create material to explain neural networks

[stevehadd]

explain back propogration highlight key hyperparameters from networtk architecture (feed forward) and back propogation training show neural networks in tensorflow to dig down into structure a bit show classification output vs regression output show construction of different architectures

• feed forward
• convolutional
• LSTM (simple example? maybe using rotors as a time series?)
• Graph NN

[stevehadd]

this has mostly been done now. Just need to finish the recurrent nerual network example, which is covered by #11 .