Fine-Grained Gap-Dependent Bounds for Tabular MDPs via Adaptive Multi-Step Bootstrap.
Haike XuTengyu MaSimon S. DuPublished in: CoRR (2021)
Keyphrases
- fine grained
- multi step
- coarse grained
- markov decision processes
- lower bounding
- single step
- access control
- state space
- upper bound
- contingency tables
- tightly coupled
- lower bound
- neural network
- data lineage
- k nearest neighbor
- optimal policy
- natural language processing
- semi supervised
- feature selection
- learning algorithm