Stationary Points of Shallow Neural Networks with Quadratic Activation Function.
David GamarnikEren C. KizildagIlias ZadikPublished in: CoRR (2019)
Keyphrases
- activation function
- neural network
- stationary points
- objective function
- hidden layer
- feed forward neural networks
- feed forward
- artificial neural networks
- back propagation
- neural nets
- fixed point
- multilayer perceptron
- constrained optimization
- nonlinear programming
- convergence analysis
- mathematical programming
- learning rate
- basis functions
- feedforward neural networks
- network architecture
- fuzzy neural network
- training phase
- extreme learning machine
- linear programming
- recurrent neural networks
- multi objective
- fault diagnosis
- neural network model
- radial basis function
- multi layer perceptron
- fuzzy rules
- machine learning
- genetic algorithm
- bayesian networks
- feature space
- evolutionary algorithm