Sign in

Arthur

@riosarthur.bsky.social
132 followers 77 following 10 posts
PostsRepliesMedia
Reposted by Arthur
Maarten van Smeden @maartenvsmeden.bsky.social · 05/10/2026
NEW PAPER Shapley value explanations have been a very popular Explainable AI technique commonly used with the aim to improve transparency of black box machine learning models. Excellently led by @alcarriero.bsky.social www.nature.com/articles/s41...
1284
Reposted by Arthur
Olivia Guest · Ολίβια Γκεστ @olivia.science · 02/10/2026
> an increase in thin papers of narrow scope, as well as ‘salami’ papers, where a single work is broken up and submitted as a set of smaller papers. There is also a marked increase in dense, AI-written papers. AI tools are making it easy for authors to flood arXiv [and others with] low-value papers.
A histogram showing the incredible increase in submissions. 

> In September of 2016, arXiv received 9,869 submissions. In September of 2024, arXiv received 20,569 submissions. This September, arXiv received 40,363 submissions, which in turn generated almost 9,000 support tickets for arXiv staff and moderators. In only the past two years, submissions have doubled.
321763
Reposted by Arthur
Andrew Gelman et al. @statmodeling.bsky.social · 23/09/2026
Here’s a causal inference problem: where did the estimate of $64 million come from in “Blake Lively Says She Suffered $64 Million Financial Fallout From Justin Baldoni Conflict . . .”? statmodeling.stat.columbia.edu/2026/09/23/h...
statmodeling.stat.columbia.edu
Here’s a causal inference problem: where did the estimate of $64 million come from in “Blake Lively Says She Suffered $64 Million Financial Fallout From Justin Baldoni Conflict . . .”? | Statistical...
032
Reposted by Arthur
Raphael Nishimura @rnishimura.bsky.social · 17/09/2026
Quer entender melhor como funciona amostragem em estatística de uma forma visual e ilustrativa? Dê uma olhada nesse aplicação que eu criei (com ajuda do Claude): websites.umich.edu/~raphaeln/si... Todo feedback é bem-vindo!
123
Reposted by Arthur
Theiss Bendixen @theissbendixen.bsky.social · 10/09/2026
"Descriptive statistics also require causal inference" So true! In the Data Analyst's Guide, we illustrate it with a fun example: a registry study on spirituality among Danes 👇 theissbendixen.com/dag-book/
0194
Reposted by Arthur
Vincent Arel-Bundock @vincentab.bsky.social · 04/09/2026
Big news! 🎉 𝚖𝚊𝚛𝚐𝚒𝚗𝚊𝚕𝚎𝚏𝚏𝚎𝚌𝚝𝚜 1.0.0 for #Rstats is out. It’s a big number and it feels like a big step. I wrote a blog on the challenges of interpreting statistical models, SPEED, cool new features, the future, and a 5 year package development and writing odyssey. arelbundock.com/posts/margin...
1642299
Reposted by Arthur
Julia M. Rohrer @dingdingpeng.the100.ci · 27/08/2026
"Statistical analyses serve to answer statistical questions. It is imperative that these questions align closely with the substantive research questions. But quite often, research questions are vague or formulated in several logically distinct ways."
0356
Reposted by Arthur
Brandon Sheffield @brandon.insertcredit.com · 26/08/2026
Men only want three things
I'm holding a blu ray triple feature of the thing (1982), the thing (from another world), and the thing (2011), in front of shelves of other disc based media
352087530
Reposted by Arthur
Vincent Arel-Bundock @vincentab.bsky.social · 26/08/2026
Interesting essay by the great Kevin Munger. asteriskmag.com/issues/15/sa...
asteriskmag.com
Saving Academics from Themselves
The university will survive AI. That doesn’t mean we’ll like what it turns into.
0146
Reposted by Arthur
Jonas @jonassmarques.com · 25/08/2026
Visando corte de gastos pra investir bastante em IA, a Microsoft cortou 400000 licenças de ONGS ao redor do mundo. Só que óbvio, não comunicaram isso direito, e aí 40 % delas perderam arquivos de forma irreversível. slate.com/technology/2026/08/micros…
slate.com
The Quiet Decision Microsoft Made That Devastated Thousands of Nonprofits
When the tech giant retired a popular software grant, years of nonprofit data vanished with it.
14110
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 22/08/2026
there are two wolves inside you. one is a bear
0386
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 14/08/2026
For enjoyers of this thread on reasons not use ratio outcome variables in statistical analysis, I have a short blog post from 2022 about why dividing an outcome by something like population density doesn't always control for population density: elevanth.org/blog/2022/09...
elevanth.org
There Are No Magic Outcome Variables
I'm very sorry but I am going to write about statistics and causal inference again. I'd much rather be doing science. I'll make it brief. I was reading a preprint about the relationship between kinsh...
13811
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 11/08/2026
Working on my book, and as always getting distracted in side quests reading old primary lit. This week it seems to be this 1988 paper by Zellner advancing an optimality argument for bayes theorem (rather than usual coherence arguments). I might try to make a geometric version of this argument...
section 3 of Zellner's 1988 paper "Optimal Information Processing and Bayes's Theorem"
6223
Reposted by Arthur
Demetri @phdemetri.bsky.social · 08/08/2026
Long ass post on exchangeability because we built a hierarchcal bayes feature dpananos.github.io/posts/2026-0...
dpananos.github.io
Exchangeability in AB Testing – Demetri Pananos Ph.D
092
Reposted by Arthur
Demetri @phdemetri.bsky.social · 05/08/2026
LMFAO
0111
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 01/08/2026
I had this idea to do a series of short lessons on learning probabilistic programming (with Stan etc), based on the Zen "Ten Bulls": 1 search 2 discovery 3 perception 4 capture 5 taming 6 riding 7 goal forgotten 8 self forgotten 9 back to the start 10 teaching. Would take some work to develop though
3684
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 30/07/2026
I don't care what the haters say, tikz is nice, can have a tiny fish for a treat
LaTeX page showing a Tikz plot for a rescaled function of exp(x)(1-x), also with a small fish silhouette for funcode to generate the tikz plot:

