How to train your d̶r̶a̶g̶o̶Neural Net (Backpropagation Intiution)

Before a neural network can do wonders it needs to be trained. Backpropagation an integral part of training of the network often goes un-understood.

Courtesy of Dreamworks Animation

With a myriad of tools available to train a neural net most people tend to skip understanding the intuition behind backpropagation. Understandably so, when the mathematics looks like this.

Andrew Ng’s Coursera course: https://www.coursera.org/learn/machine-learning

Why do we need to train? (The Not BodyBuilding Answer)

Before trying to understand Backpropagation intuitively we need to look at why do we need to train a network at all! As we saw in the Part 1 of this article a neural network is a Multi Layer Perceptron with Non-Linear Activation functions.

Neural Networks - An Intuition - Data Driven Investor - Medium

The signal input at the first layer of such a network propagates ahead into the network through weights that control the connection strength between neurons of adjacent…