EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Integrating transfer learning and uncertainty quantification for data-efficient materials discovery

Hajar Hosseini, Gloria A Sulley, Jihun Hamm, Matthew M Montemore

Journal of Physics: Materials · 01.09.2026 · 9 (3) · 035005

DOI: 10.1088/2515-7639/ae7b89 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Physics: Materials, online erschienen am 25.06.2026. Verfasst von Hajar Hosseini und 3 weiteren Autor:innen. Die Arbeit nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Science Foundation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

A major challenge in machine learning for materials lies in the simultaneous use of large, lower-accuracy datasets and small, high-accuracy datasets for data-efficient screening. Here, we show that the combination of transfer learning and uncertainty quantification enables highly efficient multi-fidelity materials screening. Using band gap prediction as a representative application, a model pre-trained on generalized gradient approximation density functional theory (GGA-DFT) data was fine-tuned on experimental and higher-level DFT datasets. The transfer learning model reduced the MAE by approximately 39% on the experimental dataset (0.833 ± 0.104 eV–0.502 ± 0.063 eV) and by 54% on the strongly constrained and appropriately normed (SCAN) dataset (0.716 ± 0.093 eV–0.331 ± 0.031 eV) relative to the respective target models trained from scratch, and matched or exceeded the best accuracy reported in the literature without requiring feature engineering. Ensemble-based uncertainty estimates guide the screening process and minimize the need for new calculations. In a photovoltaic case study, only six additional meta-GGA calculations are required to identify multiple DFT-validated candidate materials with band gaps near the target value. More broadly, this work establishes a general, data-efficient strategy applicable for accelerating materials discovery by leveraging heterogeneous datasets.

Förderung

  • National Science Foundation

Zitieren

APA

Hosseini, H., Sulley, G. A., Hamm, J., & Montemore, M. M. (2026). Integrating transfer learning and uncertainty quantification for data-efficient materials discovery. Journal of Physics: Materials, 9(3), 035005. https://doi.org/10.1088/2515-7639/ae7b89

Vancouver

Hosseini H, Sulley GA, Hamm J, Montemore MM. Integrating transfer learning and uncertainty quantification for data-efficient materials discovery. Journal of Physics: Materials. 2026;9(3):035005. doi:10.1088/2515-7639/ae7b89

BibTeX
@article{Hosseini2026,
  author = {Hosseini, Hajar and Sulley, Gloria A and Hamm, Jihun and Montemore, Matthew M},
  title = {Integrating transfer learning and uncertainty quantification for data-efficient materials discovery},
  journal = {Journal of Physics: Materials},
  year = {2026},
  volume = {9},
  number = {3},
  pages = {035005},
  doi = {10.1088/2515-7639/ae7b89}
}
Kostenlos starten →