The Sensitivity of PBIL to Its Learning Rate, and How Detailed Balance Can Remove It.
Jonathan L. ShapiroPublished in: FOGA (2002)
Keyphrases
- learning rate
- convergence rate
- learning algorithm
- evolutionary algorithm
- hidden layer
- error function
- adaptive learning rate
- convergence speed
- multilayer neural networks
- activation function
- training algorithm
- delta bar delta
- population based incremental learning
- rapid convergence
- sensitivity analysis
- step size
- weight vector
- multi objective
- real world
- global optimum
- hybrid algorithm
- global optimization
- convergence theorem
- linear combination