Definition & Use-case
An activation function is a mathematical function applied to the output of each neuron in a layer in order to introduce non-linearity to the model. The combination of the linear transformation in the neural layer and the non-linear activation, enables complex learning capabilities in the network.
