EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Adaptive Training of Grid-Dependent Physics-Informed Kolmogorov-Arnold Networks

Spyros Rigas, Michalis Papachristou, Theofilos Papadopoulos, Fotios Anagnostopoulos, Georgios Alexandridis

IEEE Access · 2024 · 12 · 176982-176998

DOI: 10.1109/access.2024.3504962 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2024. Verfasst von Spyros Rigas und 4 weiteren Autor:innen. Die Arbeit wurde bisher 59-mal zitiert (Crossref) und nennt 64 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Physics-Informed Neural Networks (PINNs) have emerged as a robust framework for solving Partial Differential Equations (PDEs) by approximating their solutions via neural networks and imposing physics-based constraints on the loss function. Traditionally, Multilayer Perceptrons (MLPs) have been the neural network of choice, with significant progress made in optimizing their training. Recently, Kolmogorov-Arnold Networks (KANs) were introduced as a viable alternative, with the potential of offering better interpretability and efficiency while requiring fewer parameters. In this paper, we present a fast JAX-based implementation of grid-dependent Physics-Informed Kolmogorov-Arnold Networks (PIKANs) for solving PDEs, achieving up to 84 times faster training times than the original KAN implementation. We propose an adaptive training scheme for PIKANs, introducing an adaptive state transition technique to avoid loss function peaks between grid extensions, and a methodology for designing PIKANs with alternative basis functions. Through comparative experiments, we demonstrate that the adaptive features significantly enhance solution accuracy, decreasing the L2 error relative to the reference solution by up to 43.02%. For the studied forward PDE problems, our methodology approaches or surpasses the results obtained from architectures that utilize up to 8.5 times more parameters, highlighting the potential of adaptive, grid-dependent PIKANs as a superior alternative in scientific and engineering applications.

Zitieren

APA

Rigas, S., Papachristou, M., Papadopoulos, T., Anagnostopoulos, F., & Alexandridis, G. (2024). Adaptive Training of Grid-Dependent Physics-Informed Kolmogorov-Arnold Networks. IEEE Access, 12, 176982-176998. https://doi.org/10.1109/access.2024.3504962

Vancouver

Rigas S, Papachristou M, Papadopoulos T, Anagnostopoulos F, Alexandridis G. Adaptive Training of Grid-Dependent Physics-Informed Kolmogorov-Arnold Networks. IEEE Access. 2024;12:176982-176998. doi:10.1109/access.2024.3504962

BibTeX
@article{Rigas2024,
  author = {Rigas, Spyros and Papachristou, Michalis and Papadopoulos, Theofilos and Anagnostopoulos, Fotios and Alexandridis, Georgios},
  title = {Adaptive Training of Grid-Dependent Physics-Informed Kolmogorov-Arnold Networks},
  journal = {IEEE Access},
  year = {2024},
  volume = {12},
  pages = {176982--176998},
  doi = {10.1109/access.2024.3504962}
}
Kostenlos starten →