Learning Procedures via Imitation, Reinforcement and Correction

In procedure learning as elsewhere in cognition, it’s not enough to use the right algorithm, one has to use it in the right way based on the data and context and affordances available.