Useful Links
People, papers, and rabbit holes that shaped how I think about all this.
How I Ended Up HereI studied with Eleftherios Garyfallidis (creator of DIPY) during my PhD, working on denoising diffusion MRI: Patch2Self (NeurIPS 2020, with Josh) and Patch2Self2 (CVPR 2024, with Petros, Agniva, and Josh). Studying noise means studying variance, which means studying leverage.
Conformal Prediction
If you want to learn conformal prediction from scratch, start here. I did.
Leverage Scores & Randomized Linear Algebra
Most of what I know about leverage scores, I learned from Petros Drineas. I started working with him during my PhD, and he opened up an entire world of randomized numerical linear algebra that I didn’t know existed.
A lot of the leverage scores literature I absorbed came through working alongside Agniva Chowdhury, who was Petros’s PhD student at Purdue and is now an AI Research Scientist at Intel.
Statistics Notes Worth Bookmarking
Cosma Shalizi (CMU Statistics) — rigorous, opinionated, and somehow entertaining. If you bookmark nothing else from this page, bookmark these.