A convergence analysis of Nesterov's accelerated gradient method in training deep linear neural networks.
Xin LiuWei TaoZhisong PanPublished in: Inf. Sci. (2022)
Keyphrases
- convergence analysis
- gradient method
- convergence rate
- neural network
- global convergence
- optimization methods
- step size
- convergence speed
- artificial neural networks
- training set
- optimality conditions
- negative matrix factorization
- semidefinite programming
- genetic algorithm
- learning rate
- linear constraints
- linear svm
- global optimization
- differential evolution
- simulated annealing
- evolutionary algorithm
- primal dual