Why activation function is needed in Neural Networks???
Today allow me to share
with you, the relationship between a spider and an activation
function
Activation function is
used in Neural network to introduce the non-linearity.
Without the proper
activation function, neural network will simply be large Linear Model.
Consider the case
where you have a data which is linearly separable
Constructing a
model to separate yellow and orange data points is very simple. A simple
logistic regression will do this job very perfectly
Now consider the
case where your data set looks like something shown below
What do you think now??
will a simple linear model be able to carve out the hyperplane in such a way to
classify both yellow and orange points.
I don’t think so. And it
is the fact that a simple linear model may not be able to find complex patterns
in a data.
So now we go to
big brother of linear model -Neural network.
Neural Network has this
super ability to both carve out linear dependency as well as nonlinear
dependency as per dataset.
How I perceive
activation function is some special kind of function which gives super power to
the simple linear model which helps model to capture complex, nonlinear
patterns.
As discussed earlier
neural network with no activation function or simply identity activation function
is just a large Linear Model.
Identity Activation
Function
Some famous activation functions used in NEural NEtorks are
1. Sigmoid Function
2. Rectified Linear Unit (ReLU)
3. Leaky ReLU
Comments
Post a Comment