A Novel Configurable High-precision and Low-cost Circuit Design of Sigmoid and Tanh Activation Function.
Feng LiuBowen ZhangGang ChenGuoliang GongHuaxiang LuWenchang LiPublished in: ICTA (2021)
Keyphrases
- activation function
- high precision
- circuit design
- low cost
- neural network
- high recall
- feed forward
- hidden layer
- artificial neural networks
- feed forward neural networks
- back propagation
- network architecture
- basis functions
- neural nets
- learning rate
- radial basis function
- sigmoid function
- chaotic neural network
- multilayer perceptron
- hidden nodes
- training phase
- feedforward neural networks
- input space
- real time
- high accuracy
- fuzzy neural network
- extreme learning machine
- achieve high precision
- recurrent neural networks
- learning tasks
- decision making
- hardware implementation
- linear combination
- bayesian networks
- learning algorithm
- genetic algorithm