wikipedia turns 25 today! the last unenshittified major website! backbone of online info! triumph of humanity! powered by urge of unpaid randos to correct each other! somehow mostly reliable! "good thing wikipedia works in practice, because it sure doesn't work in theory" - old wiki adage

At a technical university the steps of Pearl’s ladder are called stochastics, stochastic control and optimal transport

Causal inference is often hidden in plain sight. In a randomized clinical trial, the setup is such that interventional and conditional distributions coincide. That is E(X | do(T = t)) = E(X | T = t).

REMEMBERING HARRY VAN ZANTEN Botond Szabó and Aad van der Vaart in the ISBA Bulletin.

It’s giving late-game vibes of Sid Meier’s Civilization, where the player is bored and just trying to see what happens if they declare some wars before they abandon the game.

Students learn early that correlation does not imply causation. Correct, but incomplete... Certain patterns of independence and conditional dependence constrain #causal structure very strongly. A simple example:

Vaguely funny expression: a serious title

God made not only the numbers but also a sequence of independent standard normal random variables.

The (samizdat?) solution book to the classic soviet era “Problems In Mathematical Analysis” by Demidovich was called “Anti-Demidovich” by the students. 😀

The event B having happened is not the same as the receiving the information O that the event B has happened, because you also condition on receiving the information at all in the latter case. Conflating this contributes a lot to the "paradoxa" such as boy/girl paradox, Monty Hall...

Excited to share our recent paper, ”Compressing Biology,” to be presented at the Imageomics workshop at NeurIPS 2025. 🔬💻 Work led by my amazing PhD student Télio Cropsal. #cellpainting #stablediffusion #imageomics arxiv.org/abs/2510.19887

#AITHYRA, Vienna's new Biomedical AI institute, is hiring Postdocs! Come work with us. Openings in: 🔹 Generative AI 🔹 Multimodal ML 🔹 Virology 🔹 Enzyme Function Apply by Nov 20: oeaw.ac.at/aithyra/post... #PostDoc #AI #ML #Vienna #ScienceJobs

Great first day of the 3rd annual CHAIR Structured Learning Workshop @ Chalmers! 🥳 Event page & agenda: ui.ungpd.com/Events/60bfc... 1st day featuring: @betapata.bsky.social @janstuehmer.bsky.social @arnauddoucet.bsky.social @frejohk.bsky.social

Very nice by Jun Otsuka @junotk.bsky.social and Hayato Saigo: link.springer.com/article/10.1... about causal interventions/do calculus via string diagram surgery

Our process causation paper is published in Synthese! We propose that process causation (a la Salmon, Dowe, MDC new mechanists) can be modeled using a category-theoretic framework. link.springer.com/article/10.1...

Confounding is just the failure of the Doob conditioning functor from a Markov category into the associated category of backward-forward optics to be lax comonoidal

Oh, hats up to Mike Hicks who authored a policy paragraph on double-blind reviewing and amphibious type systems for POPL 2013 which has stand the test of time and is used in ICSME, OOPSLA, ECOOP, SPLASH, ICFP, PLDI, POPL, CSF, CAV... pldi12.cs.purdue.edu/others/dbr-f...

"But I am conditioning on the outcome all the time and I think I understand the world just fine!"

2025 CHAIR Structured Learning Workshop -- Apply to attend: ui.ungpd.com/Events/60bfc...

A golden Marburg Weidenhausen night accentuated by blue paper recycling bins…

So this is how I learned programming when I was twelve! 😃

Are you passionate about AI for molecular engineering? Just two weeks left to apply to the 2 PhD positions currently open in our team at Chalmers! 🎓 For details: ailab.bio/join-us

Please, please put your hyperparameter tuning procedure into the paper. For your method and the baselines

Yeah Moritz, *electricity*, that’s totally why you were doing this (bridges for random walks on random graphs)

link 📈🤖 Parallel computations for Metropolis Markov chains with Picard maps (Grazzi, Zanella) We develop parallel algorithms for simulating zeroth-order (aka gradient-free) Metropolis Markov chains based on the Picard map. For Random Walk Metropolis Markov chains targeting log-concave distributio

Don’t let your idea of good wine be shaped by a generation you wouldn’t trust to choose the tiles for an underpass

Tomorrow at #BayesComp @rseyer.bsky.social with arxiv.org/abs/2504.12190 (poster presentation, 19 Jun 2025, 5.30pm - 7.30pm local time)

It’s giving late-game vibes of Sid Meier’s Civilization, where the player is bored and just trying to see what happens if they declare some wars before they abandon the game.

Has nominative determinism gone too far h/t @benjaminschneider.ch

Right, you don't need error bars on error bars. Probabilistic uncertainty about uncertainty collapses. This is the “monadic join” in probability. Instead of a coin with random bias p ∼ π, you can flip a coin with the deterministic bias μ. Just take μ = E[p]. #statistics

Here's how the gradient flow for minimizing KL(pi, target) looks under the Fisher-Rao metric. I thought some probability mass would be disappearing on the left and appearing on the right (i.e. teleportation), like a geodesic under the same metric, but I was very wrong... What's the right intuition?

The call for papers for #CDSM2025 is out! 👇

The tragic landslide in Blatten gives me the excuse to tell you the story of how we found out Ice Ages existed. It's a cool story and the most important bit is rather similar to what's happening now.

🏹 Job alert: PhD Student for Machine Learning applications in Astronomy with Bernhard Schölkopf 📍 Tübingen 🇩🇪 🔗 More info: https://bit.ly/43ydV6l

Check this out: an accessible math font which actually looks nice! www.luciole-vision.com/en/math.html

The Berlin MATH+ Cluster of Excellence will continue to receive funding for another seven years idw-online.de/en/news852716

Matteo Giordano, Kolyan Ray Semiparametric Bernstein-von Mises theorems for reversible diffusions https://arxiv.org/abs/2505.16275

We propose Neurosymbolic Diffusion Models! We find diffusion is especially compelling for neurosymbolic approaches, combining powerful multimodal understanding with symbolic reasoning 🚀 Read more 👇

Unrelatedly, why is it „arrivederci“ and not „departderci“?

One of the perks of impostor syndrome is reading your past papers and thinking you'll never reach that level

Luke Hardcastle, Samuel Livingstone, Gianluca Baio Diffusion piecewise exponential models for survival extrapolation using Piecewise Deterministic Monte Carlo https://arxiv.org/abs/2505.05932

My Paris-based university is recruiting a researcher in machine learning, with a 5-year contract and reduced teaching load. The hosting lab (CERMICS) is an amazing environment for math & CS research. International candidates welcome, deadline next week! Link: cermics-lab.enpc.fr/2025/04/04/n...

Flow Matching in a nutshell.

I am very happy to share our latest work on the information theory of generative diffusion: "Entropic Time Schedulers for Generative Diffusion Models" We find that the conditional entropy offers a natural data-dependent notion of time during generation Link: arxiv.org/abs/2504.13612

Decisions during generative diffusion are analogous to phase transitions in physics. They can be identified as peaks in the conditional entropy rate curve!

New paper on SBI for stochastic nonlinear mixed-effects models! We propose a scalable Bayesian framework for hierarchical mixed-effects models, using amortized likelihood and posterior approximations, obtained without neural networks. 🔗 arxiv.org/abs/2504.11279

the network theorists are getting excited! that's never a good sign for you, if you are any kind of normal person