EquiformerV2 was accepted at #ICLR2024! EquiformerV2 is the result of (Yi-Lun's) careful empirical analysis of how to incorporate efficient equivariant convolutions into transformers, resulting in an architecture that scales surprisingly well across atomic datasets.

Paper shared on arXiv showing catalyst AI/ML models trained on datasets like FAIR Chemistry can generalize to solid solutions like high entropy alloys (HEA)! This is exciting because the design space of HEAs (with >5 components) is combinatorially large. arxiv.org/abs/2403.09811

We are excited to share CatTSunami: arxiv.org/abs/2405.02078! We show that a zero-shot application of machine learning models trained only on OC20 can calculate a transition state with 91% accuracy in less than a minute.

Super excited to release a set of models for computational chemistry - my last 2 years of work Orbital Materials. Post ELMo, I'm at some risk of becoming a "one trick pony" career wise, but we've managed to make pre-training work nicely for 3d crystal structures. More below!