🎉Personal update: I'm thrilled to announce that I'm joining Imperial College London @imperialcollege as an Assistant Professor of Computing @ICComputing starting January 2026. My future lab and I will continue to work on building better Generative Models 🤖, the hardest

Flow Maps are all the rage these days. In this new work we generalize them to Riemannian manifolds. SOTA results 🚀, few step inference 🔥, and generalizes many few step generative models on manifolds in one slick framework 😎. Great work with even greater collaborators

@huijiezh really nice work! i really like the \alpha-flow idea! your approach seems quite similar to our recent framework for flow maps, which you may also find interesting ( https://t.co/QBp1kELVhF) - see also the pinned post on my page, which includes very similar diagrams to yours here.

This is as close to a religious text that I would read. So comprehensive and well done!

Tired to go back to the original papers again and again? Our monograph: a systematic and fundamental recipe you can rely on! 📘 We’re excited to release 《The Principles of Diffusion Models》— with @DrYangSong, @gimdong58085414, @mittu1204, and @StefanoErmon. It traces the core

Excited to share Pearl from Genesis Molecular AI (yes, we've updated our name!): the first co-folding model to clearly surpass AlphaFold 3 on protein-ligand structure prediction. Unlike LLMs that train on vast public data, drug discovery AI faces fundamental data scarcity. Our

🚨🌶️ Did you realise you can get alignment `training’ data out of open weights models? Oops We show that models will regurgitate alignment data that is (semantically) memorised. This data can come from SFT and RL... and can be used to train your own models! 🧵

As promised after our great discussion, @chaitanyakjoshi! 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: https://t.co/kd8MFiOmuG 1/9

As the IMM paper came out in March, I implemented it myself for some project, before the true source code was made available. I am releasing my version now: https://t.co/VXOBw91GXk It contains most/all features, and should be easy to (re-)use! Hope someone finds it helpful 🙂

Excited to present our work on dense retrieval at COLM 2025! Enter BiXSE: Improving Dense Retrieval via Probabilistic Graded Relevance Distillation! We show how to train with a simple, point-wise, binary cross-entropy loss on LLM-graded data and outperform InfoNCE!

Really cool work that unifies many threads on one step generative models. This is now my go to model family.

(1/7) New paper!🚀 https://t.co/dq6yEzWyHg ✅Boltzmann distribution sampling for peptides up to 8 residues ✅4.3ms of training MD trajectories ✅Open-source codebase With @charliebtan, @leonklein26, Saifuddin Syed, @dom_beaini @mmbronstein @AlexanderTong7 @k_neklyudov Read

🔉 New paper on training better Diffusion Language Models that plan at inference time! Great work led by @pengzhangzhi1 and Zack B.!!

The @EEMLcommunity 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 @aisocietyme ❤️ (Free) registration required -- please see below!

🎉 Congrats to @Schmidt_Center postdoctoral fellow @lazar_atan and colleagues on their paper acceptance to @NeurIPSConf 2025! CurlyFM introduces Curly Flow Matching, a new way to model non-gradient field dynamics, capturing complex, periodic behaviors missed by current methods.

Curly-FM accepted to #NeurIPS2025! 🌊 See everyone in San Diego! 🌊

🌊Now coming to a #NeurIPS2025 near you.

The most important skill for a researcher is not technical ability. It's taste. The ability to identify interesting and tractable problems, and recognize important ideas when they show up. This can't be taught directly. It's cultivated through curiosity and broad reading.

🎉3 papers including 1 spotlight to #NeurIPS2025 . Congrats to all my co-authors 👏 Sadly 2 very good papers didn't make it this time (including 1 best paper at a workshop). We will fight for 🇧🇷 On a silly but personal note, paper acceptance streak ended at 21 😭.

I am thrilled to announce that our work on the generalization of flow matching has been accepted to NeurIPS as an oral!! See you in San Diego 😎

I’m recruiting fully-funded PhD students (Fall 2026) to join my new group @EmoryCS! 🎓 We’ll work on trustworthy, impactful AI at the intersection of NLP, AI Safety, Human-Centered AI & AI4Health. More details at 👉 https://t.co/q0Td4rDABR (Picture is with some incredible people