Public articles linked to the same research event.
arXiv The authors introduce OrbFlow, an SE(3)-equivariant generative model that learns a probability path over the full GTO coefficient space via flow matching and is trained with a two-phase trajectory curriculum that mitigates discretization drift during numerical integration; it reduces density error on QM9 by 13.6% relative to the previous best model, reduces error by 51% to 63% on every molecule of the MD benchmark relative to the strongest prior method sharing its basis, cuts SCF iterations by up to 68%, transfers zero-shot to unseen exchange-correlation functionals, and recovers dipole and quadrupole moments to within a few percent of DFT references without any SCF calculation.
The authors introduce OrbFlow, an SE(3)-equivariant generative model that learns a probability path over the full GTO coefficient space via flow matching and is trained with a two-phase trajectory curriculum that mitigates discretization drift during numerical integration; it reduces density error on QM9 by 13.6% relative to the previous best model, reduces error by 51% to 63% on every molecule of the MD benchmark relative to the strongest prior method sharing its basis, cuts SCF iterations by up to 68%, transfers zero-shot to unseen exchange-correlation functionals, and recovers dipole and quadrupole moments to within a few percent of DFT references without any SCF calculation.
The authors introduce OrbFlow, an SE(3)-equivariant generative model that learns a probability path over the full GTO coefficient space via flow matching and is trained with a two-phase trajectory curriculum that mitigates discretization drift during numerical integration; it reduces density error on QM9 by 13.6% relative to the previous best model, reduces error by 51% to 63% on every molecule of the MD benchmark relative to the strongest prior method sharing its basis, cuts SCF iterations by up to 68%, transfers zero-shot to unseen exchange-correlation functionals, and recovers dipole and quadrupole moments to within a few percent of DFT references without any SCF calculation.
The authors introduce OrbFlow, an SE(3)-equivariant generative model that learns a probability path over the full GTO coefficient space via flow matching and is trained with a two-phase trajectory curriculum that mitigates discretization drift during numerical integration; it reduces density error on QM9 by 13.6% relative to the previous best model, reduces error by 51% to 63% on every molecule of the MD benchmark relative to the strongest prior method sharing its basis, cuts SCF iterations by up to 68%, transfers zero-shot to unseen exchange-correlation functionals, and recovers dipole and quadrupole moments to within a few percent of DFT references without any SCF calculation.