arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.16397math.NAcs.LGcs.NA
收藏

使用 Graph Spline Networks 学习高效的代理动态模型

Learning Efficient Surrogate Dynamic Models with Graph Spline Networks

KAIST(韩国科学技术院);DiffeqML, AI4CO(DiffeqML,AI4CO);Stanford University(斯坦福大学);Mila and University of Montreal(米拉研究所和蒙特利尔大学);OMELET

Chuanbo Hua, Federico Berto, Michael Poli, Stefano Massaroli, Jinkyoo Park

AI总结:

提出 GraphSplineNets 深度学习方法,利用可微正交样条配置和空间自适应策略减小网格尺寸与迭代步数,在多种动力系统预测中显著提升了精度与加速的权衡。

摘要:

尽管物理系统的复杂模拟已广泛应用于工程和科学计算,但深度学习方法直到最近才开始解决其高昂的计算需求问题。本文提出 GraphSplineNets,一种新型深度学习方法,通过减小网格尺寸和减少深度代理模型的迭代步数来加速物理系统的预测。该方法使用两种可微正交样条配置方法,高效预测任意时空位置的响应。此外,我们引入了空间自适应配置策略,以优先从最重要区域进行采样。GraphSplineNets 在预测复杂度递增的多种动力系统时改善了精度与加速的权衡,包括热方程、阻尼波传播、Navier-Stokes 方程以及规则和非规则域中的真实海洋流。

英文摘要:

While complex simulations of physical systems have been widely used in engineering and scientific computing, lowering their often prohibitive computational requirements has only recently been tackled by deep learning approaches. In this paper, we present GraphSplineNets, a novel deep-learning method to speed up the forecasting of physical systems by reducing the grid size and number of iteration steps of deep surrogate models. Our method uses two differentiable orthogonal spline collocation methods to efficiently predict response at any location in time and space. Additionally, we introduce an adaptive collocation strategy in space to prioritize sampling from the most important regions. GraphSplineNets improve the accuracy-speedup tradeoff in forecasting various dynamical systems with increasing complexity, including the heat equation, damped wave propagation, Navier-Stokes equations, and real-world ocean currents in both regular and irregular domains.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