Response-state Learning for Transferable Vibrational Spectroscopic Characterization with Electron PriorNew

This arXiv paper introduces SENK (SO(3) Equivariant Neural Kalman Networks), a novel approach for transferable vibrational spectroscopic characterization using electron priors. The method combines an equivariant transformer backbone for Hessian, dipole-derivative, and polarizability-derivative learning with an Equivariant Neural Kalman filter to form a response-state cascade. It addresses challenges where localized stereoelectronic environments perturb intermediate response states and high-risk response units dominate characteristic spectral fingerprints.