EUVIMEDEUVIMED entdecken
Crossref

Evaluating Pre-Loss Normalization Across Multiple Loss Functions in Deep Learning for Multivariate Fluid Flow Prediction

Umut Kaya, Özgür Ekici

International Journal of Computational Methods · 2026-11 · 23 (09)

DOI: 10.1142/s0219876226500295 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Computational Methods, online erschienen am 21.04.2026. Verfasst von Umut Kaya und 1 weiteren Autor:innen. Die Arbeit nennt 39 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multi-scale flow fields can destabilize the training of data-driven models by weighting loss contributions unevenly. This study proposes a loss-independent pre-loss normalization module, used only during training, that independently scales predictions and targets prior to loss computation. Three normalization schemes are tested with four loss functions using a transformer model for laminar developing pipe flow. On the test dataset and two additional flow cases, Pre-Loss Normalization reduces the mean absolute error by 27–81% and improves the [Formula: see text]2 value of 0.186–0.630 over unnormalized training. Overall, this approach mitigates multi-scale effects and improves accuracy and robustness.

Zitieren

APA

Kaya, U., & Ekici, Ö. (2026). Evaluating Pre-Loss Normalization Across Multiple Loss Functions in Deep Learning for Multivariate Fluid Flow Prediction. International Journal of Computational Methods, 23(09). https://doi.org/10.1142/s0219876226500295

Vancouver

Kaya U, Ekici Ö. Evaluating Pre-Loss Normalization Across Multiple Loss Functions in Deep Learning for Multivariate Fluid Flow Prediction. International Journal of Computational Methods. 2026;23(09). doi:10.1142/s0219876226500295

BibTeX
@article{Kaya2026,
  author = {Kaya, Umut and Ekici, Özgür},
  title = {Evaluating Pre-Loss Normalization Across Multiple Loss Functions in Deep Learning for Multivariate Fluid Flow Prediction},
  journal = {International Journal of Computational Methods},
  year = {2026},
  volume = {23},
  number = {09},
  doi = {10.1142/s0219876226500295}
}
Kostenlos starten →