META-Learning State-based Eligibility Traces for More Sample-Efficient Policy Evaluation.
Mingde ZhaoSitao LuanIan PoradaXiao-Wen ChangDoina PrecupPublished in: AAMAS (2020)
Keyphrases
- policy evaluation
- meta learning
- eligibility traces
- least squares
- reinforcement learning
- learning tasks
- temporal difference
- monte carlo
- variance reduction
- function approximation
- model free
- model selection
- markov decision processes
- semi parametric
- policy iteration
- reinforcement learning algorithms
- sample size
- dynamic programming
- learning algorithm
- statistical inference
- state space
- machine learning methods
- decision trees
- knowledge base