\begin{tikzpicture}
	\begin{axis}[axis lines = left, xlabel = $T$, ylabel = $f$, 
			width=7cm, height=4cm, 
			tick label style={font=\tiny}, 
			label style={font=\small}       ]
		\addplot[domain=-3:27.5, red, thick, samples=70]
			{exp( (x-25)/(0.5*(30-25)) )*( 1 - (x-25)/(0.5*(30-25)) )};
		\node at (100,500) {\includegraphics[width=1.25cm]{fish.jpg}};
	\end{axis}
\end{tikzpicture}
7915
Reposted by Arthur
Emil Hvitfeldt @emilhvitfeldt.bsky.social · 28/07/2026
People have been asking me how one can use AI to help write quarto revealjs slides. Beyond pointing it at the static documentation, i have now created a number of skills that should be able to handle most tasks npx skills add emilhvitfeldt/slidecrafting.com github.com/EmilHvitfeld... #quarto
github.com
GitHub - EmilHvitfeldt/slidecrafting-book.com
Contribute to EmilHvitfeldt/slidecrafting-book.com development by creating an account on GitHub.
0244
Reposted by Arthur
Cecilia Tan @ceciliatan.com · 28/07/2026
Did they at least blame AI for making the slides?
194
Reposted by Arthur
derek guy @dieworkwear.bsky.social · 24/07/2026
Things will get worse as more students secretly use AI to do homework. There has always been a divide between those who care about education for its own sake and those who only care about grades, but AI will worsen this divide, leaving fewer people with the ability to do causal reasoning.
584586716
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 21/07/2026
When I am writing, I end up looking up lots of weird details. Today, I looked up who is responsible for the statistical independence symbol: Finnish mathematician Gustav Elfving in 1950s. He used a double orthogonal ⊥⊥, bc independence is stronger than orthogonality. Dawid's 1979 paper squished it.
Gustave Elfving, on right looking smoking hot, thinking about geometry no doubt.
4373
Reposted by Arthur
construído diferente @ddbm.video · 16/07/2026
"como que tá o trabalho aí o cliente tá cobrando" eu editando o cachorro do meu amigo pra parecer que ela tá atualizando:
6867260
Reposted by Arthur
Joseph Cox @josephcox.bsky.social · 15/07/2026
New: the highly controversial AI music generator Suno was hacked. The hacker sent us Suno source code; it shows the company scraped YouTube Music, Deezer, and Genius. In all, Suno scraped *decades* worth of music from the internet. Obviously didn't pay artists www.404media.co/hack-reveals...
6035491670
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 14/07/2026
If Haaland wanted a raccoon, he should have come to Leipzig. We have I think 1.2 million of them.
Postcard for the city of Leipzig, showing photos of invasive raccoons and nutria that live and thrive in the city.
4192
Reposted by Arthur
Andrew Gelman et al. @statmodeling.bsky.social · 10/07/2026
18 Associate Editors resign from Statistics and Computing editorial board: Problems with commercial scholarly publishing, and what does this all mean? statmodeling.stat.columbia.edu/2026/07/10/1...
statmodeling.stat.columbia.edu
18 Associate Editors resign from Statistics and Computing editorial board: Problems with commercial scholarly publishing, and what does this all mean? | Statistical Modeling, Causal Inference, and S...
065
Reposted by Arthur
maricona nível ômega 🐦‍🔥 @raphaeltf.bsky.social · 09/07/2026
ela recusou fazer PUBLI DE BET gente! vamos dar o nome aos bois ok? É DE BET QUE ELA TÁ FALANDO! ela deixou claro que “não poderia interpretar eunice paiva nos cinemas e compactuar com algo tão absurdo como bets”. sabe pq? PQ BET É UMA DESGRAÇA!!!
91378520
Reposted by Arthur
lastpositivist.bsky.social @lastpositivist.bsky.social · 04/07/2026
Three panel comic. In the first panel is a child pointing at a map being asked where is Brazil and identifying it. The second the same thing but for France. Then in the third he is asked "Where is Cabo Verde?" and places his hand over his heart as he says "here".
72412589
Reposted by Arthur
Rafe Meager (they/them) @economeager.bsky.social · 04/07/2026
I went here and everybody knew you
Metro exit sign saying Cope St
011511
Reposted by Arthur
Michael "Shapes Dude" Betancourt @betanalpha.bsky.social · 30/06/2026
"Machine learning", like "data science" and "artificial intelligence", has always been ambiguous language, appropriatable to whatever the current zeitgeist becomes, and whatever the most powerful people are trying to market. Scientists who live by the sword, risk dying by the sword.
1154
Reposted by Arthur
Charles Margossian @charlesm993.bsky.social · 26/06/2026
📘 With the release of our textbook "Bayesian Workflow" (avehtari.github.io/Bayesian-Wor...), I figured I'd also share the content of my graduate course on the topic at UBC. 🌎 charlesm93.github.io/stat547/ The course contains overlapping and complementary material, homeworks and reading.
