🚀 go-pugleaf

From: "Markus P.J. Fromherz" <fromherz@ifi.unizh.ch>
To: mod-ki@unido.informatik.uni-dortmund.de
Date: Fri, 1 Nov 1991 11:05:34 +0100
Subject: <mod-ki 0788> Prolog Forum 6.11.91
Keywords: Vortraege

mod-ki message 0788
************************************************************************
Prolog Forum                  UNI/ETH Zurich                  WS 1991/92

Wednesday, Nov. 6, 1991, 12.30 at room 21-G-74 (below the mensa)
University of Zuerich-Irchel, Winterthurerstr. 190


                         Lawrence Chachere
                     CUI, Universite de Geneve


            Advantages of a Logic-Based Representation for
               Knowledge Acquisition in Expert Systems

Knowledge acquisition is the bottleneck of expert system development.
One major component of this bottleneck is the difficulty associated with
refining information within a partially constructed knowledge base.
Thus, many recent research efforts have focused on the automation of
knowledgebase refinement techniques. The development of MINERVA by the
Knowledge Base Systems group at the University of Illinois was one such
research effort.

MINERVA is a Prolog reimplementation and enhancement of HERACLES,
the expert system shell of NEOMYCIN. The reimplementation was guided
by experience gained from the ODYSSEUS apprenticeship learning program
for HERACLES. MINERVA has an advantage over the older shell because it
accommodates inference engines for learning, in addition to the general
inference engine for problem solving.

It is emphasized that the improved shell is the result of the
restructuring of knowledge. No new knowledge was added to the system. In
this talk, I will describe how the Prolog representation of the shell
facilitates two machine learning techniques, both of which have already
produced significant results for acquiring medical domain knowledge. I
will also discuss how meta-interpreting, a special feature of Prolog,
facilitates the development of inference engines.

This work was done at the University of Illinois in Urbana.

For more information:

    Markus Fromherz                   Tel.:   +41-1-257 4313 / 4311
    Institut fuer Informatik          E-Mail: fromherz@ifi.unizh.ch
    Universitaet Zuerich-Irchel
    Winterthurerstrasse 190
    CH-8057 Zuerich
************************************************************************