Lecture 1: A Hitchhiker’s Guide to Kernels
Before we teach a machine about the cosmos, let’s agree on what “probably” means. To achieve that, we will first cover the basics of probability distributions, Jacobians, and Bayes’ theorem. After this, we will talk about the ubiquitous kernels, which are how data points decide who their closest friends are. We will end with an application on diffusion maps, using random walks through the data to perform unsupervised phase classification.