Sign in

Ethan Epperly

@ethanepperly.bsky.social
265 followers 99 following 24 posts

PhD candidate in applied math at Caltech interested in computational linear algebra he/him

PostsRepliesMedia
Ethan Epperly @ethanepperly.bsky.social · 03/09/2025
New paper out with Chris Camaño, Raphael Meyer, and Joel Tropp re-examining sketching algorithms! Included: subspace injections as an alternative to subspace embeddings, the theory and practice of sparse sketching, tensor sketching, and much more! arxiv.org/abs/2508.21189
arxiv.org
Faster Linear Algebra Algorithms with Structured Random Matrices
To achieve the greatest possible speed, practitioners regularly implement randomized algorithms for low-rank approximation and least-squares regression with structured dimension reduction maps. Despit...
121
Ethan Epperly @ethanepperly.bsky.social · 05/08/2025
New blog post up about the amazingly useful Gaussian integration by parts formula! As an application, we use it to analyze power iteration from a random start www.ethanepperly.com/index.php/20...
Gaussian integration by parts. Let z be a standard Gaussian random variable. Then E[zf(z)] = E[f'(z)].
000
Ethan Epperly @ethanepperly.bsky.social · 11/07/2025
Very excited to share that I’ve been awarded a SIAM student paper prize! I look forward to seeing any of you who will be at #SIAMAN25 in Montréal. Thanks to the committee for selecting me for this honor www.siam.org/publications...
siam.org
2025 July Prize Spotlight | SIAM
Congratulations to the SIAM prize recipients who will be recognized at AN25, ACDA25, CT25, and GD25!
051
Ethan Epperly @ethanepperly.bsky.social · 08/07/2025
Reflections on five years of blogging www.ethanepperly.com/index.php/20...
ethanepperly.com
Five Years of Blogging – Ethan N. Epperly
030
Ethan Epperly @ethanepperly.bsky.social · 16/06/2025
New blog post up about the randomized Kaczmarz algorithm. The classic RK algorithms samples rows according to their squared norms, but what happens if you sample them uniformly? The answer surprised me: Uniform sampling is often just as good or even better www.ethanepperly.com/index.php/20...
120
Ethan Epperly @ethanepperly.bsky.social · 07/06/2025
New blog post out about the new Polar Express algorithm of Amsel, Persson, Musco, and Gower for computing the matrix sign function with applications to the Muon optimizer www.ethanepperly.com/index.php/20...
ethanepperly.com
A Neat Not-Randomized Algorithm: Polar Express – Ethan N. Epperly
040
Ethan Epperly @ethanepperly.bsky.social · 24/05/2025
New blog post out in my series on Markov chains! In this post, I discuss Poincaré inequalities and their connection to mixing of Markov chains www.ethanepperly.com/index.php/20...
ethanepperly.com
Markov Musings 5: Poincaré Inequalities – Ethan N. Epperly
0142
Reposted by Ethan Epperly
Clément Canonne @ccanonne.github.io · 10/03/2025
🧩 New week, time for our weₐᵉkly quiz! Today, another thing a bit random: Pokémon! Ash and Barry want to catch 'em all: all of them. You know, Pikachu, Jigglypuff, err... Charmander? It's been a while. So, n Pokémon to catch, and no idea how long it'll take. Gotta help them out! #WeaeklyQuiz 1/
media.tenor.com
a bunch of red and white balls in the sky
Alt: A lot of red and white Pokeballs falling from the sky
1175
Ethan Epperly @ethanepperly.bsky.social · 25/02/2025
New blog post with four proofs of the Schur product theorem. Do you know a fifth? www.ethanepperly.com/index.php/20...
Proof of the Schur product theorem: The Kronecker product $A\otimes M$ of two psd matrices is psd. The entrywise product $A\circ M$ is a principal submatrix of $A\otimes M$: $$A\circ M = ((A\otimes M)_{(i+n(i-1))(i+n(i-1))} : i = 1,\ldots,n).$$All principal submatrices of a psd matrix are psd, so $A\circ M$ is psd.
110
Ethan Epperly @ethanepperly.bsky.social · 14/02/2025
New blog post up! In it, I look at the question: how accurate is sketch-and-solve method for least squares? A standard bound suggests the residual is within a 1 + O(η) factor of optimal for an embedding of distortion η. But this isn't the correct answer! www.ethanepperly.com/index.php/20...
