Dmitry Kobak
@hippopedoid
Researcher at @uni_tue. Manifold learning, contrastive learning, scRNAseq data. Excess mortality. Born but to die and reas'ning but to err.
From @boazbaraktcs: "Up to this year, I did not think of myself as a 'Zionist' any more than my U.S.-born daughter is an 'Americanist.' I did not consider the existence of my birth country as being up for debate. Yet apparently, at Harvard, it is." thecrimson.com/article/2024/1…
What does Elon Musk's online activity reveal about him? Going through nearly 40,000 posts over the past decade, and the policies implied or suggested in them, @AinslieJstone and I found a changed man
Really excited to start rolling out this scrollytelling mode for Atlas maps with an analysis of congressional tweets -- reach out if you have a story you want to tell about a big textual dataset!
What do 3.2 million @X posts from Congress show about how US legislators talk? What are they posting about going into the 2024 US presidential election? Learn what we found in 3.2M posts from Congress using Nomic Atlas:
There are three plots in this figure and all are terrible at conveying what they are intended to convey.
We updated our ICLR dataset (see github.com/berenslab/iclr…) with blind 2025 submissions to @iclr_conf. Over 10k submissions this year. I like how this embedding (SBERT + tSNE) shows which ML areas are old-school and which ones are currently booming.
We updated our ICLR dataset (see github.com/berenslab/iclr…) with blind 2025 submissions to @iclr_conf. Over 10k submissions this year. I like how this embedding (SBERT + tSNE) shows which ML areas are old-school and which ones are currently booming.
I will tolerate recurrent laryngeal nerve slander no longer! It's actually the result of several elegant solutions to difficult problems in embryology, and the length is a non-issue. A 🧵 1/13
My meager education in biology and evolution gave me the mistaken impression that evolution optimized everything. But it didn't. One example is the recurrent laryngeal nerve (RLN). It goes from behind your ear, loops down below your aorta, and then back up to the voice-box 1/2
2¹³⁶²⁷⁹⁸⁴¹−1, discovered today, is the largest known prime. It's a Mersenne prime (2ᵖ-1), which are easier to find. It took nearly 6 years for the GIMPS software to find it after the previous largest known prime. It was also the first Mersenne prime found using GPUs.
A good heuristic about whether something can in principle be solved with deep learning: can it be done by an expert human 100% intuitively, without having to consciously think about it? If yes, then pure pattern recognition is an appropriate approach to the problem, and given…
NeurIPS @NeurIPSConf requires to include the "Paper Checklist" into camera-ready papers, including (!) "Guidelines" bullet point lists which are *identical in all papers* and take ~5 pages. Isn't it ugly & useless? I'd like at least to delete the guidelines, but PCs told me no.
So @JeremyFarrar asked for an "explainer thread" on Noble Prizes in Chemistry around AlphaFold and Protein design, so here goes.
Man sits by me on train. MAN: Loads of psychopaths around here ME: Really? MAN: Loads mate ME: How'd you know? MAN: There's signs aren't there? ME: I guess? MAN: I love them (47 minutes of awkward silence.) Man leaves train, he has a bike. I realise he was saying 'cycle paths'.
Datamapplot 0.4 is out now, and has far more powerful and effective interactive plots. Here is an example of a Data Map of 2.4 million papers on ArXiv, ready to be explored.
A bit surprised by all the debate around yesterday's Nobel Prize. It is well known that all science is either physics or stamp collecting. What Hopfield and Hinton did is clearly *not* stamp collecting. Ergo it is physics.
John Hopfield has a nice article in the annual reviews of condensed matter physics. It starts off with a discussion of what physics is, which I think is totally on point.
BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Physics to John J. Hopfield and Geoffrey E. Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”
We’re launching our new group leader call: 👉 4 open positions! We're recruiting both junior and senior profiles, who use and develop AI methods and mechanistic mathematical models to address fundamental questions in biology. vib.ai/en/opportuniti…
So thrilled for the new chapter at VIB.AI and @KU_Leuven ! Beyond excited to contribute to their mission by diving deeper into the mysteries of the protein universe, working together to unlock new biology. Looking forward to what’s ahead in beautiful Leuven🚀🔬
We're thrilled to announce @Joana_MPereira will join VIB.AI as group leader! Her pioneering research combines deep learning, network analysis, and bioinformatics to investigate protein diversity and evolution. 👉 lnkd.in/eemNRkm5 Welcome aboard, Joana!
Researching #AI #MachineLearning #Robotics #HCI? Join our elite doctoral program - a partnership with MPI-IS, @uni_stuttgart & @uni_tue! Applications accepted until Nov 15, 2024 at imprs.is.mpg.de/application
Super happy to announce the release of TorchDR 0.1 : the Dimensionality Reduction (DR) library using @PyTorch. Joint work with an amazing team : @RFlamary , @mathusmassias, @nicolas_courty, @CedricCuaz, Titouan Vayer, Aurélien Garivier. Thread 🧵. Repo : github.com/TorchDR/TorchDR