Caltech's O-kRystal Learns Band Structure Corrections Beyond DFT Accuracy

Caltech's Anima Anandkumar team introduced O-kRystal, a periodic E(3)-equivariant graph neural network with dual-space orbital learning that learns corrections to individual electronic states, surpassing standard DFT accuracy with a 68.7% reduction in MAE for band structure prediction.

2026-10-18 ~ 2026-10-18 · 2 related posts

• Caltech's O-kRystal cuts PBE band energy MAE 68.7%, learning beyond-DFT corrections — AnimaAnandkumar · 2026-10-18

1 near-duplicate retellings: AnimaAnandkumar