7. Neural Networks Part II#

 
We just saw what neural networks are made of, what are the math done in a node and how to compute all values of activation units, leading to intermediary predictions.

Let’s now cover the backpropagation algorithm, which is the core of how neural networks are trained.

  • What are the cost functions for neural network?

  • How should the weights be initialized?

  • How does backpropagation work?