Johan de Aguas

Orca mom carries dead newborn calf in San Juans

It is not the first time such on-the-nose grief has been observed in this endangered population. In January, orca J35 carried a dead baby for at least five days. J35 had previously made world headlines in 2018 when she carried her dead calf for 17 days and 1,000 miles. From @kuow.org

‘World Models,’ an Old Idea in AI, Mount a Comeback | Quanta Magazine You’re carrying around in your head a model of how the world works. Will AI systems need to do the same?

"World models," internal AI simulations of reality, are again at the forefront of AI research. While today's LLMs use a fragile patchwork of heuristics, developing true world models is seen as essential for creating robust and adaptable intelligence. #MLSky

Susan Athey, Guido Imbens, Zhaonan Qu, Davide Viviano
Triply Robust Panel Estimators
https://arxiv.org/abs/2508.21536

link 📈🤖
Causal inference amid missingness-specific independencies and mechanism shifts (Aguas, Henckel, Pensar et al) The recovery of causal effects in structural models with missing data often relies on $m$-graphs, which assume that missingness mechanisms do not directly influence substantive va

Johan de Aguas, Leonard Henckel, Johan Pensar, Guido Biele
Causal inference amid missingness-specific independencies and mechanism shifts
https://arxiv.org/abs/2506.15441

@guidobiele.bsky.social

Sometimes psychologists collaborate with causal inference experts and use it for their research, here the probability of benefit or harm of ADHD medication for school outcomes: arxiv.org/abs/2502.10049
All math by @johandh2o.bsky.social

YouTube video by StarTalk Tackling the Biggest Unsolved Problems in Math with 3Blue1Brown

I loved this conversation of Neil deGrasse Tyson and Grant Sanderson about the hardest math problems: www.youtube.com/watch?v=7DEW...

Mathematical Beauty, Truth and Proof in the Age of AI | Quanta Magazine Mathematicians have started to prepare for a profound shift in what it means to do mathematics.

A new paradigm in math is coming. AI will push it toward a form of neointuitionism and hermeneutic practice, where proofs and math objects are sampled from a system, whether social or artificial, and then interpreted by humans (as in social sciences).
www.quantamagazine.org/mathematical...

Does one have to be a genius to do maths? Better beware of notions like genius and inspiration; they are a sort of magic wand and should be used sparingly by anybody who wants to see things clearly. (José Ortega y Gasset, “Notes on t…

Does one have to be a genius to do math? “NO. To make good and useful contributions to math, one does need to work hard, learn one’s field well, learn other fields and tools, ask questions, talk to other mathematicians, and think about the BIG picture"
terrytao.wordpress.com/career-advic...

Mathematical Thinking Isn’t What You Think It Is | Quanta Magazine The mathematician David Bessis claims that everyone is capable of, and can benefit greatly from, mathematical thinking.

Many people, especially those untrained, truly believe doing math is just about following strict rules to manipulate symbols on paper. It's more about imagining, playing a naïve for the obvious, and breaking the rules along the way.

www.quantamagazine.org/mathematical...

Long-term effect of pharmacological treatment on academic achievement of Norwegian children diagnosed with ADHD: a target trial emulation AbstractBackground. Attention-deficit/hyperactivity disorder (ADHD) is one of the most commonly diagnosed mental disorders in children. For many patients,

ADHD medication might improve learning much less than you would expect.

💊 Our new pretest-posttest control group design target trial emulation on the long-term effects of ADHD medication, together with Kristin Romvig Øvergaard, Arnoldo Frigessi & @guidobiele.bsky.social, has just been published. 🧵

Recovery and inference of causal effects with sequential adjustment for confounding and attrition Confounding bias and selection bias bring two significant challenges to the validity of conclusions drawn from applied causal inference. The latter can stem from informative missingness, such as in ca...

Recovery and inference of causal effects with sequential adjustment for confounding and attrition - Read on: www.degruyter.com/document/doi...

Tenure-Track Faculty in all areas related to Information Security and Artificial Intelligence (f/m/d) - extended call | CISPA | Jobs

🏹 Job alert: CISPA Helmholtz Center for Information Security is seeking Tenure-Track Faculty in all areas related to Information Security and Artificial Intelligence (f/m/d).

