Marcel Schweitzer

When using Lanczos for approximating f(A)b, how much can the result deviate from the best possible approximation hidden in the Krylov space? My new preprint goes a step towards answering this question for many important matrix functions: arxiv.org/abs/2503.04427

I'm currently looking for a PhD student to work with me on randomized methods for matrix functions: …ellenausschreibungen.uni-wuppertal.de/qisserver/rds?… (English on page 3-4)

Summary of the outcome of the ample discussion sessions at the "f(A)bulous workshop on matrix functions and exponential integrators" in Magdeburg: arxiv.org/abs/2401.16132 Dedicated to the memory of one of the greatest researchers, writers and human beings our community ever had.

Now also with link: arxiv.org/abs/2401.06684 🤦‍♂️

New work with A. Frommer, G. Ramirez-Hidalgo and M. Tsolakis on preconditioning techniques for (inverse) matrix square roots, including MPI implementation.

Great, energetic show by @TheRonnieRomero yesterday 🤘 Also, Musiktheater Piano is a really cool venue 😎

@guettel That's a very good question... I wonder the same thing now :-D

@MarcelSchweitz1 Nice work! We just posted arxiv.org/abs/2311.14206 -- now I wonder how your algorithm would compare to using sketched GMRES for solving the linear system formulation of the matrix equation (with a Kronecker structured sketch operator)

New preprint on randomized Krylov methods for large-scale Sylvester/Lyapunov equations (algorithmic aspects + theoretical analysis). Turns out to be highly competitive with state-of-the-art rational Krylov both regarding run time and memory consumption: arxiv.org/abs/2311.16019

Looking forward to a great workshop 😊

Unfassbar 🥳🏀🥇 Heute wurde Sportgeschichte geschrieben! #BasketballWM

New work with Andreas Frommer and Michele Rinelli analyzing stochastic trace estimation methods based on graph colorings: arxiv.org/abs/2308.07722

New manuscript on analysis of sketched Krylov subspace methods with Davide Palitta and Valeria Simoncini: arxiv.org/abs/2306.06481 (just in time for my talk about the topic at the @ILAS_news meeting today 🙂)

After a week of vacation: Finally had time to actively listen to some of the cool stuff released in the last two months🤘 While they're all great, @elegantweapons_'s "Horns for a Halo" is my winner (Highlight: "Ghost of you" - incredible guitar licks and mind-blowing vocals)

20 of the most important ideas everyone should know: 1. Cunningham's Law: The best way to find the right answer on the internet is not to ask the right question, but to post the wrong answer. Why? Because people are more interested in criticizing others than helping them.

@MetalHammer 1. Iron Maiden - The Number of the Beast 2. Metallica - Ride the Lightning 3. Fates Warning - Awaken the Guardian 4. Threshold - March of Progress 5. Nevermore - Dead Heart in a Dead World

Five metal albums to get to know me, go! 1. 2. 3. 4. 5.

Always getting goosebumps when @Crush40Johnny sings "The Line"! #axelrudipelltour

When measuring conditioning of matrix functions, how sensitive to perturbations is the condition number itself? And how does this change when working in a structure-preserving manner? New work on this topic together with @barslan4 and @sdrelton: arxiv.org/abs/2304.14077

Superb line-up at #rockmeetsclassic yesterday! In particular the phenomenal @TheRonnieRomero! Long live rock and roll🤘

New work on how to efficiently identify network modifications to which communicability indices react particularly sensitive: arxiv.org/abs/2303.01339

@uria_mor I sincerely hope so 😉

Should be an interesting read