The Python Software Foundation was recommended for a $1.5M grant from the National Science Foundation. The terms of the award said PSF could not work on DEI, whether or not the grant funding was used for it. PSF therefore declined the funding. Science suffers, but commitment to core values remains

⚡ Release 0.6.2 is out ⚡ github.com/skrub-data/s...

I will speak about probabilistic regressions, @skrub-data.bsky.social and skore contributors will also present their libraries. Come join us!

scikit-learn 1.8 will be the first scikit-learn release with native extensions that are officially marked as free-threading compatible. github.com/scikit-learn...

We’re happy to announce our Social Event, taking place on Tuesday 30th September at 6pm at the Cité des sciences. A perfect opportunity to unwind and connect with fellow attendees after a day of interesting talks! pydata.org/paris2025/so... pydata.org/paris2025/ti...

Looking forward to attending PyData Paris 2025! I will give a talk about probabilistic predictions for regression problems (I need to start working on my slides ;)

👋 JupyterLab and Jupyter Notebook users: What's one thing you'd love to see improved in JupyterLab, Jupyter Notebook, or JupyterLite? The team is prepping the upcoming 4.5/7.5 releases and wants to tackle some usability issues. Drop your feedback below, this will help prioritize what gets fixed!👇

Today at #EuroScipy2025, @glemaitre58.bsky.social and I presented a tutorial on pitfalls of machine learning for imbalanced classification problems. We discussed what (not) to do when fitting a classifier and obtaining degenerate precision or recall values. probabl-ai.github.io/calibration-...

Attending the @skrub-data.bsky.social tutorial by @riccardocappuzzo.com and @glemaitre58.bsky.social at #EuroScipy2025. They introduce the new DataOps feature released in skrub 0.6. Here is the repo with the material for the tutorial: github.com/skrub-data/E...

🚨What is SOTA on tabular data, really? We are excited to announce 𝗧𝗮𝗯𝗔𝗿𝗲𝗻𝗮, a living benchmark for machine learning on IID tabular data with: 📊 an online leaderboard (submit!) 📑 carefully curated datasets 📈 strong tree-based, deep learning, and foundation models 🧵

👨‍🎓🧾✨#icml2025 Paper: TabICL, A Tabular Foundation Model for In-Context Learning on Large Data With Jingang Qu, @dholzmueller.bsky.social, and Marine Le Morvan TL;DR: a well-designed architecture and pretraining gives best tabular learner, and more scalable On top, it's 100% open source 1/9

Excited to have co-contributed the SquashingScaler, which implements the robust numerical preprocessing from RealMLP!

I got 3rd out of 691 in a tabular kaggle competition – with only neural networks! 🥉 My solution is short (48 LOC) and relatively general-purpose – I used skrub to preprocess string and date columns, and pytabkit to create an ensemble of RealMLP and TabM models. Link below👇

We are excited to announce that the schedule for PyData Paris 2025 has been released! Discover the incredible lineup of speakers we have in store for you. Don't miss out on this chance to learn, connect, and grow. Save the date and get your ticket! pydata.org/paris2025/sc...

Then why does flow matching generalize?? Because it fails! The inductive bias of the neural network prevents from perfectly learning u* and overfitting. In particular neural networks fail to learn the velocity field for two particular time values. See the paper for a finer analysis 😀

📢 Last Chance! Don't miss out! Early bird tickets are available only until the end of the day. Grab yours now and save! pydata.org/paris2025/ti...

📢 Time is Running Out! Only a few days left to grab your early bird tickets! Don't wait—this special offer ends Sunday at the end of the day. Secure yours now! pydata.org/paris2025/ti...

🚨 Just 10 Days Left! 🚨 The countdown is on — only 10 days left to grab your early bird tickets for our 2025 edition! Lock in your spot at the best price before rates go up. pydata.org/paris2025/ti...

The week of September 29th, Paris will become the epicenter of #opensource scientific computing, with a great series of events. This rare alignment creates the perfect opportunity to visit and join a vibrant community of developers, maintainers, and users! Check this out (links in thread) ⬇️

There were layoffs at MS yesterday and 3 #Python core devs from the Faster CPython team were caught in them. If you know of any jobs, please send them their way: Eric Snow: www.linkedin.com/in/ericsnowc... Irit Katriel: www.linkedin.com/in/irit-katr... Mark Shannon: www.linkedin.com/in/mark-shan...

The ‘What’s New’ page for Python 3.14 now includes updates on the free-threaded (GIL-disabled) build! docs.python.org/3.14/whatsne...

I don't get any benefit from citations but an inline or footnote link is always incredibly appreciated. If you are limited to bibliography citations then the standard website format is great, "Betancourt, Michael (YEAR ACCESSED). CHAPTER NAME. Retrieved from GITHUB LINK, commit COMMIT NUMBER"

#GitHub mini tip: Use the GitHub CLI command `gh issue develop -c` to create and checkout a local branch linked to the given issue. No need to name the branch, and when you later run `gh pr create`, the pull request will be linked to the issue automatically.

Great News! We’re extending the CfP deadline to Sunday, 27th April. You have now an additional 2 weeks to submit your proposals. Take this extra time to refine your ideas and finalise your submissions ! pydata.org/paris2025/cfp

Finally a new #Shapely feature release! 🎉 Shapely 2.1.0 highlights include initial support for geometries with M or ZM values, functionality for coverage validation and simplification, and much more. For a full overview, see shapely.readthedocs.io/en/latest/re... #python #geopython #geospatial

🚀⚡ Release: 0.5.3 Check out the release notes: skrub-data.org/stable/CHANG... Highlights below ⤵️

Call for sponsors! Join us and be part of one of the best open-source events of the year. Sponsoring will showcase your support to the community and allow you to reach a wide audience of tech and data enthusiasts from academia and the industry. pydata.org/paris2025/sponsorship-opportunites#tiers

What an achievement: * All official Debian bookworm live images rebuild reproducibly lists.reproducible-builds.org/pipermail/rb...

