The learning rate is a small number, usually denoted as which is used to decrease / control the “nudge” / adjustment of the parameters of a machine learning model.
It is required because making too big adjustments might cause you to over-shoot your target. Making it too small however on the other hand will cause the model to converge too slowly.
It is one of many Hyper-Parameters in an Optimizer