Invariance-Preserving Localized Activation Functions for Graph Neural Networks.
Luana RuizFernando GamaAntonio Garcia MarquesAlejandro RibeiroPublished in: IEEE Trans. Signal Process. (2020)
Keyphrases
- activation function
- neural network
- feed forward
- artificial neural networks
- feed forward neural networks
- back propagation
- hidden layer
- hidden neurons
- neural nets
- multilayer perceptron
- network architecture
- connection weights
- neural architecture
- learning rate
- hidden nodes
- feedforward neural networks
- multi layer perceptron
- radial basis function
- fuzzy neural network
- fuzzy logic
- basis functions
- genetic algorithm
- neural network model
- pattern recognition
- input space
- recurrent neural networks
- training phase
- small number