avehtari.github.io
Bayesian Workflow book: Website – Bayesian Workflow book
Website for the Bayesian Workflow book by Gelman, Vehtari, McElreath, et al. — case studies, code, and exercises in R and Stan.
17419
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 24/06/2026
running some experiments now with priors on (ersatz-)R^2. Can really see the global shrinkage pattern, comparing prior on R^2 (left) to conventional independent priors (right). Only 2 "true" effects in this sim, easy to spot them
forest plot of marginal posterior distributions of regression coefficients
0111
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 23/06/2026
I've been trying to rewrite chapter 7 (the model comparison chapter) of my book along these lines. It's almost gaslighting to spend so much time deriving AIC, say how awesome it is, and then tell people not to use it. Still figuring out how to manage this psychodrama.
4585
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 21/06/2026
Poisson derived his distribution P(x|t)=exp(-Lt) (Lt)^x (1/x!) as a limiting case of binomial. Fine. But there is an axiomatic derivation that explains why the function looks like that. I can't smoosh it into this post, but it is pages 4-5 of this old Jaynes paper bayes.wustl.edu/etj/articles...
Denote b y h t the probabilit y there is no coun t in the time in terv al t No w the prop osition
R No coun t in t dt
is the conjunction of the t w o prop ositions
R No coun t in t No coun t in t t dt
and so b y the indep endence of dieren t time in terv als the pro duct rule giv es
h t dt h t dt
or h t h t The solution with the eviden t initial condition h is
h t e t
2536
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 17/06/2026
Remembering this WSJ graphic from 2007. Been thinking about "reverse statistics" lately, which means deriving the minimal set of assumptions that justify a result. Here I wonder if there is any function of the data that when max/min would produce such a curve. Ordinary Worst Squares?
Wall Street Journal graph from 2007 showing national corporate tax rates against tax revenue as percent of GDP. There is a curve drawn up and over the and down the other side of the points, but the curve is mainly outside the cloud of data.
1010614
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 17/06/2026
A million statisticians trying today to pronounce "Rousseeuw". I am no stranger to Belgian names, so I can tell you it is pronounced like Roooo, but round your lips and cough a little at the end Congrats to the R Core Team
3485
Reposted by Arthur
Jake Grumbach @jakemgrumbach.bsky.social · 06/06/2026
Gelman absolutely on fire on the blog statmodeling.substack.com/p/13-aspects...
69521
Reposted by Arthur
Demetri @phdemetri.bsky.social · 02/06/2026
Imagine a standing ovation for a Kaplan Meier estimator.
0151
Reposted by Arthur
al'ein @piratariaonline.bsky.social · 30/05/2026
"O sistema foi feito por universidades públicas, roda na nuvem do Serpro e é livre de comerciais e anúncios." PELO AMOR DE DEUS O FUTURO É ISOOOOOOOO
131480426
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 16/03/2026
Thanks to all for the HMM examples. The stats community here has grown to be very useful. I appreciate you all.
0373
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 08/03/2026
Last week, I mentioned this in passing in a workshop: In 1938 Enrico Fermi won a Nobel Prize for discovering two new elements of the periodic table. Lise Meitner shortly showed that Fermi was mistaken and instead had produced known lighter elements by fission. She did not win a Nobel prize.
Portrait of Lise Meitner taken in 1928. She is smoking a cigarette and looking impatient to get back to her experiments.
316251
Reposted by Arthur
Chelsea Parlett @chelseaparlett.bsky.social · 06/03/2026
Me when I can’t SELECT * FROM AMAZING_DATA.CLEAN_TABLE
My corgi nova looking suspicious
2341
Reposted by Arthur
Jairme @jairme.bsky.social · 06/03/2026
Assim hoje
529322
Reposted by Arthur
Adam L @adam-lg.bsky.social · 02/03/2026
A neat blog post walking through implementing Bayesian Decision Theory in @pymc.io with code! #statssky 📈📉 daniel-saunders-phil.github.io/imagination_...
Building a decision function over a Bayesian posterior
1349
Reposted by Arthur
Andrew Heiss @andrew.heiss.phd · 25/02/2026
New blog post! After years of annoying shenanigans, I discovered that nowadays it's *really* straightforward to get tikz diagrams in #QuartoPub to automatically turn into SVGs in HTML, docx, and Typst! Install Inkscape, install a Quarto extension, and you're done. www.andrewheiss.com/blog/2026/02...
2. Install the diagram extension in a Quarto project. Create a new Quarto project. From the terminal, run this to install the diagram extension:

