🔥Telegram İfşa

Congrats to Yi-Lun Liao on winning 2nd place in the Open Catalysis Challenge at NeurIPS 2022! opencatalystproject.org/challenge.html

Our work "Equiformer: Equivariant Graph Attention Transformer for 3D Atomistic Graphs" has been accepted as a spotlight @ #ICLR2023! A big congrats to Yi-Lun Liao whose determination and skill made this possible! (1/3)

In Equiformer, we've replaced Transformer operations with equivariant versions including equivariant attention. Equiformer achieves competitive results on QM9, MD17 and OC20 datasets, demonstrating that Transformers can be effective models for 3D atomistic graphs! (2/3)

Equiformer Paper: openreview.net/forum?id=KwmPf… Equiformer Code: github.com/atomicarchitec… (3/3)

🧵(1/6) We propose EquiformerV2 with state-of-the-art results on all OC20 tasks and AdsorbML. Joint work with Brandon Wood , Abhishek Das from FAIR Chemistry and Prof. Tess Smidt . Besides, thank AK for tweeting before. Paper: arxiv.org/abs/2306.12059 Code: github.com/atomicarchitec…

Excited to share our latest survey paper: "Artificial Intelligence for Science in Quantum, Atomistic, and Continuum Systems"! 🚀 ArXiv: arxiv.org/abs/2307.08423.

🧵(1/6) We have updated the paper of EquiformerV2 to include more results. Currently, EquiformerV2 is the best model on the large-scale OC20 and OC22 datasets. Joint work with Brandon Wood, Abhishek Das from FAIR Chemistry and Prof. Tess Smidt. Paper: arxiv.org/abs/2306.12059 Code: