Learning Tasks
Modeling state-action transitions
Learning a policy
A policy determines the action from a given state
Learning state values
How "good" is each state?
Interacts strongly with policy
(next)