The paper presents a turbulence‑aware framework for reconstructing high‑fidelity three‑dimensional flow fields from coarse‑grained data. It combines a multiscale neural operator with a state‑space model to enforce temporal consistency, incorporates Navier‑Stokes residuals as soft constraints, and uses a mixture‑of‑experts mechanism for unsupervised adaptation to unknown coarse‑graining. Experiments on canonical turbulent systems and atmospheric reanalysis data show reduced reconstruction errors and better preservation of multiscale spatial structure and coherent temporal evolution compared to existing learning‑based and classical baselines.
By Hanlin Kong, Jianan Wei, Yuntian Chen, Yujia Wang, Wenguan Wang
High-fidelity three-dimensional turbulent flow fields are crucial for scientific and engineering analyses, but direct numerical simulation or high-resolution measurements are costly. The paper introduces a turbulence-aware framework that couples a multiscale neural operator with a state space model and incorporates Navier-Stokes residuals as soft constraints to reconstruct time-resolved 3‑D flow fields from coarse-grained data. The method also uses a mixture-of-experts mechanism for unsupervised adaptation to unknown coarse-graining, and demonstrates reduced reconstruction errors and improved preservation of multiscale spatial structure and coherent temporal evolution on canonical turbulent systems and atmospheric reanalysis data.
The paper introduces Flow Divergence Sampler (FDS), a training‑free method that refines intermediate states in flow‑matching models by using the divergence of the marginal velocity field to detect and correct misguidance toward low‑density regions. FDS operates during inference, requires no additional training, and can be applied as a plug‑and‑play module with standard solvers and existing flow backbones. Experiments show that FDS consistently improves fidelity in tasks such as text‑to‑image synthesis and inverse problems.
By Yeonwoo Cha, Jaehoon Yoo, Semin Kim, Yunseo Park, Jinhyeon Kwon, Seunghoon Hong
The paper presents a Deep Operator Network that reconstructs full‑field 4D mitral regurgitation hemodynamics from sparse planar velocity data and boundary pressure traces. The network is pretrained on a URANS database of eleven orifice phantoms and then fine‑tuned on new cases, achieving rapid predictions in minutes. While the adaptation improves flow topology in the observed plane, reconstruction error increases sharply with distance from that plane, limiting physical consistency in the surrounding volume.
By Jakob Marcel Hoffmann, Yosuke Hasegawa, Alexander Stroh
The paper introduces a kernel‑based operator learning method that preserves key physical properties—such as incompressibility, periodicity, and turbulence—of the incompressible Navier–Stokes equations. By mapping input functions to expansion coefficients in a property‑preserving kernel basis, the method guarantees that predicted velocity fields analytically maintain these properties. The authors provide theoretical convergence guarantees, develop efficient computational techniques for large‑scale training, and demonstrate significant accuracy and speed improvements over neural operators on 2D and 3D flow benchmarks.
By Ramansh Sharma, Matthew Lowery, Houman Owhadi, Varun Shankar
arXiv:2610.07033v1 Announce Type: new
Abstract: Predicting transient urban winds is fundamental to understanding urban microclimates and designing climate-resilient cities. Building-resolving large-e...
By Yidi Wang, Yunhe Zhang, Jiawei Gu, Ziyue Qiao, Pengyang Wang
arXiv:2608. 04222v1 Announce Type: cross Abstract: Turbulence is a central testbed for machine learning on physical dynamics because its governing laws are known exactly.
By Yilong Dai, Yiming Sun, Yiheng Chen, Shengyu Chen, Peyman Givi, Xiaowei Jia, Runlong Yu
arXiv:2507. 22082v2 Announce Type: replace-cross Abstract: Direct numerical simulation (DNS) accurately resolves all spatio-temporal scales of wall-bounded turbulence but becomes prohibitively expensive as the Reynolds number increases.
By Anuraj Maurya
The paper presents a lightweight retraining strategy for a parameterized Reduced Order Model (ROM) that achieves full‑model accuracy using only a fraction of the computational effort and sparse observations. The ROM architecture combines a Variational Autoencoder for dimensionality reduction with a transformer network that evolves latent states while accounting for the Reynolds number as an external control variable. By leveraging the probabilistic VAE, the method generates trajectory ensembles and uncertainty estimates, and adapts to out‑of‑sample parameters through sparse data assimilation with an ensemble Kalman filter, focusing retraining on the autoencoder to correct latent manifold distortions.
By Isma\"el Zighed, Andrea N\'ovoa, Luca Magri, Taraneh Sayadi
arXiv:2602.23188v2 Announce Type: replace
Abstract: We propose an efficient retraining strategy for a parameterized Reduced Order Model (ROM) that attains accuracy comparable to full retraining while...
By Isma\"el Zighed, Andrea N\'ovoa, Luca Magri, Taraneh Sayadi
arXiv:2605. 00941v4 Announce Type: replace Abstract: Flow matching has become a leading framework for generative modeling, but quantifying the uncertainty of its samples remains an open problem.
By Jiarui Xing, Song Wang, Jian Wang
arXiv:2604. 23874v3 Announce Type: replace-cross Abstract: The differentiable physics paradigm may be leveraged as an a-posteriori approach for discovering turbulence closure models by embedding a neural network parameterization directly inside the solver and optimizing it given potentially sparse target data.
By Ashwin Suriyanarayanan, Dibyajyoti Chakraborty, Romit Maulik