@genbio_workshop @LucasPinede @junonam_ @RGBLabMIT @sihyun_yu Unfortunately both @LucasPinede and I won't be at ICML in person this year. But our poster will be at the workshop 😂 Feel free to DM us or email our lead author (lucasp at about the paper! Preprint coming very soon, please tune in! (4/n, N=4).

@genbio_workshop @LucasPinede @junonam_ @RGBLabMIT At low noise, emulator scores ≈ physical forces. Even w/o alignment loss, emulator embeddings drift toward MLIP's. Inspired by @sihyun_yu’s REPA, we add a loss to encourage this — and training becomes faster & better. (3/n)

@genbio_workshop @LucasPinede @junonam_ @RGBLabMIT Boltzmann emulators can sample conformations cheaper than MD but are expensive to train. They need equilibrium data, which is very hard to get. But… MLIPs like MACE are trained on tons of noneq. data. What if we transfer knowledge from force models to generative models? (2/n).

🚀 Come check our poster at ICML @genbio_workshop!.We show that pretrained MLIPs can accelerate training of Boltzmann emulators — by aligning their internal representations. Coauthors @LucasPinede, @junonam_, @RGBLabMIT (1/n)

RT @satyanadella: Understanding protein motion is essential to understanding biology and advancing drug discovery. Today we’re introducing….

Happy to have contributed to this awesome project!.

RT @FrankNoeBerlin: Huge shoutout to the outstanding team: Sarah Lewis, Tim @tmhmpl, José @josejimlun, Michael Gastegger, Yu @YuuuXie, Andr….

RT @FrankNoeBerlin: BioEmu now published in @ScienceMagazine !! . What is BioEmu? Check out this video: .

RT @jacksonnn18: Our review paper is now live! 🚀.We dive into how in-situ cryoET allows us to not only view proteins inside their native ce….

RT @ChoYehlin: 🚀 Excited to release BoltzDesign1!. ✨ Now with LogMD-based trajectory visualization. 🔗 Demo: Feedbac….

RT @ChoYehlin: I'll be attending @iclr_conf and the @gembioworkshop! 🧠✨If you're around and want to meet up or chat, feel free to DM me! 😊.

RT @HannesStaerk: New paper (and #ICLR2025 Oral :)): .ProtComposer: Compositional Protein Structure Generation with 3D Ellipsoids https://t….

RT @jonkhler: Spending the Easter days prepping for #ICLR2025 in #singapore 😍🤩. If you want to meet/catch up on @cusp_ai #AI4Science #Machi….

Excited for #ICLR2025! I'll be there 4/25~ to organize the GEM Bio Workshop @gembioworkshop to bridge experiments and ML in bio 🥰 If you want to meet up on ML for proteins / dynamics / simulation, DM me! . I'm curious to learn about academic/industry opportunities out there!.

RT @ask1729: 1/ Machine learning force fields are hot right now 🔥: models are getting bigger + being trained on more data. But how do we ba….

RT @MSFTResearch: Meet BioEmu-1 from Microsoft Research. This deep learning model can generate thousands of protein structures per hour, un….

RT @FrankNoeBerlin: The BioEmu-1 model and inference code are now public under MIT license!!! . Please go ahead, play with it and let us kn….

RT @MSFTResearch: Microsoft researchers introduce MatterGen, a model that can discover new materials tailored to specific needs—like effici….

RT @HannesStaerk: Awesome paper!.BioEmu from MSR: Scalable emulation of protein equilibrium ensembles with generative deep learning https:/….

RT @DhimanRay16: Just out in @JChemPhys! Integrating Weighted Ensemble (@WestpaSoftware) with OPES-Flooding for accurate and efficient calc….