A survey on modern trainable activation functions.
Andrea ApicellaFrancesco DonnarummaFrancesco IsgròRoberto PrevetePublished in: CoRR (2020)
Keyphrases
- activation function
- neural network
- feed forward
- artificial neural networks
- hidden layer
- back propagation
- complex valued
- feed forward neural networks
- neural nets
- neural architecture
- learning rate
- basis functions
- hidden nodes
- network architecture
- radial basis function
- multilayer perceptron
- continuous functions
- autoregressive
- fuzzy neural network
- fuzzy logic
- genetic algorithm
- input space
- recurrent neural networks
- prior knowledge