Wide-minima Density Hypothesis and the Explore-Exploit Learning Rate Schedule.
Nikhil IyerV. ThejasNipun KwatraRamachandran RamjeeMuthian SivathanuPublished in: CoRR (2020)
Keyphrases
- learning rate
- convergence rate
- learning algorithm
- error function
- multilayer neural networks
- scheduling problem
- rapid convergence
- activation function
- adaptive learning rate
- hidden layer
- natural gradient
- training algorithm
- convergence speed
- convergence theorem
- weight vector
- machine learning
- high accuracy
- multi class
- evolutionary algorithm
- feature selection
- neural network
- delta bar delta