Supervised Competitive Learning Part II: SCL with Fuzzy Logic

Authors

Ce Wang, Washington University in St Louis
Takayuki D. Kimura, Washington University in St Louis
Thomas H. Fuller Jr., Washington University in St Louis

Document Type

Technical Report

Publication Date

1993

Filename

WUCS-93-44.pdf

Technical Report Number

WUCS-93-44

Abstract

Supervised Competitive Learning (SCL) is described in an accompanying paper [1]; SCL assembles a set of learning modules into a supervised learning system to address the stability-plasticity dilemma. That paper reported results using backpropagation networds as the learning modules (SCL/BP). Here (Part II) we report SCL results using learning modules based on fuzzy logic (SCL/FZ). Although its learning algorithm is very different from that of backpropagation networks, fuzzy logic also suffers the stability-plasticity dilemma. A simulator on handwritten digit and gesture recognition was constructed to demonstrate the utility of SCL/FZ; it is recognized 98% of test digits, and 91% of test gestures. In this paper, we also compare SCL/BP with SCL/FZ for recognizing handwritten digits.

Comments

Permanent URL: http://dx.doi.org/10.7936/K7GF0RTX

Recommended Citation

Wang, Ce; Kimura, Takayuki D.; and Fuller, Thomas H. Jr., "Supervised Competitive Learning Part II: SCL with Fuzzy Logic" Report Number: WUCS-93-44 (1993). All Computer Science and Engineering Research.
https://openscholarship.wustl.edu/cse_research/539