SAU: Smooth Activation Function Using Convolution with Approximate Identities.
Koushik BiswasSandeep KumarShilpak BanerjeeAshish Kumar PandeyPublished in: ECCV (21) (2022)
Keyphrases
- activation function
- neural network
- artificial neural networks
- hidden layer
- feed forward
- back propagation
- feed forward neural networks
- network size
- learning rate
- multilayer perceptron
- neural nets
- feedforward neural networks
- radial basis function
- sigmoid function
- network architecture
- chaotic neural network
- basis functions
- hidden nodes
- rbf neural network
- control system
- machine learning