Dmitry Kovalev

[Dmitry Kovalev @dakovalev1]

AC made certain requests regarding my #NeurIPS2026 submission in the initial metareview that directly violate the official reviewing guidelines, while simultaneously declaring those requests impossible to satisfy, effectively rendering the author-reviewer discussion pointless before it had even started.

Today, I tried to report this to the SACs but was basically told to fuck off. How am I supposed to report violations like this? Or should I just accept that NeurIPS reviewing has always been, and always will be, nonsense?

[Peter Richtarik @peter_richtarik]

Former PhD student revolutionizing our understanding of Adam and AdamW!

[Konstantin Mishchenko @konstmish]

That's a nice paper, very neat.

[Pasquale Minervini @PMinervini]

How do you like to spend your time?

[Dmitry Kovalev @dakovalev1]

@predict_addict @srchvrs @LucaAmb @peter_richtarik The so-called “factual issues” are blatant lies and insults to several people. Keep going.

[Dmitry Kovalev @dakovalev1]

@predict_addict @srchvrs @LucaAmb @peter_richtarik As soon as you start watching yours.

[Dmitry Kovalev @dakovalev1]

@predict_addict @srchvrs @LucaAmb @peter_richtarik Maybe you should try reading with your eyes instead of your ass.

I am not a PhD student, the original thread clearly contains lifelong consequences, and I am not defending pollution of science.

[Dmitry Kovalev @dakovalev1]

@predict_addict @srchvrs @LucaAmb There are life-long consequences that can be considered an effective lifetime ban.

[Dmitry Kovalev @dakovalev1]

@nomoneyforfilm @miniapeur Just out of curiosity, do you read every single bibtex entry that you copy-paste from google scholar, line by line?

[Dmitry Kovalev @dakovalev1]

@ChShersh The definition is incorrect. Did you generate it with AI?

[Peter Richtarik @peter_richtarik]

I am pleased to announce that together with 3 of my (then present and now former) KAUST PhD students

- Samuel Horváth @sam_hrvth (now Assistant Professor at MBZUAI, Abu Dhabi),
- Dmitry Kovalev @dakovalev1 (now Research Scientist at Yandex, Moscow),
- Konstantin Mishchenko @konstmish (now Research Scientist at Meta, Paris)

and Sebastian U Stich @SebastianUStich (now Faculty at CISPA, Saarbrücken; who was visiting us at KAUST at the time)

we have won the 

2023 Charles Broyden Prize

for the paper

Stochastic distributed learning with gradient quantization and double-variance reduction

the paper: https://t.co/hTJpLM3ScW

previous prizes: https://t.co/375niVZSsG

*****

The Charles Broyden Prize is an annual international award honoring the best paper published in the journal Optimization Methods and Software (OMS) during the preceding year. Established in 2009, the prize commemorates the life and work of British mathematician Charles George Broyden (1933–2011), a pioneer in numerical optimization known for his namesake methods and his role in the development of the BFGS algorithm.

*****

I consider this prize to be shared with the authors of the original DIANA paper

Konstantin Mishchenko @konstmish, Eduard Gorbunov @ed_gorbunov, Martin Takáč @TakacMartin and P.R. Distributed learning with compressed gradient differences, https://t.co/Snlj4jGgF5,

which was the inspiration for this follow up work that polished, generalized and improved it, and the authors of the previous SEGA work,

Filip Hanzely, Konstantin Mishchenko @konstmish and P.R.. SEGA: Variance reduction via gradient sketching, Advances in Neural Information Processing Systems 31:2082-2093, 2018,

which can be seen as a single-node variant of DIANA, as well as the authors of the JacSketch paper, which was the inspiration for SEGA:

Robert M. Gower @gowerrobert, P.R. and Francis Bach @BachFrancis, Stochastic quasi-gradient methods: Variance reduction via Jacobian sketching, Mathematical Programming 188:135–192, 2021 (https://t.co/gtOm87M7YP)

One can keep going like this, since every new discovery builds on prior work, but I'll stop here.

So, once again, congrats to the authors of the award-winning paper, as well as to the authors of all these prior works!

[Konstantin Mishchenko @konstmish]

I find this new paper on adaptive methods very impressive. It develops a new accelerated method with adaptive estimation of smoothness. Its unique feature is that the method keeps adapting to local geometry as long as we run it. It's not trivial, so let me explain in a thread.

[Konstantin Mishchenko @konstmish]

Several research groups have released papers on the convergence of Muon, mostly looking at either as a Frank-Wolfe method with momentum or a trust region procedure. I find this one to be particularly easy to read.

[Dmitry Kovalev @dakovalev1]

@lpr_tw

[Dmitry Kovalev @dakovalev1]

@peter_richtarik @TheSandyCoder @KAUST_News lol

[Dmitry Kovalev @dakovalev1]

@vigimadi @ioannZH Не зависимо от того, как вы относитесь к т.н. репарациям, надо быть безмозглым выродком, чтобы не видеть разницы между выплатами в пользу граждан РФ и выплатами в пользу других государств.

[Dmitry Kovalev @dakovalev1]

@Resurrexit_Luci @pushistyyy @JohnRMetzner @TimothyDSnyder This attitude will get you in trouble one day.