College of Computing and Informatics
Research
The goal of this thesis is to develop a psychologically plausible theory of human categorization.
The human mind is adept at finding patterns within the world around it. When some new stimulus is encountered, inferences can quickly be made by categorizing it with similar stimuli in memory, ultimately speeding up the decision-making process. However, this sometimes comes at the cost of accuracy as people tend to inflate intra-category similarities while overestimating inter-category differences. This becomes a particular problem in social categorizations as these imperfect inferences can lead to prejudice and discrimination. To best mitigate these negative effects, it is imperative that we learn more about the cognitive processes underlying categorization and explore their impact on decision making. Therefore, the goal of this thesis is to develop a psychologically plausible theory of human categorization and implement it using a cognitive architecture - a computational framework that incorporates many prominent theories of human cognition.