Reinforcement learning can be formalized terms of ___ in which the agent initially only knows the set of possible ___ and the set of possible actions
1
Markov decision processes, objects
2
Hidden states, objects
3
Markov decision processes, states
4
Objects, states