Provably Efficient Algorithm for Nonstationary Low-Rank MDPs.
Yuan ChengJing YangYingbin LiangPublished in: NeurIPS (2023)
Keyphrases
- non stationary
- low rank
- learning algorithm
- dynamic programming
- matrix completion
- convex optimization
- matrix decomposition
- finite horizon
- markov decision processes
- computationally efficient
- input data
- dimensionality reduction
- probabilistic model
- k means
- reinforcement learning
- matrix factorization
- singular values
- pairwise