EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Time-delayed transformers for data-driven modeling of low-dimensional dynamics

Albert Alcalde, Markus Widhalm, Emre Yılmaz

Machine Learning: Science and Technology · 01.08.2026 · 7 (4) · 045038

DOI: 10.1088/2632-2153/ae8e2d ↗

Kurz gefasst

Fachartikel in der Zeitschrift Machine Learning: Science and Technology, online erschienen am 04.08.2026. Verfasst von Albert Alcalde und 2 weiteren Autor:innen. Die Arbeit nennt 54 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Union’s Horizon Europe MSCA, German Aerospace Center, Saxon State Ministry for Economic Affairs, Labor and Transport" and "Federal Ministry for Economic Affairs and Climate Action".

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We propose the time-delayed transformer (TD-TF), a simplified TF architecture for data-driven modeling of unsteady spatio-temporal dynamics. TD-TF bridges linear operator-based methods and deep sequence models by showing that a single-layer, single-head TF can be interpreted as a nonlinear generalization of TD dynamic mode decomposition. The architecture is deliberately minimal, consisting of one self-attention layer with a single query per prediction and one feedforward layer, resulting in linear computational complexity in sequence length and a small parameter count. Numerical experiments demonstrate that TD-TF matches the performance of linear baselines on near-linear systems, while significantly outperforming them in nonlinear and chaotic regimes, where it accurately captures long-term dynamics. Validation studies on synthetic signals, unsteady aerodynamics, the Lorenz’63 system, and a reaction-diffusion model show that TD-TF preserves the interpretability and efficiency of linear models while providing substantially enhanced expressive power for complex dynamics.

Förderung

  • European Union’s Horizon Europe MSCA
  • German Aerospace Center
  • Saxon State Ministry for Economic Affairs, Labor and Transport" and "Federal Ministry for Economic Affairs and Climate Action"

Zitieren

APA

Alcalde, A., Widhalm, M., & Yılmaz, E. (2026). Time-delayed transformers for data-driven modeling of low-dimensional dynamics. Machine Learning: Science and Technology, 7(4), 045038. https://doi.org/10.1088/2632-2153/ae8e2d

Vancouver

Alcalde A, Widhalm M, Yılmaz E. Time-delayed transformers for data-driven modeling of low-dimensional dynamics. Machine Learning: Science and Technology. 2026;7(4):045038. doi:10.1088/2632-2153/ae8e2d

BibTeX
@article{Alcalde2026,
  author = {Alcalde, Albert and Widhalm, Markus and Yılmaz, Emre},
  title = {Time-delayed transformers for data-driven modeling of low-dimensional dynamics},
  journal = {Machine Learning: Science and Technology},
  year = {2026},
  volume = {7},
  number = {4},
  pages = {045038},
  doi = {10.1088/2632-2153/ae8e2d}
}
Kostenlos starten →