A Translator for the Conversion of Graphical Geometric Figures to First-Order Logic Description

摘要

The Abstractional Concept Mapping Theory is a computational model of analogical reasoning allowing for the explicit representation of arbitrary comparisons of similarity or analogy between two given descriptions in first-order logic. The work presented here is a first step in the implementation of this theoretical model. Given that geometric analogies involve a relatively simple knowledge domain, we have chosen to first test the analogy model on four-part geometric analogy problems. In order to facilitate the testing of such analogies, we present here the implementation of a heuristically-based automatic translator for the translation of groups of geometric figures into first-order logic description. Such descriptions can then serve as input to the analogy model for testing.