跳到正文
arXiv cs.AI· Zijiang Yang, Xiaomeng Wu, Chen Kuang, Dongmei Fu·· 23 小时前AI 评分26

PANO:用并行潜在空间神经算子求解 PDE

PANO: Parallel Latent Space Neural Operator for Solving PDEs

AI 导读

研究者提出并行潜在空间神经算子 PANO,将观测间的物理交互建模为多个共享参数并行潜在空间交互结果的加性组合,避免单一压缩模式丢失物理信息。PANO 还引入 Coupled Assignment Projection(CAP)构建双向一致的编解码映射,并用 Pairwise Latent Space Attention Layer(PaLA)实现跨空间自适应交互。

正文 · 原文

View PDF HTML (experimental)

Abstract:Aggregating observations into a limited number of latent tokens and learning physical interactions in the latent space has emerged as an effective approach for Transformer-based neural operators to solve Partial Differential Equation (PDE) problems involving large-scale observations. However, existing methods typically compress observations using a single aggregation pattern, making physical information not retained by that pattern invisible to the latent space and thereby constraining physical interactions within it. To address this limitation, we propose PArallel Latent Space Neural Operator (PANO), a novel neural operator for solving PDEs that formulates physical interactions among observations as an additive combination of interaction results from parallel latent spaces with shared parameters. Specifically, PANO transforms observations into multiple distinct compressed representations through independent mappings, learns physical interactions in parallel based on these representations, and finally fuses the decoded results to predict the target physical field. Within PANO, we further propose Coupled Assignment Projection (CAP) to construct bidirectionally consistent encoding and decoding mappings, and Pairwise Latent Space Attention Layer (PaLA) to jointly update latent representations of each latent space through adaptive cross-space interactions. Extensive experiments on nine benchmarks spanning diverse physical systems, geometries, and observation scales demonstrate that PANO achieves state-of-the-art performance. Notably, PANO reduces equivalent plastic strain prediction errors on Unilateral Stamping by 28.3% (or more) and lift-coefficient prediction errors on AirCraft by 21.4% (or more) compared with existing methods. Code is available at this https URL.
Comments: 39 pages, 24 figures, 23 tables
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.22782 [cs.LG]
  (or arXiv:2611.22782v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2611.22782

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Zijiang Yang [view email]
[v1] Tue, 17 Nov 2026 13:46:57 UTC (8,520 KB)

来源:arXiv cs.AI · arxiv.org