QUT NRC Technical Reports

[Prof Joachim Diederich]

                Computers that learn vs. Users that learn:
                 Experiments with adaptive e-mail agents

                          Joachim Diederich*
                          Elizabeth M. Gurrie**
                          Markus Wasserschaff***

                    Neurocomputing Research Centre*
                      School of Information Systems**
                  Queensland University of Technology
                      Brisbane Q 4001 Australia


       German National Research Center for Computer Science (GMD)***
            Institute for Applied Information Processing (FIT)
                            P.O. Box 1316
                     D-5205 St. Augustin 1, Germany


                         QUTNRC-95-01-01.ps.Z

                               Abstract

        The classification, selection and organization of electronic
messages (e-mail) is a task that can be supported by a neural information
processing system.  The objective is to select those incoming messages for
display that are most important for a particular user, and to propose
actions in anticipation of the user's decisions.  The artificial neural
networks (ANNs) extract relevant information from incoming messages during a
training period, learn the response to the incoming message, i.e., a
sequence of user actions, and use the learned representation for the
proposal of user actions.  We test the system by comparing simple recurrent
networks (SRNs, Elman,1990) and recurrent cascade correlation networks (RCC,
Fahlman, 1991) by use of a sequence production task. The performance of both
network architectures in terms of network size and learning speed for a
given data set is examined.  Our results show that (1) RCC generates smaller
networks with better performance compared to SRNs and (2) learns
significantly faster than SRNs.

Submitted for publication. This is an extended version of the IJCAI-93
paper.

***************************************************************************

                  A Survey And Critique of Techniques
                          For Extracting Rules
                From Trained Artificial Neural Networks

                           Robert Andrews* **
                          Joachim Diederich*
                           Alan B. Tickle* **

                   Neurocomputing Research Centre*
                     School of Information Systems**
                 Queensland University of Technology
                     Brisbane Q 4001 Australia

                         QUTNRC-95-01-02.ps.Z

                              Abstract


	It is becoming increasingly apparent that without some form of
explanation capability, the full potential of trained Artificial Neural
Networks (ANNs) may not be realised. This survey gives an overview of
techniques developed to redress this situation. Specifically the survey
focuses on mechanisms, procedures, and algorithms designed to insert
knowledge into ANNs (knowledge initialisation), extract rules from
trained ANNs (rule extraction), and utilise ANNs to refine existing
rule bases (rule refinement). The survey also introduces a new taxonomy
for classifying the various techniques, discusses their modus operandi,
and delineates criteria for evaluating their efficacy.

Keywords: rule extraction from neural networks, rule refinement using neural
networks, knowledge insertion into neural networks, fuzzy neural networks,
inferencing, rule generation

Accepted for: Knowledge-Based Systems. Special Issue on Knowledge-Based
Neural Networks (Editor: Prof LiMin Fu).

These papers are available from

ftp.fit.qut.edu.au

cd to /pub/NRC/tr/ps