Normalisation is the technique in ML to make input values or input values to a hidden layer closer so they have similar meanings, and also it’s required for the network to learn because a network can’t learn infinite range, ie. without normalisation.
