Semantic Network based appraoche for automatic generalization problem

Abstract

One of the crucial goals for research in artificial intelligence is to understand the nature of human learning and implement learning capabilities in machine. Machine learning is the field of artificial Intelligence that is concerned with developing computational learning theories and constructing learning systems. Human learning is a very complex process and it has different forms (concept learning, word learning, behavior learning, etc.), from which we are interested in concept learning. Most of concept learning researches that have been done in artificial intelligence has consisted of either: using largely analytic techniques to classify inputs or supplying programs with examples and sometimes counter examples of a specified concepts and these programs determine the definition of concepts. Our work falls under the second category. Our purpose is to construct a concept learning algorithm that can learn effectively from limited number of examples using semantic networks. And surpass the limits encountered in the existent approaches.