Low-rank MDPs with Continuous Action Spaces.
Miruna OprescuAndrew BennettNathan KallusPublished in: AISTATS (2024)
Keyphrases
- action space
- low rank
- markov decision processes
- state space
- state and action spaces
- reinforcement learning
- linear combination
- continuous state spaces
- missing data
- convex optimization
- real valued
- matrix factorization
- singular value decomposition
- high order
- low rank matrix
- kernel matrix
- continuous state
- markov decision process
- semi supervised
- rank minimization
- high dimensional data
- action selection
- markov decision problems
- policy iteration
- continuous action
- stochastic processes
- dynamic programming
- single agent
- optimal policy
- function approximators
- average reward
- average cost
- markov chain
- least squares
- recommender systems
- high dimensional