Adaptive Blending Units: Trainable Activation Functions for Deep Neural Networks.
Leon René SütfeldFlemming BriegerHolger FingerSonja FüllhaseGordon PipaPublished in: CoRR (2018)
Keyphrases
- activation function
- neural network
- artificial neural networks
- hidden units
- feed forward neural networks
- feed forward
- hidden layer
- back propagation
- neural nets
- hidden neurons
- complex valued
- neural architecture
- multilayer perceptron
- network architecture
- connection weights
- radial basis function
- basis functions
- multi layer perceptron
- rbf neural network
- feedforward neural networks
- learning rate
- neural network model
- hidden nodes
- multi layer
- pattern recognition
- rbf network
- fuzzy neural network
- recurrent neural networks
- small number
- prior knowledge
- expert systems
- training data
- feature extraction