Building Self Driving Cars Course # 5 — Neural Networks (Part 2)

Welcome to the Self-driving car course part 4.2. In the last section, we had a look at error functions and understood that in order to use gradient decent out predictions should be on a continuous scale rather than a discrete scale. When I say predictions on a discrete scale we usually get predictions as a yes or a no. But in continuous predictions, we get the likelihood of the event which is a value between 0 and 1.

Original post posted at https://theautonomousmachines.com/building-self-driving-cars-course-5-neural-networks-part-2/