Definition
The strategy of dynamically adjusting the learning rate during the training process of a neural network, often involving cyclical or stepwise changes, to improve convergence speed and final performance.
In short: The strategy of dynamically adjusting the learning rate during the training process of a neural network, often involving cyclical or stepwise changes, to improve convergence speed and final perform...