GitHub Actions now supports free-threaded Python! I wrote up how to add it your workflows so you can start testing free-threaded Python 3.13 and 3.14 with either actions/setup-python or actions/setup-uv. hugovk.dev/blog/2025/fr... #Python #FreeThreaded #GitHub #GitHubActions #CI #testing

🔥🔥🔥 CV Folks, I have some news! We're organizing a 1-day meeting in center Paris on June 6th before CVPR called CVPR@Paris (similar as NeurIPS@Paris) 🥐🍾🥖🍷 Registration is open (it's free) with priority given to authors of accepted papers: cvprinparis.github.io/CVPR2025InPa... Big 🧵👇 with details!

🎓Paper time!✨ #ICLR spotlight. Concluding of 5 years of research on missing values handling for prediction: Beware of diminishing returns in imputation for prediction. 1/8

we released olmo 32b today! ☺️ 🐟our largest & best fully open model to-date 🐠right up there w similar size weights-only models from big companies on popular benchmarks 🐡but we used way less compute & all our data, ckpts, code, recipe are free & open made a nice plot of our post-trained results!✌️

Two big LLM releases today - Cohere's Command A simonwillison.net/2025/Mar/13/... and Ai2's OLMo 2 OLMo claims to be "the first fully-open model (all data, code, weights, and details are freely available) to outperform GPT3.5-Turbo and GPT-4o mini", which feels notable allenai.org/blog/olmo2-32B

Loky 3.5.0 is out! Loky provides an extended version of Python's `concurrent.futures.ProcessPoolExecutor` that leverages cloudpickle to work within interactive Jupyter sessions on all platforms and reuse existing workers to hide the overhead of starting new workers each time.

I have the intuition that TabPFN approximates amortized Bayesian inference with a Solomonoff prior via in-context learning. Perplexity agrees :) www.perplexity.ai/search/is-it... I wonder if this theoretical "universality" is one of the reasons for its empirical success.

🎓📑✨ Preprint ( #AIstats ): survival analysis and competing risks with scalable stochastic solvers that recovers well outcome probabilities. TL;DR: using stochastic optimization, our approach gives survival models both faster and predicting better event probabilities 🧵 1/7

Recently merged in scikit-learn's main branch: display the maximum predicted class probability in 2D continuous feature spaces (mostly for didactic purposes): scikit-learn.org/dev/auto_exa... The linked example has been updated to include some conclusions we can draw from this plot.

It has been great to make these under-the-hood improvements for speed and efficiency. Binder is becoming both more performant and more cost effective! You should notice this on mybinder.org as well as any community-deployed BinderHubs.

The call for proposals for PyData Paris 2025 is now open! Don't delay, get your submissions early. CfP Deadline: Sunday 13th April 2025 📍 Cité des sciences et de l'industrie 📆 September 30th and October 1st. pydata.org/paris2025/cfp

📢PSA: #NeurIPS2024 recordings are now publicly available! The workshops always have tons of interesting things on at once, so the FOMO is real😵‍💫 Luckily it's all recorded, so I've been catching up on what I missed. Thread below with some personal highlights🧵

📢We are excited to announce the keynote speakers for our 2025 conference: - Alenka Frim from United.Cloud - Ralf Gommers from @quansight.com - Anita Graser from the Austrian Institute of Technology Join us at PyData Paris at Cité des Sciences from Sep 30 to Oct 1, 2025. medium.com/@PyDataParis...

Quick TIL: if you have uv installed and you want to try out ANY Python package in an ipython REPL, you can do so just by running uvx --with llm --with sqlite-utils ipython Pass as many --with package options as you like I wrapped it in a shell script, "itry llm" til.simonwillison.net/python/itry

🚀 Open-R1's latest: OpenR1-Math-220k dataset with 220k verified math problems, running on 512 H100s! The Qwen-7B model nears DeepSeek's performance, while community finds quality reasoning possible with just ~1000 samples. Check it out: https://huggingface.co/blog/open-r1/update-2

We've built a simulated driving agent that we trained on 1.6 billion km of driving with no human data. It is SOTA on every planning benchmark we tried. In self-play, it goes 20 years between collisions.

SmolLM2 paper by @loubnabnl.hf.co A detailed guide for building SOTA small LMs. While most LM papers skim over data, they show how data is the secret sauce behind strong performance.

It is hard to overstate how cool and powerful is flex attention. @chhillee.bsky.social pytorch.org/blog/flexatten… TL;DR: it is an implementation of the attention operator in pytorch that allows in particular to efficiently "carve" the attention matrix. 1/3

It’s Em-el-el-ememememmm … 😂

Early stopping on validation loss? This leads to suboptimal calibration and refinement errors—but you can do better! With @dholzmueller.bsky.social, Michael I. Jordan, and @bachfrancis.bsky.social, we propose a method that integrates with any model and boosts classification performance across tasks.

Today, we are publishing the first-ever International AI Safety Report, backed by 30 countries and the OECD, UN, and EU. It summarises the state of the science on AI capabilities and risks, and how to mitigate those risks. 🧵 Full Report: assets.publishing.service.gov.uk/media/679a0c... 1/21