(Colourful intuitive diagram) Easiest way to quickly grasp artificial neural network back propagation

[thoughtfuhk]

 

(High quality version: https://i.imgur.com/iQEDfXq.png)

"Giant intuitive diagram showing how an artificial neural network works; by allowing error signals (aka changes in costs) to “trickle” backwards from the output layer.

For example: “Trickling” backwards simply means a value (aka cost) computed at some neuron \(j\) in layer \(L\) is input to values generated by computations of some prior neuron \(k\) in layer \(L-1\). Some cost computed from neuron \(k\) then becomes input to values generated by computations of some prior neuron \(m\), in layer \(L-2\)."

 

Links:

Amazon: 
https://www.amazon.com/dp/B077FX57ZZ

Free copy with equations that are nicely coloured differently than their surrounding text content (instead of equations with the same colouring as their surrounding text content)…on research gate :
https://www.researchgate.net/publication/321162382_Artificial_Neural_Nets_For_Kids

Free copy on quora: 
https://www.quora.com/What-is-the-most-intuitive-explanation-of-artificial-neural-networks/answer/Jordan-Bennett-9

 

Edited November 24, 2017 by thoughtfuhk