Remarkable that Igor Savelyev credits Lenin with the definition of matter! Being a Soviet scientist was a tricky business -- even physics books needed to pass the censor.

Modeling and numerics: could we do something better? by Juho Timonen. www.generable.com/post/modelin...

Personalized dosing decisions beyond clinical trials: www.generable.com/post/improvi...

Nonparametric statistics: Gaussian processes and their approximations by Nikolas Siccha www.generable.com/post/nonpara...

If you are interested in meta-analysis (Bayesian or non-Bayesian), check out @rlgrant.bsky.social Robert Grant's nice StanBio Connect talk available at www.youtube.com/watch?v=HbZ1... in which he also advertises his soon-to-be shipping book "Bayesian meta-analysis" www.routledge.com/Bayesian-Met...

StanBio Connect 2025 free online conference May 30 stanbio.org ... Stan in biomedicine including drug discovery and development, bioinformatics, medical devices, health economics, real-world evidence, and decision-making under uncertainty

Things I read listened to, and saw in 2024: ericnovik.com/2025/03/book...

The Scientific Method Vs. AI: Is There Still Room For Theory? My take here: www.forbes.com/councils/for...

StanConnects are a series of one-day online conferences for the application of Stan to specific areas. This year, we're kicking off with Stan for biology (broadly defined), co-organized by @ericnovik.bsky.social and @vianeylb.bsky.social, on 30 May '25!! stanbio.org

“In the fall of 1972, President Nixon announced that the rate of increase of inflation was decreasing. This was the first time a sitting president had used the third derivative to advance his case for re-election.” (Notices of the AMS, Oct. 1996, 43(10), p. 1108.)

Turns out there Bayesian BioPharm Meeting 2025. bnp14.org/bbp2025/. Anyone going?

My daughter's pork belly ramen is the best I've ever had.

I was just talking to an experienced probabilistic modeler who said: "After working with a few probabilistic programming languages, Stan is my fave: everything is typed and everything just works." It's a pain to get things in and out of Stan, but it's usually worth it.

2025 is a 2, 3, 4, 5 kind of year. Why? Because 2025 = 3^4 * 5^2

Modeling Public Opinion Over Time and Space: Trust in State Institutions in Europe, 1989-2019 by Kołczyńska et al. ojs.ub.uni-konstanz.de/srm/article/...

"When you are playing a point, it's the most important thing in the world. But when it's behind you, it's behind you..." -- Roger Federer

Thrilled to share our new @pnas.org paper using active learning 🤖 to optimise disease testing 🦠 and support public health 🌍. Press release by @imperialsph.bsky.social: imperial.ac.uk/news/259573/... Paper: www.pnas.org/doi/10.1073/... Joint work with some stellar collaborators!

If you are interested in doing a #PhD with me at Imperial College London and qualify as a home student, please reach out (before end of 2024)! Potential topics: spatial statistics, applied deep generative models, probabilistic programming and more.

Why is a bullshitter worse than a liar? Because he is not even wrong! en.wikipedia.org/wiki/On_Bull... en.wikipedia.org/wiki/Not_eve...

A lot of people heard about John von Neumann but not about Alexander Craig Aitken, but Aitken was quite remarkable particularly if you are a statistician. For example, he independently derived Cramer-Rao lower bound, a very nice result in (frequentist) statistics. en.wikipedia.org/wiki/Alexand...

Traveling salesmen are still out of luck, but the commercial movers finally caught a break with this breakthrough! phys.org/news/2024-12...

Statistical wars inside the frequentist camp: According to Neyman, though, the argument began because Fisher demanded that Neyman lecture only from Fisher’s book. When Neyman refused, Fisher promised to oppose him “in all my capacities.” - The Theory That Would Not Die

Ahh, Indox zero.

Why do CS people write nice-looking code while statisticians don't? For a computer scientist, code is the product; for a statistician, code is a byproduct.

A new release of Stan is coming discourse.mc-stan.org/t/cmdstan-st.... Highlights: New constraints for stochastic matrices and zero-sum vectors Easier user-defined constraints Improved diagnostics New distribution, beta_negative_binomial

Lisbon looks very nice this morning.

What are pragmatic Bayesians? We all want to use optimization. Why? Because it's fast and we want to move on with our lives. In complex settings, optimization doesn't work well or at all.

Running Markov Chain Monte Carlo on Modern Hardware and Software by Sountsov, Carroll, and Hoffman, showing some new performance benchmarks. arxiv.org/abs/2411.04260

Roger Penrose convinced me to study some complex analysis, and this is from the book Visual Complex Analysis by Needham, which is a gem. (I don't know Penrose personally, but I tried reading The Road to Reality)

Note to self: In pretty much any field, it helps to think like an economist -- people and organizations will behave according to incentives that you (the people or the person) put in place. Corollary: be careful not to hang your incentives on a single number/metric, as it can probably be gamed.

CmdStan & Stan 2.36 release candidate is now available with new constrained types, improved diagnostics, and a new distribution for heavy-tailed count data. See more at discourse.mc-stan.org/t/cmdstan-st... and test with your own models! #Bayesian #mcmc_stan

Meta-analysis with only one study? No problem. stat.columbia.edu/~gelman/rese...

Looks like a nice package for Causal Inference in Python: causalpy.readthedocs.io

Causes of death globally, 2019. Source: Our World in Data ourworldindata.org/causes-of-de...