Computationally Efficient PAC RL in POMDPs with Latent Determinism and Conditional Embeddings.
Masatoshi UeharaAyush SekhariJason D. LeeNathan KallusWen SunPublished in: CoRR (2022)
Keyphrases
- computationally efficient
- reinforcement learning
- partially observable markov decision processes
- markov decision processes
- optimal policy
- state space
- continuous state
- latent space
- function approximation
- partially observable
- vector space
- finite state
- policy gradient
- policy search
- sample complexity
- reinforcement learning algorithms
- learning algorithm
- sample size
- latent variables
- multi agent
- low dimensional
- dimensionality reduction
- markov decision process
- upper bound
- transfer learning
- machine learning
- reinforcement learning methods
- policy iteration
- supervised learning
- model free
- pac learning
- temporal difference
- dynamic programming
- markov chain
- manifold learning
- probabilistic model
- action selection
- action space
- computational complexity
- actor critic
- dynamical systems