0 alternative definitions
A technique used in training neural networks to prevent overfitting by randomly deactivating a subset of neurons during each training iteration.
