en-de
en-es
en-fr
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
Reinforcement Learning Theory
Published on Feb 25, 200713378 Views
The tutorial is on several new pieces of Reinforcement learning theory developed in the last 7 years. This includes:\\ 1. Sample based analysis of RL including E3 and sparse sampling.\\ 2. Generalizat
Related categories
Chapter list
Modern Reinforcement Learning Theory00:05
Reinforcement Learning is Always Relevant00:22
The answer to: \"Is this an RL problem?\" is always \"yes\"02:47
Outline06:22
What is a sample complexity guarantee?07:08
The E3 guarantee12:04
E3 Theorem Statement17:23
The Known(h) MDP29:03
The Known(h) MDP 0131:10
The Known(h) MDP 0232:07
The Known(h) MDP 0332:17
The Unknown(h) MDP 33:05
Dynamic Program36:01
E3(h) Explicit Explore or Exploit Algorithm41:46
The proof uses (5!) MDPs46:04
Proof Sketch:48:05
R-Max(h) Modification55:48
Delayed Q-learning01:01:00
Outline01:04:35
The Limits of Sample Complexity: A lower bound01:04:57
Proof01:05:36
Implications01:08:43
Attempt 1: Factored - E301:09:35
Attempt 2: Metric - E301:13:57
Do we really want the guarantee these algorithms provide?01:19:21