Widely Linear Kernels for Complex-valued Kernel Activation Functions.
Simone ScardapaneSteven Van VaerenberghDanilo ComminielloAurelio UnciniPublished in: ICASSP (2019)
Keyphrases
- complex valued
- activation function
- kernel function
- kernel methods
- real valued
- neural network
- artificial neural networks
- back propagation
- feature space
- input space
- feed forward
- support vector
- neural nets
- hidden layer
- recurrent neural networks
- learning rate
- multilayer perceptron
- multi valued
- network architecture
- high dimensional
- basis functions
- support vector machine
- radial basis function
- linear equations
- bp neural network
- feature extraction