Modeling Human-like Concept Learning with Bayesian Inference over Natural Language

06/05/2023

∙

by   Kevin Ellis, et al.

∙

0

∙

share

We model learning of abstract symbolic concepts by performing Bayesian inference over utterances in natural language. For efficient inference, we use a large language model as a proposal distribution. We fit a prior to human data to better model human learners, and evaluate on both generative and logical concepts.