Jigyasa Nigam

Classical force fields can now learn both bonded and nonbonded parameters consistently from first principles, achieving ab initio accuracy without empirical nonbonded terms.

Remarkably, even incomplete or ill-conditioned descriptors can yield accurate atomic reconstructions, challenging the conventional wisdom about the necessity of high-dimensional feature sets.

Ditch the geometry-to-property map: this work uses the external potential as the primary input for machine learning models, unlocking a scalable and equivariant approach to predicting electronic structure.