Kafnets: kernel-based non-parametric activation functions for neural networks.
Simone ScardapaneSteven Van VaerenberghAurelio UnciniPublished in: CoRR (2017)
Keyphrases
- input data
- activation function
- neural network
- input space
- artificial neural networks
- back propagation
- hidden layer
- feed forward
- feed forward neural networks
- hidden neurons
- neural nets
- multilayer perceptron
- neural architecture
- network architecture
- connection weights
- feedforward neural networks
- radial basis function
- parametric models
- hidden nodes
- learning rate
- pattern recognition
- multi layer perceptron
- genetic algorithm
- kernel methods
- basis functions
- training phase
- fuzzy neural network
- support vector machine
- training data
- neural network model
- multi layer
- rbf neural network
- fuzzy logic
- prediction model
- expert systems
- global exponential stability