Previous research has explored the equivalence between various theories of human category learning and probability density estimation. At the same time, there has been rapidly growing evidence of important underlying connections between (a) causal learning and beliefs; and (b) human category learning and structure. In this talk, I will bring these two strands together by showing how to represent human category learning and inference in the framework of probabilistic graphical models. Time permitting, I will also present several applications of these "re-representations" to empirical datasets.
Copyright. All Rights Reserved.