ethanepperly.com
Note to Self: How Accurate is Sketch and Solve? – Ethan N. Epperly
010
Ethan Epperly @ethanepperly.bsky.social · 16/12/2024
Delightful little tale by Nick Trefethen: people.maths.ox.ac.uk/trefethen/ba...
people.maths.ox.ac.uk
010
Ethan Epperly @ethanepperly.bsky.social · 12/12/2024
What is your favorite proof of the Cauchy–Schwartz inequality? I wrote about my favorite proof, which uses matrix theory, in a new blog post. Check it out! Also included: a matrix theoretic proof of Jensen’s inequality for 1/x www.ethanepperly.com/index.php/20...
ethanepperly.com
My Favorite Proof of the Cauchy–Schwarz Inequality – Ethan N. Epperly
021
Ethan Epperly @ethanepperly.bsky.social · 09/12/2024
Did you know that randomized Nyström approximation of A is equivalent to running the randomized SVD on A⁰ᐧ⁵? This and other surprising facts on this week's blog post on the "Gram correspondence" www.ethanepperly.com/index.php/20...
ethanepperly.com
Low-Rank Approximation Toolbox: The Gram Correspondence – Ethan N. Epperly
2186
Ethan Epperly @ethanepperly.bsky.social · 08/12/2024
This whole “advent of research” series of posts by David is really excellent, but I love this one in particular
010
Ethan Epperly @ethanepperly.bsky.social · 05/12/2024
New blog post up! The randomized Kaczmarz algorithm doesn’t converge for inconsistent systems of linear equations, but—as an estimator for the least-squares solution—it does have an exponentially decreasing bias www.ethanepperly.com/index.php/20...
ethanepperly.com
Randomized Kaczmarz is Asympotically Unbiased for Least Squares – Ethan N. Epperly
020
Ethan Epperly @ethanepperly.bsky.social · 02/12/2024
New paper out with Gil Goldshlager and Rob Webber! In it, we show that *tail averaging* can be used to improve the accuracy of the randomized Kaczmarz method for solving least-squares problems. The resulting method, TARK, outcompetes other row-access methods for least squares
Error for different randomized Kaczmarz methods applied to a least-squares problem. Tail-averaged randomized Kaczmarz (TARK) outcompetes the existing methods
120
Reposted by Ethan Epperly
Sam Power @spmontecarlo.bsky.social · 30/11/2024
Cross-posting this - please join us! Mailing List link: groups.google.com/g/internatio... YouTube Channel link: www.youtube.com/@MonteCarloS...
1205
Reposted by Ethan Epperly
Anupam Gupta @anupamg.bsky.social · 30/11/2024
A reminder about NY Theory Day in a week! Fri Dec 6th! Talks by Amir Abboud, Sanjeev Khanna, Rotem Oshman, and Ron Rothblum! At NYU Tandon! sites.google.com/view/nyctheo... Registration is free, but please register for building access. See you all there!
sites.google.com
Home
About The New York Theory Day is a workshop aimed to bring together the theoretical computer science community in the New York metropolitan area for a day of interaction and discussion. The Theory Da...
1459
Reposted by Ethan Epperly
Quanquan Gu @quanquangu.bsky.social · 23/11/2024
Just created the Starter Pack for Optimization Researchers to help you on your journey into optimization! 🚀 Did I miss anyone? Tag them or let me know what to add! go.bsky.app/VjpyyRw
13378
Ethan Epperly @ethanepperly.bsky.social · 21/11/2024
New blog post up presenting some beautiful *exact formulas* for sketched least squares with a Gaussian embedding. These beautiful formulas appear to have only been published as recently as 2020; see post for details! www.ethanepperly.com/index.php/20...
130
Ethan Epperly @ethanepperly.bsky.social · 05/12/2023
Very excited to be attending #NeurIPS2023 next week where I’ll be presenting my work “Kernel quadrature with randomly pivoted Cholesky” with Elvira Moreno. I’ve written a little blog post to explain what kernel quadrature is and what our approach is to it!
ethanepperly.com
Five Interpretations of Kernel Quadrature – Ethan N. Epperly
050