A semiparametric statistical approach to model-free policy evaluation.
Tsuyoshi UenoMotoaki KawanabeTakeshi MoriShin-ichi MaedaShin IshiiPublished in: ICML (2008)
Keyphrases
- policy evaluation
- model free
- semi parametric
- statistical inference
- reinforcement learning
- policy iteration
- temporal difference
- least squares
- function approximation
- reinforcement learning algorithms
- monte carlo
- markov decision processes
- statistical methods
- density estimation
- regression model
- neural network
- active learning
- variance reduction