A Tale of Two-Timescale Reinforcement Learning with the Tightest Finite-Time Bound.
Gal DalalBalázs SzörényiGugan ThoppePublished in: AAAI (2020)
Keyphrases
- reinforcement learning
- state and action spaces
- upper bound
- function approximation
- lower bound
- worst case
- model free
- state space
- reinforcement learning algorithms
- markov decision processes
- machine learning
- real valued functions
- multi agent
- dynamic programming
- temporal difference
- data sets
- temporal difference learning
- action space
- learning algorithm
- optimal policy
- supervised learning
- finite number
- action selection
- real time
- finite automata
- least squares
- reinforcement learning methods
- decision trees
- support vector