Demystifying AI

Welcome to This Explainer. If you’ve ever looked at artificial intelligence and thought, “Wow, that’s just a total impenetrable black box,” well, you are absolutely in the right place. Today, we are gonna pop the hood and see exactly how this thing actually works. We’re going from the super high-level concepts of machine learning all the way down to the actual math and code that makes it all happen. So let’s kick things off with a pretty fundamental question: When an AI algorithm spits out some startlingly precise prediction or suddenly recognizes a hidden pattern in massive amounts of data, I mean, it kind of feels like magic, right? Like some sort of computational alchemy. But, spoiler alert for today, it is entirely not magic. We’re peeling back the layers to show you that underneath all the hype, it’s really just a rigorous, hard mathematical framework. No wands required! To get there, we’ll be following a quick 4-step roadmap today. First, the big machine learning paradigms. Then, the mathematical engine running them. Followed by optimization and learning, and finally, we’ll look at building AI with PyTorch versus TensorFlow.