Statistical Machine Learning Almanac

The Statistical ML Almanac is where my more technical writing lives: synthesized course notes, reference pages, and the occasional standalone concept post. It is not a blog to read in order — think of it as a cheat sheet with proofs. Pages get revised as I learn more, so nothing here should be considered finished.

Right now there are three shelves:

STAT 241A: Theoretical Statistics — write-ups of Jason Lee’s Fall 2026 lectures, one post per lecture, titled by main idea rather than date. Start with Learning as Optimization: the Excess Risk Decomposition, which frames everything the course is about.

STAT 210B: Results Toolbox — individual results from Nikita Zhivotovskiy’s STAT 210B, stated the way they appear in my notes, each with the one proof step worth remembering. These exist to be linked from other posts: when a 241A page says “by Bernstein,” the toolbox has the exact statement being invoked.

Concepts — standalone background posts, like PCA, that the course notes lean on but that stand on their own.

Everything is in the sidebar. If a page cites a source (Bach, Tengyu Ma’s CS229M notes), the attribution is at the top of that page — several posts follow those sources closely and say so.