Lean CNNs for mapping electron charge density fields to material properties

Published

January 27, 2025

Author(s)

Pranoy Roy, Kamal Choudhary, Surya Kalidindi

Abstract

This work introduces a lean convolutional neural network (CNN) framework, with a drastically reduced number of fittable parameters ($<$81k) compared to the benchmarks in current literature, to capture the underlying low-computational cost (i.e., surrogate) relationships between the electron charge density (ECD) fields and their associated effective properties. These lean CNNs are made possible by adding a pre-processing step (i.e., a feature engineering step) that involves the computation of the spatial correlations (specifically, 2-point spatial correlations in this work) of the ECD fields. The viability and benefits of the proposed lean CNN framework are demonstrated by establishing robust structure-property relationships involving the prediction of effective material properties using the feature-engineered ECD fields as the only input. The framework is evaluated on a dataset of crystalline cubic systems consisting of 1805 atomic structures spanning 61 different elemental species and 10 space groups across face-centered cubic (FCC), body-centered cubic (BCC), simple cubic (SC) crystal structures.

Citation

Journal of Materials Engineering

Pub Type

Journals

Download Paper

https://doi.org/10.1007/s40192-024-00389-9

Local Download

Keywords

DFT, ML, AI

Citation

Roy, P. , Choudhary, K. and Kalidindi, S. (2025), Lean CNNs for mapping electron charge density fields to material properties, Journal of Materials Engineering, [online], https://doi.org/10.1007/s40192-024-00389-9, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=958962 (Accessed October 9, 2025)