quarto install extension pandoc-ext/diagram

3. Tell your document to use the extension. Add this to the YAML header of a Quarto document:

filters:
  - diagram

4. Add a .tikz block to your document. Add a tikz code block to your document like this:

```{.tikz}
\begin{tikzpicture}[>={stealth}]
  \node (x) at (0,0) {$X_{t}$};
  \node (y) at (2,0) {$Y_{t + 1}$};
  \node (z) at (1,1) {$Z$};
  \path[->] (z) edge (x);
  \path[->] (z) edge (y);
  \path[->] (x) edge (y);
\end{tikzpicture}
```

5. Render! Render the document to PDF and you should see a diagram. Render it to HTML or .docx or Typst and it should use inkscape to convert it to SVG. Magic.
210717
Reposted by Arthur
Demetri @phdemetri.bsky.social · 22/02/2026
A more definitive answer on A/B testing run times if you're Baysian dpananos.github.io/posts/2026-0...
dpananos.github.io
Demetri Pananos Ph.D - How Long To Run an A/B Test as a Bayesian
1112
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 18/02/2026
Almost done with my next lecture, so going back now to work on some new memes
Mental gymnastics meme.
TOP: Regression mental gymnastics. The effect is significant...even when I adjust for...everything we have measured.
BOTTOM: Causal Inference Mental Gymnastics.
unmeasured confounds could explain the effect
the sample might suffer from selection bias
differential measurement error!
SUTVA violation!
415218
Reposted by Arthur
Chelsea Parlett @chelseaparlett.bsky.social · 17/02/2026
I should start an educational series about “AI” targeted towards non-tech people who date men in cities that have high concentrations of “tech-bros” so that they’re unimpressed with the ones that don’t actually know what they’re talking about 😒
4211
Reposted by Arthur
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 12/02/2026
mel gibson passion of christ meme. Gibson looking comfortable, labeled HUMAN GENETICIST. Actor in bloody christ garb labeled PLANT GENETICIST
36114
Reposted by Arthur
Julia M. Rohrer @dingdingpeng.the100.ci · 02/02/2026
Finally listened to this lecture -- highly recommended if you've ever been confused about fixed effects vs. random effects models. Love that the framing starts from group-level confounding -- here's a problem, what could be solutions (rather than: here's a model, what are its features).
212414