A Machine Learning Inversion Scheme for Determining Interaction from Scattering

TL;DR

This paper presents a machine learning approach to infer effective interactions in condensed matter from scattering data, outperforming traditional methods in accuracy and efficiency, and applicable to highly correlated systems.

Contribution

Introduces a novel machine learning inversion scheme that probabilistically infers effective potentials from scattering spectra without restrictive model assumptions.

Findings

Successfully inferred effective potentials in colloidal suspensions

Outperformed existing parametric methods in accuracy and efficiency

Applicable to highly correlated condensed matter systems

Abstract

We outline a machine learning strategy for determining the effective interaction in the condensed phases of matter using scattering. Via a case study of colloidal suspensions, we showed that the effective potential can be probabilistically inferred from the scattering spectra without any restriction imposed by model assumptions. Comparisons to existing parametric approaches demonstrate the superior performance of this method in accuracy, efficiency, and applicability. This method can effectively enable quantification of interaction in highly correlated systems using scattering and diffraction experiments.