Convergence rates for gradient descent in the training of overparameterized artificial neural networks with biases.
Arnulf JentzenTimo KrögerPublished in: CoRR (2021)
Keyphrases
- convergence rate
- artificial neural networks
- conjugate gradient
- training algorithm
- feed forward neural networks
- learning rate
- back propagation
- benchmark classification problems
- levenberg marquardt
- step size
- evolutionary artificial neural networks
- convergence speed
- neural network
- multi layer perceptron
- objective function
- primal dual
- gaussian kernels
- feedforward artificial neural networks
- genetic algorithm
- mutation operator
- cost function
- training process
- global convergence
- stochastic gradient descent
- genetic algorithm ga
- training set
- hidden layer
- neural network model
- dynamic programming
- reinforcement learning