Challenge : Write less than 100 word paragraph on any 5 important topics related to Deep Learning Models
Activation functions are introduced in the neural network to capture non-linearities in the input data. It converts the weighted sum of a node's input with an addition of bias, to the node's output, eventually providing an advantage to the network on controlling output of the nodes, compared to a network without activation function which essentially works as linear regression model.
Some most used activation functions are: