Improving the Integration of Neuro-Symbolic Rules with Case-Based Reasoning

Authors:
Jim Prentzas;Ioannis Hatzilygeroudis;Othon Michail
Affiliations:
School of Engineering Dept of Computer Engin. & Informatics, University of Patras, Patras, Greece 26500 and Dept of Informatics & Computer Technology, TEI of Lamia, Lamia, Greece 35100;School of Engineering Dept of Computer Engin. & Informatics, University of Patras, Patras, Greece 26500;School of Engineering Dept of Computer Engin. & Informatics, University of Patras, Patras, Greece 26500
Venue:
SETN '08 Proceedings of the 5th Hellenic conference on Artificial Intelligence: Theories, Models and Applications
Year:
2008

Citing 4
Cited 0

Case-based reasoning integrations

AI Magazine
KBS Maintenance as Learning Two-Tiered Domain Representation

ICCBR '95 Proceedings of the First International Conference on Case-Based Reasoning Research and Development
An Efficient Hybrid Rule Based Inference Engine with Explanation Capability

Proceedings of the Fourteenth International Florida Artificial Intelligence Research Society Conference
Integrating (rules, neural networks) and cases for knowledge representation and reasoning in expert systems

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we present an improved approach integrating rules, neural networks and cases, compared to a previous one. The main approach integrates neurules and cases. Neurules are a kind of integrated rules that combine a symbolic (production rules) and a connectionist (adaline unit) representation. Each neurule is represented as an adaline unit. The main characteristics of neurules are that they improve the performance of symbolic rules and, in contrast to other hybrid neuro-symbolic approaches, they retain the modularity of production rules and their naturalness in a large degree. In the improved approach, various types of indices are assigned to cases according to different roles they play in neurule-based reasoning, instead of one. Thus, an enhanced knowledge representation scheme is derived resulting in accuracy improvement. Experimental results demonstrate its effectiveness.