AEDR: An Adaptive Mechanism to Achieve Online Learning Rate Dynamically.
Jinjing ZhangFei HuLi LiXiaofei XuZhanbo YangPublished in: ICTAI (2017)
Keyphrases
- learning rate
- convergence rate
- learning algorithm
- update rule
- multilayer neural networks
- error function
- hidden layer
- rapid convergence
- online learning
- delta bar delta
- bp neural network algorithm
- adaptive learning rate
- activation function
- weight vector
- convergence speed
- natural gradient
- middle layer
- step size
- feed forward
- kernel function
- multi class
- pairwise
- machine learning
- neural network