Provable convergence of Nesterov's accelerated gradient method for over-parameterized neural networks.
Xin LiuZhisong PanWei TaoPublished in: Knowl. Based Syst. (2022)
Keyphrases
- gradient method
- convergence rate
- neural network
- step size
- convergence speed
- log likelihood function
- learning rate
- pattern recognition
- genetic algorithm
- convex formulation
- optimization methods
- semidefinite programming
- artificial neural networks
- back propagation
- negative matrix factorization
- genetic algorithm ga
- feature selection
- natural gradient learning