LOSSGRAD: automatic learning rate in gradient descent.
Bartosz WójcikLukasz MaziarkaJacek TaborPublished in: CoRR (2019)
Keyphrases
- learning rate
- error function
- update rule
- convergence rate
- learning algorithm
- natural gradient
- hidden layer
- convergence speed
- adaptive learning rate
- multilayer neural networks
- rapid convergence
- conjugate gradient
- machine learning
- weight vector
- activation function
- cost function
- loss function
- objective function
- data mining
- delta bar delta