Improved Overparametrization Bounds for Global Convergence of Stochastic Gradient Descent for Shallow Neural Networks.
Bartlomiej PolaczykJacek CyrankaPublished in: CoRR (2022)
Keyphrases
- global convergence
- stochastic gradient descent
- neural network
- step size
- convergence rate
- global optimum
- convergence speed
- optimization methods
- least squares
- loss function
- matrix factorization
- differential evolution
- online algorithms
- genetic algorithm
- artificial neural networks
- random forests
- worst case
- image restoration
- evolutionary algorithm
- learning rate
- feature space
- regularization parameter
- multiple kernel learning
- weight vector
- feature extraction