1 Preface
The backpropagation is the mathematical process by which the gradients of loss function with respect to weights of all layers of a dense network were computed analytically.
The previous writeup on this was a very brief presentation. Thus, this indepth derivation was made from the class-notes of Graphical Models and Deep Learning by Dr. Sumitra.
This writeup was made as a chapter of the lecture on neural networks, the link is here.