{A page about understanding Deep Learning overall, including some of the big concepts and metaphors. Right now, a notes dump}

Overview Metaphor:

Other Metaphors


People often ask how much linear algebra do I need to know to be able to do deep learning. This is the amount you need. And if you’re not familiar with this, that’s fine. You need to know about matrix products. You don’t need to know a lot about them, you just need to know like computationally what are they and what do they do. You’ve got to be very comfortable with if a matrix of size blah times a matrix of size blah gives a matrix or size blah (i.e. how do the dimensions match up). So if you have 3, and they remember in numpy and PyTorch, we use @ times 3 by 5 gives a vector of size 5. L4