A Cognitive Modeling Approach to Learning of Ill-defined Categories

Mukesh Rohatgi

A vast majority of concepts learned by human beings arc ill-defined. Such concepts elude precise definitions and are represented by nonverbalizable descriptions. Literature from cognitive psychology provides an insight into processes used by human beings for learning ill-defined categories. This paper describes the implementation of an Adaptive Concept Learning (ACL) system designed learn ill-defined categories by simulating human learning behavior. The contents of this paper stand in contrast to the extant concept learning systems described in machine learning literature. The extant systems are generally designed to learn well defined concepts that can be represented with the help of verbalizable rules. In contrast, the system described in this paper is designed to learn ill-defined categories. Therefore, the implementation of ACL is a more realistic simulation of concept learning bel]avior observed in human beings.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.