Probabilistic learning rate scheduler with provable convergence.
Dahlia DevapriyaThulasi TholetiJanani SureshSheetal KalyaniPublished in: CoRR (2024)
Keyphrases
- learning rate
- convergence rate
- rapid convergence
- convergence theorem
- convergence speed
- adaptive learning rate
- update rule
- learning algorithm
- error function
- weight update
- hidden layer
- global convergence
- weight vector
- delta bar delta
- multilayer neural networks
- step size
- natural gradient
- uniform convergence
- line search
- activation function
- training algorithm
- feature selection