I am always facinated by machine learning algorithms or math/physics theorems that seem nonsensical - but proven true. It is a very gratifying to understand the details of an algorithm, but the process is often time-consuming. The authors of an algorithm frequently assumed prior knowledge, and omitted explanations for intermediary steps. An interested reader needs to look up multiple resources to understand the full logic.
Inspired by https://github.com/alex/what-happens-when, I would like to start my own journey of explaining (almost) every step of a few machine learning algorithms. Hopefully others who go through the repository will feel some gaps in their knowledge are filled.