Learning Thermoelectric Transport from Crystal Structures via Multiscale Graph Neural Network

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zeng, Yuxuan, Cao, Wei, Zuo, Yijing, Lyu, Fang, Xie, Wenhao, Peng, Tan, Hou, Yue, Miao, Ling, Wang, Ziyu, Shi, Jing
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918426979598336
author Zeng, Yuxuan
Cao, Wei
Zuo, Yijing
Lyu, Fang
Xie, Wenhao
Peng, Tan
Hou, Yue
Miao, Ling
Wang, Ziyu
Shi, Jing
author_facet Zeng, Yuxuan
Cao, Wei
Zuo, Yijing
Lyu, Fang
Xie, Wenhao
Peng, Tan
Hou, Yue
Miao, Ling
Wang, Ziyu
Shi, Jing
contents Graph neural networks (GNNs) are designed to extract latent patterns from graph-structured data, making them particularly well suited for crystal representation learning. Here, we propose a GNN model tailored for estimating electronic transport coefficients in inorganic thermoelectric crystals. The model encodes crystal structures and physicochemical properties in a multiscale manner, encompassing global, atomic, bond, and angular levels. It achieves state-of-the-art performance on benchmark datasets with remarkable extrapolative capability. By combining the proposed GNN with \textit{ab initio} calculations, we successfully identify compounds exhibiting outstanding electronic transport properties and further perform interpretability analyses from both global and atomic perspectives, tracing the origins of their distinct transport behaviors. Interestingly, the decision process of the model naturally reveals underlying physical patterns, offering new insights into computer-assisted materials design.
format Preprint
id arxiv_https___arxiv_org_abs_2512_06697
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Learning Thermoelectric Transport from Crystal Structures via Multiscale Graph Neural Network
Zeng, Yuxuan
Cao, Wei
Zuo, Yijing
Lyu, Fang
Xie, Wenhao
Peng, Tan
Hou, Yue
Miao, Ling
Wang, Ziyu
Shi, Jing
Materials Science
Applied Physics
Graph neural networks (GNNs) are designed to extract latent patterns from graph-structured data, making them particularly well suited for crystal representation learning. Here, we propose a GNN model tailored for estimating electronic transport coefficients in inorganic thermoelectric crystals. The model encodes crystal structures and physicochemical properties in a multiscale manner, encompassing global, atomic, bond, and angular levels. It achieves state-of-the-art performance on benchmark datasets with remarkable extrapolative capability. By combining the proposed GNN with \textit{ab initio} calculations, we successfully identify compounds exhibiting outstanding electronic transport properties and further perform interpretability analyses from both global and atomic perspectives, tracing the origins of their distinct transport behaviors. Interestingly, the decision process of the model naturally reveals underlying physical patterns, offering new insights into computer-assisted materials design.
title Learning Thermoelectric Transport from Crystal Structures via Multiscale Graph Neural Network
topic Materials Science
Applied Physics
url https://arxiv.org/abs/2512.06697