Thank you to all who joined us @ #MoML2024! Here's to another year of record-breaking attendance & incredible posters! Thanks to J&J, Takeda, Genesis Therapeutics, & Dimension for making MoML possible! Congrats again to Soojung Yang & Juno Nam on winning the Best Paper Prize!

Check out new work arxiv.org/abs/2411.07233 by Alexandra, Agnish, Aditya and Cal on generative diffusion but with correlated or ``active" noise.

Today we are excited to welcome Roberto Covino to give this months Chalmers AI4Science seminar. Join us in Analysen on the Chalmers Johanneberg Campus this afternoon at 3pm or on zoom. For more details see psolsson.github.io/AI4ScienceSemi…

We often think of an "equilibrium" as something standing still, like a scale in perfect balance. But many equilibria are dynamic, like a flowing river which is never changing—yet never standing still. These dynamic equilibria are nicely described by so-called "detailed balance"

Thrilled to announce Boltz-1, the first open-source and commercially available model to achieve AlphaFold3-level accuracy on biomolecular structure prediction! An exciting collaboration with Jeremy Wohlwend, Saro and an amazing team at MIT and Genesis Therapeutics. A thread!

New Paper Alert! "Thermodynamic Interpolation: A generative approach to molecular thermodynamics and kinetics" introduces Thermodynamic Interpolation (TI) for generating and transforming equilibrium statistics with temperature control! 🌡️ led by Selma Moqvist and Weilong Chen

Zero-shot extrapolation for out-of-distribution (OOD) chemical property prediction is an important step towards high-performance materials discovery. Check out our spotlight at the #NeurIPS AI for Accelerated Materials Design Workshop! openreview.net/pdf?id=HkfnueE…

Super excited to preprint our work on developing a Biomolecular Emulator (BioEmu): Scalable emulation of protein equilibrium ensembles with generative deep learning from Microsoft Research AI for Science. #ML #AI #NeuralNetworks #Biology #AI4Science biorxiv.org/content/10.110…

Our paper has been selected as a spotlight at #NeurIPS2024 AI for Materials. We uncover why generative models are well-suited for materials synthesis prediction and propose a diffusion-based approach, When/Where: Sat 14 Dec 8:15a, West 211-214 openreview.net/forum?id=hy39q…

📢New preprint out! We constrain the molecular generation space to follow the "symmetry" of patented molecules that are likely to be synthesizable. Achieved with "symmetry-aware" fragment decomposition, and a constrained Monte Carlo Tree Search generator. arxiv.org/abs/2410.08833

Excited to be at #NeurIPS2024 🚀 I will share prelim results: Improving long-term rollout of neural operators with flow matching-inspired correction ml4physicalsciences.github.io/2024/files/Neu… Learning PDEs (for frontal polymerization) with differentiable simulations openreview.net/pdf?id=5tzkzJ2… 1/2

📢Muratahan Aykol, Ekin Dogus Cubuk and colleagues from Google DeepMind introduce a computational approach to predict the most likely crystallization products from amorphous precursors, which has the potential to help with the synthesis of new materials. nature.com/articles/s4358…

Thrilled to share my recent work with Sergey Ovchinnikov, Gabriel Rocklin, Kotaro Tsuboyama, and Justas Dauparas: biorxiv.org/content/10.110… (1/9)

📢Out now! Hao Tang and colleagues from DMSE at MIT and MIT Nuclear Science and Engineering introduce MEHnet, a deep learning method for molecular electronic structures that can predict a host of molecular properties. nature.com/articles/s4358… 🔓rdcu.be/d4Zd5

🧵(1/5) Have you ever wanted to combine different pre-trained diffusion models but don't have time or data to retrain a new, bigger model? 🚀 Introducing SuperDiff 🦹‍♀️ – a principled method for efficiently combining multiple pre-trained diffusion models solely during inference!

Amongst many recent discrete diffusion, I found DDPD very interesting. Its unique in a way it naturally decomposes the task into that planner and denoiser (so I implemented minDDPD for imagenet)

In DDPD, planner decides which tokens to denoise, and denoiser decides what to replace it with. Model's knowledge is decomposed to guessing which part is incoherent and how its incoherent. Left is planner's prediction on 'whats wrong'. Right is denoising state. You can see its

I wrote a thing about "RL or control as Bayesian inference", which encompasses - RLHF and controlled generation in LLMs - Finetuning or guidance in diffusion models - Diffusion samplers from general unnormalized densities - Sequential Monte Carlo sampling for all of the above

A General Framework for Inference-time Scaling and Steering of Diffusion Models Introduces Feynman-Kac steering, an inference-time steering framework for sampling diffusion models guided by a reward function. It generates multiple samples (particles) like best-of-n (importance

Microsoft researchers introduce MatterGen, a model that can discover new materials tailored to specific needs—like efficient solar cells or CO2 recycling—advancing progress beyond trial-and-error experiments. msft.it/6012U8zX8