The program simple_nn.rs feed binary numbers into a neural net with two hidden layers of 8 neurons each and train the network to output the correct decimal number using a 7-segment display.
In the note Back propagtion the reader is walked through the theory of an ultra simple neural net in great detail.
The same ultra simple net is simulated with numbers in this spread sheet
The aim is to bridge the gab for absolute beginners, who knows simple calculus, but are new to machine learning.
Reading the note and playing with the spread sheet hopefully can help gain familiarity with simple neural nets.