Hardware implementation of evolvable block-based neural networks utilizing a cost efficient sigmoid-like activation function.
Vishnu P. NambiarMohamed Khalil HaniRiadh SahnounMuhammad N. MarsonoPublished in: Neurocomputing (2014)
Keyphrases
- activation function
- hardware implementation
- cost efficient
- neural network
- feed forward neural networks
- efficient implementation
- signal processing
- feed forward
- artificial neural networks
- hidden neurons
- back propagation
- hidden layer
- multilayer perceptron
- neural nets
- single layer
- feedforward neural networks
- backpropagation algorithm
- hidden nodes
- general purpose processors
- sigmoid function
- image processing algorithms
- field programmable gate array
- network architecture
- radial basis function
- extreme learning machine
- learning rate
- basis functions
- hidden units
- trained neural network
- data fusion
- rough sets