In the previous chapter, I mentioned that everything you know about neural networks is wrong. Here, I repeat that claim. Most literature out there on a neural network starts with a comparison with biological neurones and ends there. This leads readers to often assume that it is. I'd like to make a point that artificial neural networks are nothing like their biological namesake.
Instead, in the last chapter, I spent a significant amount of the chapter describing linear algebra, and explained that the twist is that you can express almost any machine learning (ML) problem as linear algebra. I shall continue to do so in this chapter.
Rather than think of artificial neural networks as analogies of real-life ...
