ABSTRACT
Categories are crucial for a large number of cognitive activities, such as classification, inference, problem solving, and explanation. They provide an important means for allowing us to benefit from past experiences. Because of this importance and involvement across a wide variety of intelligent activities, category learning has long been a central research topic in cognitive science, cognitive psychology, and machine learning. Most of the research on category learning has focused on classification learning, how to assign items to categories. The focus on classification has led to finding a strong influence of feature diagnosticity, those features that distinguish the categories being learned. There has been a variety of research examining the different ways people learn and use categories, which addresses these limitations of the focus on classification. In the decoding task, learners are able to classify later coded messages on the number relations learned during decoding, even when the relations are fairly abstract.
