PRISM: Periodic Representation with multIscale and Similarity graph Modelling for enhanced crystal structure property prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Solé, Àlex, Mosella-Montoro, Albert, Cardona, Joan, Aravena, Daniel, Gómez-Coca, Silvia, Ruiz, Eliseo, Ruiz-Hidalgo, Javier
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911286286090240
author Solé, Àlex
Mosella-Montoro, Albert
Cardona, Joan
Aravena, Daniel
Gómez-Coca, Silvia
Ruiz, Eliseo
Ruiz-Hidalgo, Javier
author_facet Solé, Àlex
Mosella-Montoro, Albert
Cardona, Joan
Aravena, Daniel
Gómez-Coca, Silvia
Ruiz, Eliseo
Ruiz-Hidalgo, Javier
contents Crystal structures are characterised by repeating atomic patterns within unit cells across three-dimensional space, posing unique challenges for graph-based representation learning. Current methods often overlook essential periodic boundary conditions and multiscale interactions inherent to crystalline structures. In this paper, we introduce PRISM, a graph neural network framework that explicitly integrates multiscale representations and periodic feature encoding by employing a set of expert modules, each specialised in encoding distinct structural and chemical aspects of periodic systems. Extensive experiments across crystal structure-based benchmarks demonstrate that PRISM improves state-of-the-art predictive accuracy, significantly enhancing crystal property prediction.
format Preprint
id arxiv_https___arxiv_org_abs_2511_20362
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PRISM: Periodic Representation with multIscale and Similarity graph Modelling for enhanced crystal structure property prediction
Solé, Àlex
Mosella-Montoro, Albert
Cardona, Joan
Aravena, Daniel
Gómez-Coca, Silvia
Ruiz, Eliseo
Ruiz-Hidalgo, Javier
Machine Learning
Materials Science
Crystal structures are characterised by repeating atomic patterns within unit cells across three-dimensional space, posing unique challenges for graph-based representation learning. Current methods often overlook essential periodic boundary conditions and multiscale interactions inherent to crystalline structures. In this paper, we introduce PRISM, a graph neural network framework that explicitly integrates multiscale representations and periodic feature encoding by employing a set of expert modules, each specialised in encoding distinct structural and chemical aspects of periodic systems. Extensive experiments across crystal structure-based benchmarks demonstrate that PRISM improves state-of-the-art predictive accuracy, significantly enhancing crystal property prediction.
title PRISM: Periodic Representation with multIscale and Similarity graph Modelling for enhanced crystal structure property prediction
topic Machine Learning
Materials Science
url https://arxiv.org/abs/2511.20362