Reinforcement Learning Crash Course - Eligibility Traces & Function Approximation
Reinforcement Learning Crash Course by Viviane Clay
0:00:00 Averaging n-step Returns (lambda return)
0:01:40 Recap: n-step return & discount factors
0:04:04 Lambda Exercise
0:05:08 Exercise Solution
0:06:00 Forward View (Lambda Return Algorithm)
0:07:30 Backward View (TD Lambda)
0:08:13 Excursion: Value Function Approximation (DRL)
0:13:25 Eligibility Trace
0:16:08 TD Lambda Pseudocode
0:19:26 Eligibility Traces Illustrated
0:22:08 Properties of TD Lambda