shapebias

Induction, overhypotheses, and the shape bias: Some arguments and evidence for rational constructivism

Demonstrated that Bayesian hierarchical models can capture how humans "learn to learn" new concepts