Linear

The linear activation function simply returns its input without applying any modification:

$$g(x) = x$$

Its derivative is then given by:

$$\frac{\partial g}{\partial x} = 1$$

This activation function is used for instance in the last layer of a regression network.