EUVIMEDEUVIMED entdecken
Crossref

Node transfer for multi-fidelity and multimodal machine learning for predicting experimental bandgaps

Shuai Li iD, Wen-Cheng Yao iD, Bin-Bin Xie iD, Lin Shen iD, Ling Chen iD, Wei-Hai Fang iD

The Journal of Chemical Physics · 14.04.2026 · 164 (14)

DOI: 10.1063/5.0320627 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Journal of Chemical Physics, online erschienen am 08.04.2026. Verfasst von Shuai Li und 5 weiteren Autor:innen. Beteiligte Einrichtung: Key Laboratory of Theoretical and Computational Photochemistry of Ministry of Education, College of Chemistry, Beijing Normal University, Beijing 100…. Die Arbeit nennt 44 Literaturangaben. Gefördert von: Key Science and Technology Project of Jinhua City, Zhejiang Provincial Natural Science Foundation of China, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Bandgap is a key property of materials. In recent years, machine learning has become a powerful tool to predict the experimental bandgaps of compounds before synthesis, but there is still much room for improving the prediction accuracy. Here, we build a machine learning framework that consists of multi-fidelity and multimodal learning models to integrate heterogeneous data sources obtained from first-principle calculations and x-ray diffraction spectra. A new information-fusion strategy named node transfer is proposed. Compared to the widely used Δ-learning strategy, it naturally extends two-fidelity to multi-fidelity learning and facilitates heterogeneous multimodal integration. Node transfer consistently outperforms Δ-learning across two-fidelity, multi-fidelity, and multimodal benchmarks under fine-tuning. The best model involves XRD-based descriptors and encoded descriptors pre-trained based on four computational datasets using different functionals. It achieves a mean absolute error of 0.258 eV, a 26.3% reduction vs the single-fidelity baseline of 0.350 eV. In all prediction tasks, only the chemical composition of the crystal is required as input for the constructed machine learning models, which is free of structural information and, therefore, applicable to materials design before experiments or first-principle calculations.

Autor:innen und Institutionen

  • Key Laboratory of Theoretical and Computational Photochemistry of Ministry of Education, College of Chemistry, Beijing Normal University, Beijing 100875, China.
  • Hangzhou Institute of Advanced Studies, Zhejiang Normal University, Hangzhou 311231, Zhejiang, China.
  • Yantai-Jingshi Institute of Material Genome Engineering, Yantai 265505, Shandong, China.
  • State Key Laboratory of New Textile Materials and Advanced Processing, Wuhan Textile University, Wuhan 430200, Hubei, China.

ORCID-Profile: Shuai Li, Wen-Cheng Yao, Bin-Bin Xie, Lin Shen, Ling Chen, Wei-Hai Fang

Förderung

  • Key Science and Technology Project of Jinhua City
  • Zhejiang Provincial Natural Science Foundation of China
  • National Natural Science Foundation of China
  • Robotic AI-Scientist Platform of the Chinese Academy of Sciences

Zitieren

APA

Li, S., Yao, W. C., Xie, B. B., Shen, L., Chen, L., & Fang, W. H. (2026). Node transfer for multi-fidelity and multimodal machine learning for predicting experimental bandgaps. The Journal of Chemical Physics, 164(14). https://doi.org/10.1063/5.0320627

Vancouver

Li S, Yao WC, Xie BB, Shen L, Chen L, Fang WH. Node transfer for multi-fidelity and multimodal machine learning for predicting experimental bandgaps. The Journal of Chemical Physics. 2026;164(14). doi:10.1063/5.0320627

BibTeX
@article{Li2026,
  author = {Li, Shuai and Yao, Wen-Cheng and Xie, Bin-Bin and Shen, Lin and Chen, Ling and Fang, Wei-Hai},
  title = {Node transfer for multi-fidelity and multimodal machine learning for predicting experimental bandgaps},
  journal = {The Journal of Chemical Physics},
  year = {2026},
  volume = {164},
  number = {14},
  doi = {10.1063/5.0320627}
}
Kostenlos starten →