Automatic learning rate optimization by higher-order derivatives.
Xiao-Hu YuLi-Qun XuPublished in: ICNN (1997)
Keyphrases
- learning rate
- higher order
- convergence rate
- error function
- learning algorithm
- convergence speed
- rapid convergence
- adaptive learning rate
- markov random field
- hidden layer
- optimization method
- optimization algorithm
- multilayer neural networks
- pairwise
- activation function
- global optimization
- evolution strategy
- weight vector
- global convergence
- delta bar delta
- optimization methods
- machine learning
- constrained optimization
- training algorithm