On the Global Optimum Convergence of Momentum-based Policy Gradient.
Yuhao DingJunzi ZhangJavad LavaeiPublished in: AISTATS (2022)
Keyphrases
- global optimum
- global convergence
- policy gradient
- faster convergence
- gradient method
- step size
- simulated annealing
- optimization method
- convergence rate
- objective function
- search space
- learning rate
- optimal solution
- actor critic
- convergence speed
- reinforcement learning
- function approximation
- optimization methods
- optimal control
- reinforcement learning algorithms
- least squares
- average reward
- planning problems
- wavelet transform
- multi objective
- neural network