Neural Networks
Initialization
Forward propogation
Loss Functions
The right initialization is instrumental in getting neural networks to works right and neglecting it was one of the key reasons nerual nets didn’t work well for so many years