Learning rate
The learning rate is a hyperparameter in machine learning algorithms that controls the step size at each iteration while moving toward a minimum of a loss function. It plays a critical role in model training by influencing the speed and quality of convergence.