Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning.
Ronald J. WilliamsPublished in: Mach. Learn. (1992)
Keyphrases
- reinforcement learning
- learning algorithm
- computational cost
- optimization problems
- computationally efficient
- neural network
- statistical approaches
- image segmentation
- expert systems
- black box
- statistical methods
- combinatorial optimization
- statistical measures
- model free
- convergence rate
- information theoretic
- benchmark datasets
- theoretical analysis
- dynamic programming
- computational complexity
- clustering algorithm
- machine learning