Andrew Ng’s two series on NN and RNN respectively.

Start from a logistic regression on a small training sample to grasp NN knowing the below loss and cost functions: