Introduction
Birds inspired aeroplanes, burdock burrs inspired Velcro, and the brain inspired artificial neural networks (ANNs). An ANN is a machine learning model made of many small, simple units — artificial neurons — connected together. Each unit does something very simple, but millions of them working together can recognise faces, translate languages, drive cars and write text.
ANNs are the heart of deep learning: "deep" simply means networks with many layers of neurons.
Just as aeroplanes don't flap their wings, modern ANNs aren't faithful copies of the brain. The brain was the starting idea; since then, the field has followed whatever works best mathematically. In this topic we look at that starting idea, the bumpy history of the field, and why deep learning exploded in the 2010s.