Peter Richtarik (@peter_richtarik) 's Twitter Profile
Peter Richtarik

@peter_richtarik

Federated Learning Guru. Tweeting since 20.5.2020. Lived in 🇸🇰🇺🇸🇧🇪🇬🇧🇸🇦

ID: 4555244481

linkhttps://richtarik.org calendar_today21-12-2015 08:56:16

3,3K Tweet

7,7K Followers

647 Following

Francesco Orabona (@bremen79) 's Twitter Profile Photo

I have an opening for a post-doc position: I am looking for smart people with a strong CV in optimization and/or online learning All my ex post-docs (Kwang-Sung (Kwang) Jun, Mingrui Liu, and El Mehdi SAAD) became assistant professors, I'd like to continue this trend 😉 Please share it!

Peter Richtarik (@peter_richtarik) 's Twitter Profile Photo

This is not only NOT "next level", this is a completely natural and expected requirement for any reputable AI/ML venue. Reviewers who don't get this do not understand science well enough to play in the NHL of AI. Period.

Peter Richtarik (@peter_richtarik) 's Twitter Profile Photo

A related tech report by my student from a month ago: Hessian of Perplexity for Large Language Models by PyTorch autograd (Open Source) arxiv.org/abs/2504.04520

Peter Richtarik (@peter_richtarik) 's Twitter Profile Photo

We are excited about this work; and are thankful to the outstanding prior efforts of the folks who developed / analyzed Muon and Scion, including Keller Jordan Yuchen Yin Vlado Boza Jiacheng You, Franz Cesista, Laker Newhouse, Jeremy Bernstein, Thomas Pethick, Wanyun Xie, Kimon

Federated Learning One World Seminar (FLOW) (@flow_seminar) 's Twitter Profile Photo

📢: The 127th FLOW talk is on Wednesday (May 21) at **1 pm UTC.** Please register for our mailing list: bit.ly/3WVplLU. FLOW Calendar: bit.ly/3E9iIT4.

📢: The 127th FLOW talk is on Wednesday (May 21) at **1 pm UTC.**               

Please register for our mailing list: bit.ly/3WVplLU. 
FLOW Calendar: bit.ly/3E9iIT4.
Robert M. Gower 🇺🇦 (@gowerrobert) 's Twitter Profile Photo

Are you interested in the new Muon/Scion/Gluon method for training LLMs? To run Muon, you need to approximate the matrix sign (or polar factor) of the momentum matrix. We've developed an optimal method *The PolarExpress* just for this! If you're interested, climb aboard 1/x

Are you interested in the new Muon/Scion/Gluon method for training LLMs? 
To run Muon, you need to approximate the matrix sign (or polar factor) of the momentum matrix. We've developed an optimal method *The PolarExpress* just for this! If you're interested, climb aboard 1/x
Peter Richtarik (@peter_richtarik) 's Twitter Profile Photo

Hi from Vienna! Attending the Third Vienna Workshop on Computational Optimization 2025 (VWCO25)... ps-mathematik.univie.ac.at/e/index.php?ev…

Hi from Vienna! Attending the Third Vienna Workshop on Computational Optimization 2025 (VWCO25)...

ps-mathematik.univie.ac.at/e/index.php?ev…
Peter Richtarik (@peter_richtarik) 's Twitter Profile Photo

On Monday June 16, I will be in Bratislava, Slovakia - visiting the Kempelen Institute of Intelligent Technologies (KInIT). I will give a talk at KiNiT at 12:30. Hope to see some old friends, and hope to make new ones!

On Monday June 16, I will be in Bratislava, Slovakia - visiting the Kempelen Institute of Intelligent Technologies (<a href="/KInIT_sk/">KInIT</a>). I will give a talk at KiNiT at 12:30. Hope to see some old friends, and hope to make new ones!