Constrained RTRL To Reduce Learning Rate and Forgetting Phenomenon.
Fabrice DruauxEmmanuel RogueAlain FaurePublished in: Neural Process. Lett. (1998)
Keyphrases
- learning rate
- convergence rate
- learning algorithm
- error function
- rapid convergence
- hidden layer
- weight vector
- convergence speed
- delta bar delta
- multilayer neural networks
- adaptive learning rate
- training algorithm
- convergence theorem
- activation function
- particle swarm optimization
- linear programming
- reinforcement learning
- neural network