Sample-Based Planning and Learning with Function Approximation
Abstract¶
We give a short tutorial on sample-based planning and learning with function approximation for reinforcement learning. The focus is on explaining the standard setups and algorithm design principles. The core algorithmic ideas explained are approximate backwards induction, the (Bellman) eluder dimension and the optimism principle.
Links¶
Knowledge Base
External
Similar Papers¶
1
78%
2
77%
3
76%
4
75%
5
74%