How to Train a Neural Net?

When you're training a neural net, what you're actually doing is figuring out what weights you want to have on each of the connections between the different neurons in your head. And so there's like an equation that you can put all these inputs into it, and then it spits out either a zero or a one. Well, so the equation itself is fixed, but the weights that are associated with each of the inputs, those are free to move around. So when I'm training my neural net, I'm not actually changing the equations that govern how the neural net works. I'm just changing the weights of the different pixels, or I should say the different inputs.