sticklen@cpswh.cps.msu.edu (Jon Sticklen) (09/22/89)
AVIGNON 90
Expert Systems & their Applications
Tenth International Workshop
Avignon - France, May 28 - June 1, 1990.
Specialized Conference:
Second Generation Expert Systems
Combining Different Reasoning Techniques
CALL FOR PAPERS
The term "Second Generation Expert Systems" is used to characterize
knowledge-based systems able to solve problems by combining different
types of reasoning. Such systems often use multiple representations of
the problem to develop different problem-solving strategies.
The first generation expert systems were largely based on heuristic,
associational rules. To overcome their limitations, a new line of
research was begun into the use of deeper knowledge, often referred to as
"model-based", "causal" or "qualitative" reasoning. Since model-based and
heuristic approaches appear to be largely complementary, recent work has
begun to combine these two reasoning processes into a single
problem-solver.
Another thread of research has been aimed at elaborating "task-specific
architectures": problem solving models which attempt to mirror the
epistemological function and control operators found in the problem
domain. Research has been conducted into designing particular problem
solvers by combining multiple generic or primitive task-specific
architectures.
More generally, Second Generation Expert Systems combines representations
and reasoning techniques that complement one another in some way, such as
heuristic and model-based, quantitative and qualitative or classification
and abduction.
Following the first specialized conference on Second Generation Expert
Systems in 1989, AVIGNON '90 will devote one of its specialized
conferences to the study of practical and theoretical problems raised by
the cooperation of different reasoning techniques.
Topics
The Program Committee is seeking papers on the theme of combining
different reasoning types:
- architectures integrating heuristic and model-based reasoning;
- reasoning with multiple models;
- multi-expert, multi-agent cooperation;
- cooperation of distributed problem-solvers;
- combining multiple task-specific architectures;
- applications of these techniques to real-world problems
(e.g. diagnosis, design, scheduling);
- combining different types of reasoning to enhance knowledge
acquisition or explanation;
- the use of qualitative, model-based, causal or temporal reasoning
to supplement heuristic reasoning;
- integrating qualitative and quantitative reasoning;
- etc.
Papers describing applications should outline the strengths as well as the
weaknesses of the implemented systems. In particular, examples and
analysis of failures will be appreciated in order to delineate the
applicability of the methods used.
Theoretical papers should be clearly related to previous work and should
enlighten the advantages and originality of the proposed approach.
Submission
Authors should submit 6 copies of their papers before December 15, 1989 to
AVIGNON '90 general chairman:
Jean-Claude Rault
EC2
269-287, rue de la Garenne ; 92000 Nanterre ; France
tel: 33 - 1 - 47.80.70.00 ; fax: 33 - 1 - 47.80.66.29
Paper should be a minimun of 2000 words to a maximun of 5000 words (about
10 pages single-spaced). Each submission should contain the following
information: title of paper; full name of all authors; complete address of
first author (including telephone, fax number and e-mail address if
available); abstract of 100-200 words; list of key-words.
Each submission will be reviewed by at least three referees.
Notifications of acceptance or rejection will be mailed from February 20, 1990.
Program Committee
Chairman : Jean-Marc David
Renault ; Service Systhmes Experts
9-11, Av. du 18 Juin 1940 ; 92500 Rueil Malmaison ; France.
e-mail: david@renault.uucp ; tel: 33 - 1 - 47.32.02.53 ;
fax: 33 - 1 - 47.51.29.91
Alice Agogino (University of California; Berkeley, USA); Bert Bredeweg
(University of Amsterdam; The Netherlands); B. Chandrasekaran (Ohio State
University; Columbus, USA); Marie-Odile Cordier (Universite de Rennes;
France); Jean-Luc Dormoy (Etudes et Recherches EDF; Clamart, France);
Jacques Ferber (Universite Paris 6; France); Massimo Gallanti (CISE;
Segrate, Italy); Walter Hamscher (Price Waterhouse; Menlo Park, USA);
Jean-Paul Krivine (Sedco-Forex Schlumberger; Montrouge, France); Benjamin
Kuipers (University of Texas; Austin, USA); Roy Leitch (Heriot-Watt
University; Edinburgh, UK); Robert Milne (Intelligent Applications;
Livingston, UK); Richard Pelavin (Philips Laboratories; Briarcliff Manor,
USA); Olivier Raiman (IBM Scientific Center; Paris, France); Reid Simmons
(Carnegie Mellon University; Pittsburgh, USA); Luc Steels (Vrije
Universiteit; Brussels, Belgium); Jon Sticklen (Michigan State University;
East-Lansing, USA); Pietro Torasso (Universita di Torino; Italy); Louise
Trave-Massuyes (LAAS - CNRS; Toulouse, France); Walter Van de Velde (Vrije
Universiteit; Brussels, Belgium).
---jon---
-------------------------------------------------------------
Jon Sticklen
Artificial Intelligence/Knoweldge Based Systems Group
Computer Science Department
Michigan State University
East Lansing, MI 48824-1027
517-353-3711
-------------------------------------------------------------