Invariance-Preserving Localized Activation Functions for Graph Neural Networks.
Luana RuizFernando GamaAntonio G. MarquesAlejandro RibeiroPublished in: CoRR (2019)
Keyphrases
- activation function
- neural network
- feed forward
- artificial neural networks
- back propagation
- feed forward neural networks
- hidden layer
- neural architecture
- neural nets
- hidden neurons
- connection weights
- multilayer perceptron
- feedforward neural networks
- radial basis function
- multi layer perceptron
- pattern recognition
- fuzzy logic
- fuzzy neural network
- hidden nodes
- basis functions
- multi layer
- feature space
- knn
- network architecture
- input space
- linear combination
- genetic algorithm