Tutorial T05 - Symbolic Reasoning with Subsymbolic Tasks

Presented by Dr J. Zeleznikow

Eighth Australian Joint Conference on Artificial Intelligence (AI'95)

14 November 1995

Canberra, Australia

Hosted By
Department of Computer Science
University College, The University of New South Wales
Australian Defence Force Academy, Canberra, ACT 2600, Australia

Last modified Thu Sep 21 10:43:43 1995

Extended Title

Merging Symbolic Reasoning with Subsymbolic Tasks to build Intelligent Decision Support Systems

Introduction

Traditional expert systems used deductive reasoning to model expert'(s) knowledge. Such systems have failed to directly use data, do not provide alternative and possibly competing strategies, and perform inadequate argumentation and explanation. We propose the use of induction case based reasoning and neural networks, and an integration of these strategies, to build second generation expert systems.

These issues will be considered with reference to examples in administration, government and law.

Syllabus

• Introduction
• Deductive reasoning and its inadequacies
• A comparison of symbolic and statistical reasoning
• Symbolic reasoning: rule induction
• Rule induction - an example from government
• Symbolic reasoning: case based reasoning - the process and paradigms
• Examples of case based reasoners: HYPO (law), PERSUADER (negotiation), FLORENCE (health)
• Benefits and drawbacks of case based reasoning
• Integrating deductive and case based reasoning: principles and examples (CABARET and IKBALS)
• Architectures for integrated systems - blackboard systems and distributed artificial intelligence
• Subsymbolic approaches - their benefits and inadequacies
• Neural networks - an introduction
• Providing explanation for neural networks
• Integrating deductive reasoning and neural networks - the Split Up system
• Conclusion

Presenter

Dr John Zeleznikow is a senior lecturer in the Department of Computer Science and Computer Engineering and Head of the Database Research Laboratory, Applied Computing Research Institute, La Trobe University. At La Trobe, he supervises a team which includes more than a dozen researchers and has received financial support from the Australian Research Council and industry. He is the author of more than fifty refereed articles and with Dan Hunter has recently written Zeleznikow, J. and Hunter, D., 1994, 'Building Intelligent Legal Information Systems: Knowledge Representation and Reasoning in Law', Kluwer Computer/Law Series, 13 ISBN: 9065448330.

Dr Zeleznikow has taught and conducted research in computer science, information systems, law and mathematics in Australia, France, the Netherlands and the United Kingdom. He was tutorial chairman of the sixteenth International Conference on Very Large Databases and the fourteenth International Conference on Software Engineering. He was chairman and organiser of the IJCAI91 workshop on Integrating Artificial Intelligence and Databases and DS-5, an IFIP workshop on the semantics of interoperable databases. He will be chairman of the sixth International Conference on Artificial Intelligence and Law (1997). In 1994, Dr Zeleznikow presented tutorials at

• Fourth National Conference on Law, Computers and Artificial Intelligence - Exeter, United Kingdom
• Seventh Australian Joint Conference on Artificial Intelligence - Armidale, New South Wales
• Seventh International Conference on Legal Knowledge Based Systems - Amsterdam, the Netherlands
• Fourteenth British Conference on Expert Systems - Cambridge, United Kingdom