Projection by Convolution: Optimal Sample Complexity for Reinforcement Learning in Continuous-Space MDPs.
Davide MaranAlberto Maria MetelliMatteo PapiniMarcello RestelliPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- sample complexity
- continuous space
- sequential decision problems
- learning problems
- markov decision processes
- dynamic programming
- learning algorithm
- supervised learning
- state space
- theoretical analysis
- optimal policy
- special case
- optimal control
- discrete space
- active learning
- lower bound
- reward function
- partially observable
- upper bound
- training examples
- machine learning
- mathematical morphology
- data sets
- markov decision problems
- learning process