EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Information decomposition for disentangled and interpretable manifold learning of fluid flows via variational autoencoders

Zhiyuan WANG, Iacopo TIRELLI, Stefano DISCETTI, Andrea IANIRO

Chinese Journal of Aeronautics · 2026-11 · 39 (11) · 104342

DOI: 10.1016/j.cja.2026.104342 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Chinese Journal of Aeronautics, 2026-11. Verfasst von Zhiyuan WANG und 3 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Horizon Europe, Spain Ministry of Science Innovation and Universities.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We introduce an information-theoretic framework that uses variational autoencoders (VAEs) to extract compact, physically interpretable manifolds from high-dimensional flow-field data. To this end, the Kullback--Leibler (KL) divergence in the variational objective is decomposed into three complementary information-theoretic terms: the index-code mutual information, the total correlation, and the dimension-wise KL divergence. These terms explicitly regulate data compression, latent disentanglement, and geometric regularization. This establishes a principled basis for targeted latent-space design, allowing enhanced interpretability without sacrificing information capacity, a common drawback of heavily regularized VAE variants. The approach is evaluated on two synthetic unsteady flow datasets. First, we consider a flow around a cylinder in a channel with variable cylinder position, diameter, and Reynolds number. Later, we also consider the flow around a NACA 0012 airfoil at varying angles of attack and subjected to strong vortex gusts with variable intensity, position, and length scale. Comparisons with Principal Component Analysis, Isometric Feature Mapping, and $\beta$-VAE demonstrate clear advantages in disentanglement and physical interpretability. The learned latent coordinates successfully separate distinct physical effects. Moreover, the proposed method demonstrates strong robustness to variations in the loss-weighting parameters, despite involving a larger number of such parameters.

Förderung

  • Horizon Europe
  • Spain Ministry of Science Innovation and Universities

Zitieren

APA

WANG, Z., TIRELLI, I., DISCETTI, S., & IANIRO, A. (2026). Information decomposition for disentangled and interpretable manifold learning of fluid flows via variational autoencoders. Chinese Journal of Aeronautics, 39(11), 104342. https://doi.org/10.1016/j.cja.2026.104342

Vancouver

WANG Z, TIRELLI I, DISCETTI S, IANIRO A. Information decomposition for disentangled and interpretable manifold learning of fluid flows via variational autoencoders. Chinese Journal of Aeronautics. 2026;39(11):104342. doi:10.1016/j.cja.2026.104342

BibTeX
@article{WANG2026,
  author = {WANG, Zhiyuan and TIRELLI, Iacopo and DISCETTI, Stefano and IANIRO, Andrea},
  title = {Information decomposition for disentangled and interpretable manifold learning of fluid flows via variational autoencoders},
  journal = {Chinese Journal of Aeronautics},
  year = {2026},
  volume = {39},
  number = {11},
  pages = {104342},
  doi = {10.1016/j.cja.2026.104342}
}
Kostenlos starten →