Building models of learning and expertise with CHREST

Abstract

CHREST (Chunk Hierarchy and REtrieval STructures) is a complete computational architecture implementing processes of learning and perception. CHREST models have successfully simulated human data in a variety of domains, such as the acquisition of syntactic categories, expertise in programming and in chess, concept formation, implicit learning, and the acquisition of multiple representations in physics for problem solving. In this tutorial, we describe the learning, perception and attention mechanisms within CHREST as well as key empirical data captured by CHREST models. Apart from the theoretical material, this tutorial also introduces participants to an implementation of CHREST and its use in a variety of domains. Material and examples are provided so participants can adapt and extend the CHREST architecture.