Evolving parsimonious networks by mixing activation functions.

Alexander Hagg Maximilian Mensing Alexander Asteroth

Published in: GECCO (2017)

Keyphrases

activation function
network size
neural network
artificial neural networks
feed forward
complex valued
hidden layer
neural architecture
back propagation
feed forward neural networks
neural nets
social networks
basis functions
multilayer perceptron
radial basis function
learning rate