We're going to talk about how neurons in a neural network learn by getting their math adjusted, called backpropagation, and how we can optimize networks by finding the best combinations of weights to minimize error. [12:08]