> Even when looking at something as barebones as perceptrons I was thinking something like "it is trying to approximate a non-linear function" (which is what i

> Even when looking at something as barebones as perceptrons I was thinking something like "it is trying to approximate a non-linear function" (which is what it is in the case of MLPs).