und
Arbeitsphase. Die beiden Phasen laufen waehrend der gesamten Laufzeit des
Systems parallel, damit das Netzwerk jederzeit in der Lage ist, sich
veraendernden Bedingungen anzupassen.
Kapitel 2 geht naeher auf Lernparameter, sinnvolle Defaultwerte, und auf
Auswirkungen von Parameteraenderungen ein. Kapitel 3 stellt den Simulator
und seine Bedienung vor, und in Kapitel 4 wird noch die Pro-
grammierschnittstelle der Simulationssoftware beschrieben. Kapitel 5
schliesslich erlaeutert noch offengebliebene Probleme und enthaelt einige
abschliessende Bemerkungen.
Die Software wurde auf Apple-Macintosh Rechnern in THINK Pascal
implementiert. Fuer den Netzwerksimulator NetSim ist ein Mathematischer
Coprozessor erforderlich. Die Applikation ist auf Macintosh II, IIxi,
IIcx, IIfx und Quadra 950 getestet und lauffaehig.
Fragen, Hinweise und Anregungen koennen an
Herman Keuchel
Parkstrasse 27b
67655 Kaiserslautern
oder an
Uwe R. Zimmer (siehe unten)
gerichtet werden.
-----------------------------------------------------
-----
Uwe R. Zimmer ---
University of Kaiserslautern - Computer Science Department |
Research Group Prof. v. Puttkamer |
6750 Kaiserslautern - Germany |
-------------------------------------------------------------- |
P.O.Box:3049 | Phone:+49 631 205 2624 | Fax:+49 631 205 2803 |