🤔 What if a simple initialization trick in Transformers wasn’t just an engineering hack, but a theoretically optimal solution? 🚀 Thrilled to share our latest research: "𝐎𝐧 𝐙𝐞𝐫𝐨-𝐈𝐧𝐢𝐭𝐢𝐚𝐥𝐢𝐳𝐞𝐝 𝐀𝐭𝐭𝐞𝐧𝐭𝐢𝐨𝐧: 𝐎𝐩𝐭𝐢𝐦𝐚𝐥 𝐏𝐫𝐨𝐦𝐩𝐭 𝐚𝐧𝐝 𝐆𝐚𝐭𝐢𝐧𝐠

6. Position: Graph Learning Will Lose Relevance Due To Poor Benchmarks East Exhibition Hall A-B #E-604 Thu 17 Jul 11 a.m. PDT Maya Bechler-Speicher Ben Finkelshtein Fabrizio Frasca Bryan Perozzi Michael Galkin Mathias Niepert Christopher Morris et al.

If you are in the Bay Area, consider attending our workshop, "Graph Learning Meets TCS," at the Simons Institute (simons.berkeley.edu/workshops/grap…).

This will be great newgraphperspectives.com.

🚨 New preprint: How well do universal ML potentials perform in biomolecular simulations under realistic conditions? There's growing excitement around ML potentials trained on large datasets. But do they deliver in simulations of biomolecular systems? It’s not so clear. 🧵 1/

BUT: These improvements do not consistently translate into more accurate physical observables in simulations. Densities, radial distribution functions, and conformational ensembles show inconsistent trends with model size and long-range electrostatics. 7/

Xirtam Esrevni I assume this refers to the ICTP architecture. Short answer: not quite. 1/

NNPs are becoming very accurate for gas-phase small molecules. But can current approaches scale to the condensed phase, or will scale inevitably create new challenges? Today on the blog, we review how scientists are incorporating long-range forces into NNPs.

Hey it happened again! 2 tenure-track positions at UC Chile, in the maths/stats department. We have good growing links with the dept of computational engineering too, so people wanting to fill the bridge from maths to CS are welcome. mathjobs.org/jobs/list/27012

Consider submitting a long abstract about any recent works you might have to the Molecular Machine learning conference at MIT :) moml.mit.edu MIT Jameel Clinic for AI & Health

Which foundation MLIP is best suited for your application? The MLIPX framework can help you answering this question. Checkout the Paper iopscience.iop.org/article/10.108… and github.com/basf/mlipx #MLIPX

😄

The EEML is coming to Podgorica 🇲🇪 on 8 November! Mark your calendars 🚀 Beyond excited to share that we're organising the Montenegrin ML Workshop (MMLW'25), part of EEML Workshop Series, together with Montenegrin AI Association ❤️ (Free) registration required -- please see below!

We’ve received A LOT OF submissions this year 🤯🤯 and are excited to see so much interest! To ensure high-quality review, we are looking for more dedicated reviewers. If you'd like to help, please sign up here docs.google.com/forms/d/e/1FAI…

New preprint & open-source! 🚨 “SimpleFold: Folding Proteins is Simpler than You Think” (arxiv.org/abs/2509.18480). We ask: Do protein folding models really need expensive and domain-specific modules like pair representation? We build SimpleFold, a 3B scalable folding model solely

Call for Papers –Machine Learning for Simulations in Biology & Chemistry (Simbiochem) Workshop @ EurIPS 2025 📍 Copenhagen, Denmark (Dec 6th or 7th) 📅 Submission Deadline: 10th October, 2025 👉 Submit your paper: simbiochem.com Machine learning has transformed

I just read the Equilibrium Matching (EqM) paper; it’s excellent and insightful work! Interestingly, we recently published a related method called Adaptive Equilibrium Flow Matching (AEFM). Leaving out “adaptive” reveals strong conceptual parallels between the two approaches.

Really interesting comparison between recent equilibrium flow matching and equilibrium matching papers arxiv.org/abs/2507.16521

"Equivariance matters even more at larger scales" ~ arxiv.org/abs/2510.09768 All the more reason we need scalable architectures with symmetry awareness. I know this is an obvious ask but I'm still confident that scaling and inductive bias need not be at odds. This paper

As promised after our great discussion, CHAITANYA JOSHI! Your inspiring post led to our formal rejoinder: the Platonic Transformer. What if the "Equivariance vs. Scale" debate is a false premise? Our paper shows you can have both. 📄 Preprint: arxiv.org/abs/2510.03511 1/9