Definitions for "Activation Function"
is the transform applied to the weighted sum of inputs plus offset for computing the output of a neuron. Also known as the squashing function.
A function by which new output of the basic unit is derived from a combination of the net inputs and the current state of the unit (the total input).
Artificial Neural Networks: A neuron activation function computes the output (also called the activation) of an artificial neuron from a weighted sum of the inputs. A variety of different activation functions are used, including, among others, sigmoid, signum, hyperbolic tangent, sine, Gaussian and inverted Gaussian functions. Neuron activation functions also are commonly referred to as neuron transfer functions.