Convergence rate of inertial Forward-Backward algorithm beyond Nesterov's rule.
Vassilis ApidopoulosJean-François AujolCharles DossalPublished in: Math. Program. (2020)
Keyphrases
- convergence rate
- forward backward
- convergence speed
- step size
- global convergence
- cost function
- dynamic programming
- variable step size
- semidefinite programming
- primal dual
- learning rate
- k means
- hybrid algorithm
- kalman filter
- premature convergence
- numerical stability
- learning algorithm
- convex hull
- hidden markov models
- significant improvement
- mutation operator
- search space
- levenberg marquardt
- objective function
- simplex method
- reinforcement learning
- faster convergence rate