A Heuristic Reinforcement Learning Based on State Backtracking Method.
Min FangHao LiXiaosong ZhangPublished in: Web Intelligence (2012)
Keyphrases
- dynamic programming
- reinforcement learning
- preprocessing
- cost function
- high precision
- classification method
- combinatorial optimization
- detection method
- significant improvement
- experimental evaluation
- support vector machine
- high accuracy
- clustering method
- segmentation method
- state space
- constraint satisfaction
- beam search
- synthetic data
- neural network
- computational cost
- pairwise
- computational complexity
- image sequences