📍 St. Ingbert, near Saarbrücken 🇩🇪
📅 Apply by April 8
🔗 More info: bit.ly/4i12JV7

30 Charts That Show How Everything Changed in March 2020 It can be easy to forget, or look away from, the pain and disruption of the pandemic. The numbers will be there to remind us.

Decades from now, the Covid-19 pandemic will be visible in the historical data of nearly anything measurable today. Here’s an incomplete collection of charts that capture that break — across the economy, health care, education, work, family life and more.

I think one misconception about CI is that is about theory-building. It’s not. It’s about the logical tools to deduce hypothetical changes on already built world-models/theories. Neither CI nor the used models/theories (as in physics) imply the laws of nature.

Karl Pearson was a ferviente opponent of causal language because it invites metaphysics into science. Similar criticism of “subjectivity” have faced Quantum Mechanics and Bayesian Statistics. More than objectivity, utility should be valued

Internally valid RCTs are relevant, even essential, for scientific discovery. When externally valid, they provide valuable insights for policy design targeting reference populations. Science is not only about discovery: it’s also about world-model building, testing and explaining

JOB OPENING at University of Bristol:
-- Lecturer in Statistics or Machine Learning
-- (x 2 positions available)
-- application deadline: 31 March, 2025

Details at www.bristol.ac.uk/jobs/find/de...

Come join us!

I already advertised for this document when I posted it on arXiv, and later when it was published.

This week, with the agreement of the publisher, I uploaded the published version on arXiv.

Less typos, more references and additional sections including PAC-Bayes Bernstein.

arxiv.org/abs/2110.11216

Choose France - CNRS AI Rising Talents As part of the French Strategy in Artificial Intelligence (AI), the Choose France - CNRS AI Rising Talents program offers exceptionally tal

Job alert: full-time researcher position for AI with CNRS here in France. Initial contract for 5 years (tenure track). Attractive starting package and better salary than regular positions. I have been w/ @cnrs.fr since 2013, it's great!
Applications due 31/03
www.ins2i.cnrs.fr/en/cnrsinfo/...

🏹 Job alert: @eng.cam.ac.uk is seeking a Uni Asst. Professor in #MachineLearning in the Computational and Biological Learning Lab (CBL) to lead a research programme and teach.

📍 Cambridge, England
📅 Apply by 16 March 2025
🔗 Get details at bit.ly/4khlntY

It's time to Make Causality Less Sectarian Again

(Was it less sectarian before though?)

#causalsky #episky #statsky #mlsky

I make an attempt to present causality as a framework here: bit.ly/41v2ix7

Thoughts?

Hardcore causal experimentalists are weirdly religious. They renounce math and notation, and seriously think randomization alone can explain the laws of nature. Bless their hearts.

We are now accepting applications for Postdoctoral Fellows (3 years) or Research Fellows (5 years) in ICT. Deadline 2 February at 11:59 PM Finnish time
Find out more information @ shorturl.at/uMk4i #postdocjob #postdoc #artificialintelligence #datascience #bioinformatics #algorithms #AcademicSky

I am hiring in my machine learning group, DL February 2, 2025

Postdocs in Helsinki: www.aalto.fi/en/open-posi...

Topic: combinations of the following:
- foundations of (prob) ML
- beyond current distribution shift models
- AI-assistance with computationally rational user models
- appl: eng bio etc

Internship in our group at Mila in reinforcement learning + graphs for reducing energy use in buildings.

More info and submit an application by Jan 13 here:
forms.gle/TCChXnvSAHqz...

Questions? Email donna.vakalis@mila.quebec with [intern!] in the subject line.

Or in their conservative presentation:

This statement will be wrong no more than 100(1-CI)% of times.

Causal interference sounds like an interesting field though.

Looking for a great place to pursue a Ph.D. at the crossroads of data science, machine learning & mathematical optimization?

Apply to Iowa’s Business Analytics program!

For more information: tippie.uiowa.edu/phd/phd-busi...

If you would like to work with me in MIP+ML: {first}-{last}@uiowa.edu