Linear Stochastic Approximation: Constant Step-Size and Iterate Averaging.
Chandrashekar LakshminarayananCsaba SzepesváriPublished in: CoRR (2017)
Keyphrases
- step size
- stochastic approximation
- steepest descent method
- convergence rate
- cost function
- convergence speed
- monte carlo
- policy iteration
- temporal difference
- variable step size
- learning rate
- wavelet coefficients
- feature vectors
- feature space
- image compression
- temporal difference learning
- reinforcement learning
- multiscale