Bounded activation functions for enhanced training stability of deep neural networks on visual pattern recognition problems.
Shan Sung LiewMohamed Khalil HaniRabia BakhteriPublished in: Neurocomputing (2016)
Keyphrases
- activation function
- pattern recognition problems
- neural network
- feed forward neural networks
- backpropagation algorithm
- feedforward neural networks
- pattern recognition
- continuous functions
- training phase
- back propagation
- hidden layer
- multi layer perceptron
- artificial neural networks
- feed forward
- neural nets
- multilayer perceptron
- connection weights
- radial basis function
- neural architecture
- learning rate
- training process
- combination of multiple classifiers
- network architecture
- multi layer
- fuzzy neural network
- basis functions
- training examples
- fuzzy logic
- bp neural network
- supervised learning
- asymptotic stability
- recurrent neural networks
- neural network model
- genetic algorithm