Convergence Rates of a Momentum Algorithm with Bounded Adaptive Step Size for Nonconvex Optimization.
Anas BarakatPascal BianchiPublished in: ACML (2020)
Keyphrases
- convergence rate
- step size
- variable step size
- learning rate
- global convergence
- faster convergence
- cost function
- convergence speed
- steepest descent method
- primal dual
- line search
- learning algorithm
- optimization algorithm
- gradient method
- evolutionary programming
- levenberg marquardt
- objective function
- global optimization
- stochastic gradient descent
- search space
- optimization method
- stopping criterion
- least mean square
- simulated annealing
- adaptive filter
- computational complexity
- blind source separation
- global optimum
- approximation algorithms
- optimization problems