跳到正文
arXiv cs.AI· Gabriele Accarino, Juan Nathaniel, Carla Roesch, Pierre Gentine, Sara Shamekh, Duncan Watson-Parris, Viviana Acquaviva·· 4 小时前AI 评分41

用于多尺度物理仿真的 Wavelet Flow Matching(WFM)

Wavelet Flow Matching for Multi-Scale Physics Emulation

AI 导读

研究者提出 Wavelet Flow Matching(WFM),一种在多尺度小波空间中直接进行最优传输 flow matching 的生成式 PDE 仿真器,无需单独预训练 autoencoder。WFM 利用 U-Net 的层级结构联合预测无损小波表示的传输速度,在三个混沌流体动力学系统上实现更优的频谱一致性和不弱于 SOTA 的长时程精度。该工作被 NeurIPS 2026 接收。

正文 · 原文

View PDF HTML (experimental)

Abstract:Accurate emulation of multi-scale physical systems governed by PDEs demands models that remain stable over long autoregressive rollouts while preserving fine-scale structures. Deterministic emulators produce overly smoothed predictions, while generative approaches better capture details but are costly. Latent-space generative models have emerged as a compromise but with the additional cost of separately pre-trained autoencoders. We propose Wavelet Flow Matching (WFM), a generative emulator that overcomes current trade-offs between cost and skill by performing optimal-transport flow matching directly in the multi-scale wavelet space. Rather than learning a latent compression, WFM leverages the hierarchical structure of a U-Net to jointly predict the transport velocities of a prescribed, lossless wavelet representation. On three challenging systems of chaotic fluid dynamics, WFM achieves superior spectral coherence and competitive or better long-horizon accuracy compared to state-of-the-art models. Our results position the wavelet space as an effective training-free representation for generative emulation of complex physical dynamics.
Comments: Accepted at NeurIPS 2026; camera-ready version. Please check the project webpage for links to the source code, pre-trained weights, and datasets: this https URL
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Fluid Dynamics (physics.flu-dyn)
MSC classes: 42C40, 65T60, 35Q68,
ACM classes: I.4.10; I.5.4; I.2.10; G.3
Cite as: arXiv:2605.16573 [cs.LG]
  (or arXiv:2605.16573v2 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2605.16573

arXiv-issued DOI via DataCite

Submission history

From: Gabriele Accarino [view email]
[v1] Fri, 15 May 2026 19:24:31 UTC (21,887 KB)
[v2] Fri, 13 Nov 2026 00:51:03 UTC (22,282 KB)

来源:arXiv cs.AI · arxiv.org