Learning lattice parameters from powder X-ray diffraction data using invariants

We introduce a smooth, invariant representation of the reciprocal lattice which avoids the discontinuities of unit-cell parameterizations, and show it improves machine-learned lattice parameter prediction from powder X-ray diffraction data.

Towards quantum crystallographic description of incommensurately modulated structures

A modification of the structure-factor equation for incommensurately modulated structures is proposed to enable structure description with non-spherical atomic form factors.