Scaling Laws in Particle Physics Data! This is a result I've been itching to share and it's finally out. One of the big open questions is how much better AI-based methods at particle colliders can still become. 1/4
Congrats Shubhendu!
We're looking for a few profiles:
1. generative models (come work with me!): jobs.ashbyhq.com/cuspai/b8108...
2. molecular simulation: jobs.ashbyhq.com/cuspai/3bb2f...
3. materials foundation models: jobs.ashbyhq.com/cuspai/90a8f...
4. semiconductors: jobs.ashbyhq.com/cuspai/f6a81...
2/2
Happy New Year!
Do your plans for 2026 include...
- working with a great team lead by @wellingmax.bsky.social and @aronwalsh.github.io?
- living in Amsterdam, Berlin, London, or Cambridge?
- using fun tools from ML and material science?
- solving important problems?
Then join us at CuspAI!
1/2
Or join our semiconductor team!
Still plenty of ML (especially generative models) in this role, but with a focus on semiconductor modelling and design.
jobs.ashbyhq.com/cuspai/f6a81...
Same great team, but working directly w/ Aron Walsh in London.
Happy to chat!
2/2
Come work with us at @cuspai.bsky.social in the generative model team!
Excited about flow / diffusion models and chemistry? Looking for impact?
jobs.ashbyhq.com/cuspai/b8108...
Join a great team lead by @wellingmax.bsky.social and Aron Walsh, work in Amsterdam / Cambridge / London / Berlin.
1/2
Today at NeurIPS (SD), come meet the @cuspai.bsky.social team and learn about our work!
Find us at 5:30pm at booth 1343 (Renaissance Philanthropy / UK government)
As we go into the Thanksgiving holiday, I wanted to express my thanks to my collaborators @johannbrehmer.bsky.social @glouppe.bsky.social, Juan Pavez, @smsharma.bsky.social. Recently, I was awarded the Pritzker Prize for AI in Science for work on SBI. That wouldn't have never happened without them.
I'll be at NeurIPS next week – together with my @cuspai.bsky.social colleagues @jonkhler.argmin.xyz, @hannahopenshaw.bsky.social, Friso de Kruiff, and @wellingmax.bsky.social.
If you'd like to chat about ML for material discovery, generative models, or start-ups made in Europe, ping me!
It was great to work on the ODAC25 paper with the Meta FAIR Chemistry and Georgia Tech. A leap forwards in modelling direct air carbon capture with metal organic frameworks, with much better data and larger models.
Paper: arxiv.org/abs/2508.03162
Data and models: huggingface.co/facebook/ODA...
Are you tired of context-switching between coding models in @pytorch.org and paper writing on @overleaf.com?
Well, I’ve got the fix for you, Neuralatex! An ML library written in pure Latex!
neuralatex.com
To appear in Sigbovik (subject to rigorous review process)
An unexpected surprise. The 2025 Breakthrough Prize in Fundamental Physics honors over 13,000 researchers whose labors have led to the precise description the Higgs mechanism, … breakthroughprize.org/News/91 @CERN
On March 7th, we’re Standing Up for Science—
and against political censorship, autocracy, and fascism.
Science stands at a crossroads. This is a wider fight for truth, for democracy, and for the future.
We hope you join us.
www.standupforscience2025.org
📣 Hiring! I am looking for PhD/postdoc candidates to work on foundation models for science at @ULiege, with a special focus on weather and climate systems. 🌏 Three positions are open around deep learning, physics-informed FMs and inverse problems with FMs.
Thanks a lot, Guillaume!
Excellent talk by @johannbrehmer.bsky.social
On “Does equivariance matter at scale?” At NeurReps workshop
arxiv.org/abs/2410.23179
www.neurreps.org
If you're in Vancouver and want to chat about these papers, material discovery, or anything else, come by the posters or ping me!
6/6
You might not be surprised to hear that equivariance improves data efficiency.
But did you expect equivariant models to also be more *compute*-efficient? Learning symmetries from data costs FLOPs!
arxiv.org/abs/2410.23179
With Sönke Behrends, @pimdh.bsky.social, and @taco-cohen.bsky.social.
5/6
On Saturday at 11:00 at the @neurreps.bsky.social workshop, I'll talk about our investigation into the relevance of equivariance at scale.
We studied empirically how equivariant and non-equivariant architectures scale as a function of training data, model size, and training steps.
4/6
Combining L-GATr with Riemannian flow matching, they also constructed the first Lorentz-equivariant generative model.
arxiv.org/abs/2405.14806
With @jonasspinner.bsky.social, Victor Bresó, @pimdh.bsky.social, Tilman Plehn, and Jesse Thaler.
3/6
On Thursday from 11:00 to 14:00, I'll be cheering on @jonasspinner.bsky.social and Victor Bresó at poster 3911.
They built L-GATr 🐊: a transformer that's equivariant to the Lorentz symmetry of special relativity. It performs remarkably well across different tasks in high-energy physics.
2/6
Just arrived in Vancouver for #NeurIPS.
I'm looking forward to meeting old and new friends, learning a thing or two, and presenting some recent work:
1/6
A common question nowadays: Which is better, diffusion or flow matching? 🤔
Our answer: They’re two sides of the same coin. We wrote a blog post to show how diffusion models and Gaussian flow matching are equivalent. That’s great: It means you can use them interchangeably.
The sbi package is growing into a community project 🌍 To reflect this and the many algorithms, neural nets, and diagnostics that have been added since its initial release, we have written a new software paper 📝 Check it out, and reach out if you want to get involved: arxiv.org/abs/2411.17337
Thrilled to announce that L-GATr is going to NeurIPS 2024! Plus, there is a new preprint with extended experiments and a more detailed explanation.
Code: github.com/heidelberg-h...
Physics paper: arxiv.org/abs/2411.00446
CS paper: arxiv.org/abs/2405.14806
1/7
Screenshot of Google scholar
Milestone: our review paper “The Frontier of Simulation-Based Inference” coauthored with @glouppe.bsky.social & @johannbrehmer.bsky.social hit 1000 citations. I’m very excited about the potential for these methods to transform science!
www.pnas.org/doi/10.1073/...
simulation-based-inference.org
The snow is gently falling outside the window, the models are training, what could be better? Two articles cool to read:
Does Equivariance matter at scale? (@johannbrehmer.bsky.social et al.) arxiv.org/abs/2410.23179
Denoising Diffusion Bridge Models (Linqi Zhou et al.) arxiv.org/pdf/2309.16948
Would love to be on the list as well, thanks!
Oops, sorry and thanks!
It's been a while, but I still like SBI — can you add me, too?
