Soon @ @icmlconf ! 🌪️Normalizing flows for rigid bodies and the rotation group SO(3)! 🌪️ We found a way to design smooth normalizing flows for rigid body systems, like ice 👇 Joint work with amazing @inve_michele @pimdehaan and @FrankNoeBerlin 1/ https://t.co/FJIGsajwrb

I decided that Twitter is becoming more and more useless for anything professional/science related. As such I will start to include more shit posting and recreational stuff from now on. Hope it's not taken as an offense by all the serious scientists following me so far :)

😅🌹✊

We’ve identified a “Collaboration Gap” in today’s top AI models. Testing 32 leading LMs on our novel maze-solving benchmark, we found that models that excel solo can see their performance *collapse* when required to collaborate – even with an identical copy of themselves. A \🧵

Yet another NequIP in the top 10 with EquFlash, this time with some very clever accelerations! Bringing the total to ….? I’ll leave it to you how to count :) One question this raises is what a lot of folks have told me recently both on here and in private: they find it

Context:

Bit funny that a) this is something new for US twitter (reminds me of the electric kettles some years ago) but also b) since when is Germany a Slavic country? 😅 (not that I would mind but the label feels oddly wrong!)

Job postings here :) https://t.co/u4las635t3

Super interesting work!

meine liebste grammatikalische Konstruktion ist das „politische Wir“, das in Sätzen wie „wir müssen wehrfähig werden“ alle meint außer die sprechende Person, aber in Sätze wie „wir müssen unseren Wohlstand bewahren“ nur die sprechende Person, sehr interessant

"Equivariance matters even more at larger scales" ~ https://t.co/hSdOxIP3UD 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

🤦‍♂️

just as a clarification before someone rightfully complains: not involved in AF2 itself! just fishing in the same pond at the same time before things were cool :D

so after AF2 again working in a very hot field 😅

Reinforcement Learning (RL) has long been the dominant method for fine-tuning, powering many state-of-the-art LLMs. Methods like PPO and GRPO explore in action space. But can we instead explore directly in parameter space? YES we can. We propose a scalable framework for

Hiring for BioEmu project @MSFTresearch AI for Science. Berlin DE or Cambridge UK. https://t.co/M4cNZrVo49 - #MachineLearning Researcher https://t.co/19D9MpdeCY - Data Engineer https://t.co/72S24Pa5Yq - Applied Researcher https://t.co/cP3jwgOCoi - Principal Applied Researcher

predictions are hard, especially about the future... 🤷🏼‍♂️

Sometimes the simple things in life are just good. ❤️

Call for blog posts here:

More good news! 🎉 Just like in past years, accepted blog posts can also be presented as posters at @iclr_conf 🇧🇷 #ICLR2026 #MachineLearning

Good news should be shared among friends - the blog post track of #ICLR2026 is coming again! We're looking forward to your submissions :) #MachineLearning #ICLR