Sergey Bartunov

Stay tuned 🎧

Come and talk to us about current opportunities to join our interdisciplinary team at @icmlconf on 21-27 July. Share your details here: bit.ly/3VSq6qx #ICML2024

We’ll be at @icmlconf 21-27 July - if you'll be there we’d love to chat to you about our rational drug design work and opportunities to join our team. Register here bit.ly/3VSq6qx to arrange a time. #ICML2024

There is actually no need of topology and the answer is quite trivial:

What do polar coordinates, polar matrix factorization, & Helmholz decomposition of a vector field have in common? They’re all implied by Brenier’s Theorem: a cornerstone of Optimal Transport theory. It’s a fundamental decomposition result & deserves to be better known. 1/5

Tomas Mikolov, the OG and inventor of word2vec, gives this thoughts on the test of time award, and the current state of NLP, and chatGPT. 🍿

@y0b1byte That’s because cars are not the optimal form of transport…

It happened again! Came to see amazing progress in autonomous driving, but there's only weird math in there.

Excited to share our @nature paper on #FunSearch (Search in Functional space)! tl;dr: we find new results in maths by evolving code with a pretrained (no finetuning involved) medium-size LLM acting as crossover/mutation operator deepmind.google/discover/blog/… storage.googleapis.com/deepmind-media… 1/n

Removed, enjoy !

@pfau Yeah, that's clearly progress, but sampling lots of solutions doesn't impress me yet. I think it's fair to allow much smaller number of samples though

Feels like this release got overshadowed by the Gemini announcement. Remember that Codeforces is one of the domains in which GPT-4 did abysmally poorly (at least on held out problems...it just memorized the answer to problems in the training set).

@round_piglets Have we finished with decolonisation of mathematics yet? I suggest we don't move further without ensuring the proper foundation for AI

👍🏻👍🏻👍🏻👍🏻

Pushkin house tonight

I just published the story of how I created the world’s first No-Limit Holdem poker solver and made $500k by age 23 @olegostroumov/worlds-first-poker-solver-6b1dbe80d0ee" target="_blank" rel="nofollow noopener">medium.com/@olegostroumov… I had to keep the story secret since 2013, but now you can read how I went from near broke to reshaping world's toughest poker games

@mosicr @pmarca Don't really want to start a tasteless argument on Hitler, let me just leave an obviously non neutral reference here winstonchurchill.org/publications/f… One may wonder why did Churchill had such a nuanced (to put it mildly) attitude towards Hitler, perhaps there could be a reason

@sbos @pmarca

One plausible reading of history is the persistent inability of smarter people to talk dumber people into doing what the smarter people want -- for better and/or worse.

@mosicr @pmarca And it's not that some smart people didn't have a tactical interest in Hitler convincing dumb people

@pmarca Hitler was smart, convinced dumb people.

@Manka2084 Как так вышло?

С днем когда в баре разбилось бухла на милион деняк

@hr0nix Which ones?

Sad to see Turing Award winners not being able to grasp simple technical terms like "autoregressive".

Both the company and I reported this to @Glassdoor. We explicitly said that there is only one person who can be identified as an author of the review and that would be me, except I didn't write it (I have never left a company review on Glassdoor or a similar service)

The review is very rude, negative, factually incorrect and clearly had an intent of damaging the company's and my reputations. While I obviously entirely happy there, otherwise I would still be working there, I strongly condemn and disagree with what the review says.

Can someone please explain me all of this like to a 5 years old? Step by step.