Abstract

I will show how some of the technical themes introduced in the last two days -- hierarchical Bayesian models, nonparametric Bayes, and Bayesian semi-supervised learning -- can be used to build models of core aspects of human inductive learning, including word learning and property induction.