Demystifying Learning Rate Polices for High Accuracy Training of Deep Neural Networks.
Yanzhao WuLing LiuJuhyun BaeKa Ho ChowArun IyengarCalton PuWenqi WeiLei YuQi ZhangPublished in: CoRR (2019)
Keyphrases
- learning rate
- backpropagation algorithm
- training algorithm
- rapid convergence
- high accuracy
- neural network
- multilayer neural networks
- hidden layer
- adaptive learning rate
- feed forward neural networks
- activation function
- training speed
- convergence rate
- feedforward neural networks
- error function
- learning algorithm
- training process
- convergence speed
- back propagation
- fuzzy neural network
- weight vector
- delta bar delta
- multi layer perceptron
- fuzzy logic
- deep architectures
- artificial neural networks
- feed forward
- convergence theorem
- genetic algorithm
- machine learning