Pinned post
John Gardner
@jla_gardner
Developing #ML models and methods for chemistry, and open-source tools to accelerate them. Interning @valence_ai, prev @MSFTResearch. PhD student @VLDGroupOx
485 Following 432 Followers 211 Posts
Latest posts 5
RetroChimera has just been published in @Nature! Retrosynthesis is a hard problem (what an understatement) and so it’s extremely exciting to be making meaningful progress along the path to solving it computationally.
Custom-made molecules are advancing medicine, materials, and agriculture, but producing them is slow and expensive. A new Nature paper highlights RetroChimera, a predictive model that helps accelerate chemical synthesis, helping researchers explore a wide range of molecules.: msft.it/6012a5OgyView quoted post on X
Excited to share the pre-print we’ve been working on for the last ~4 months:
“Distillation of atomistic foundation models across architectures and chemical domains”
Deep dive thread below! 🤿🧵
“Distillation of atomistic foundation models across architectures and chemical domains”
Deep dive thread below! 🤿🧵
Extremely excited to be sharing the output of my internship in @MSFTResearch's #AIForScience team: "Understanding multi-fidelity training of machine-learned force-fields" 🤖🧪
Having played around with these new models, the Egret-1 series are extremely nice!
They're reasonably fast - I can drive short MD simulations with these on my MacBook - while still being very accurate on important bits of chemistry.
Awesome job @RowanSci 😊
They're reasonably fast - I can drive short MD simulations with these on my MacBook - while still being very accurate on important bits of chemistry.
Awesome job @RowanSci 😊
Today at @RowanSci, we're releasing Egret-1, a family of open-source neural network potentials.View quoted post on X
On lots of benchmarks, the Egret-1 models match or exceed the accuracy of quantum chemistry for organic & biomolecular simulation.
Here's me and @manntis4 with a brief video intro: