EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multifrequency wavefield solutions for variable velocity models using meta-learning enhanced low-rank physics-informed neural network

Shijun Cheng, Tariq Alkhalifah

Geophysical Journal International · 11.11.2025 · 244 (1)

DOI: 10.1093/gji/ggaf417 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Geophysical Journal International, online erschienen am 17.10.2025. Verfasst von Shijun Cheng und 1 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: King Abdullah University of Science and Technology.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Physics-informed neural networks (PINNs) face significant challenges in modelling multifrequency wavefields in complex velocity models due to their slow convergence, difficulty in representing high-frequency details and lack of generalization to varying frequencies and velocity scenarios. To address these issues, we propose Meta-LRPINN, a novel framework that combines low-rank parametrization using singular value decomposition (SVD) with meta-learning and frequency embedding. Specifically, we decompose the weights of PINN’s hidden layers using SVD and introduce an innovative frequency embedding hypernetwork (FEH) that links input frequencies with the singular values, enabling efficient and frequency-adaptive wavefield representation. Meta-learning is employed to provide robust initialization, improving optimization stability and reducing training time. Additionally, we implement adaptive rank reduction and FEH pruning during the meta-testing phase to further enhance efficiency. Numerical experiments, which are presented on multifrequency scattered wavefields for different velocity models, demonstrate that Meta-LRPINN achieves much faster convergence speed and much higher accuracy compared to baseline methods such as Meta-PINN and vanilla PINN. Also, the proposed framework shows strong generalization to out-of-distribution frequencies while maintaining computational efficiency. These results highlight the potential of our Meta-LRPINN for scalable and adaptable seismic wavefield modelling.

Förderung

  • King Abdullah University of Science and Technology

Zitieren

APA

Cheng, S., & Alkhalifah, T. (2025). Multifrequency wavefield solutions for variable velocity models using meta-learning enhanced low-rank physics-informed neural network. Geophysical Journal International, 244(1). https://doi.org/10.1093/gji/ggaf417

Vancouver

Cheng S, Alkhalifah T. Multifrequency wavefield solutions for variable velocity models using meta-learning enhanced low-rank physics-informed neural network. Geophysical Journal International. 2025;244(1). doi:10.1093/gji/ggaf417

BibTeX
@article{Cheng2025,
  author = {Cheng, Shijun and Alkhalifah, Tariq},
  title = {Multifrequency wavefield solutions for variable velocity models using meta-learning enhanced low-rank physics-informed neural network},
  journal = {Geophysical Journal International},
  year = {2025},
  volume = {244},
  number = {1},
  doi = {10.1093/gji/ggaf417}
}
Kostenlos starten →