Is Reinforcement Learning More Difficult Than Bandits? A Near-optimal Algorithm Escaping the Curse of Horizon.
Zihan ZhangXiangyang JiSimon S. DuPublished in: COLT (2021)
Keyphrases
- learning algorithm
- k means
- experimental evaluation
- reinforcement learning
- times faster
- detection algorithm
- high accuracy
- dynamic programming
- cost function
- expectation maximization
- optimization algorithm
- recognition algorithm
- computational cost
- high dimensional
- preprocessing
- optimal solution
- model free
- stochastic approximation
- reinforcement learning algorithms
- optimal or near optimal
- hybrid algorithm
- ant colony optimization
- segmentation algorithm
- linear programming
- simulated annealing