arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.15526cs.LGcs.AIphysics.comp-ph
收藏

在PINN中建立信任:通过有限差分方法进行误差估计

Building Trust in PINNs: Error Estimation through Finite Difference Methods

Department of Artificial Intelligence, Fraunhofer Heinrich Hertz Institute(人工智能部门,弗劳恩霍夫海因里希·赫兹研究所);Department of Electrical Engineering and Computer Science, Technische Universität Berlin(电气工程与计算机科学系,柏林技术大学);BIFOLD - Berlin Institute for the Foundations of Learning(BIFOLD—柏林学习与数据基础研究所);Centre of eXplainable Artificial Intelligence, Technological University Dublin(可解释人工智能中心,都柏林技术大学)

Aleksander Krasowski, René P. Klausen, Aycan Celik, Sebastian Lapuschkin, Wojciech Samek, Jonas Naujoks

AI总结:

本文提出一种轻量级后处理方法,通过有限差分法生成PINN预测的点误差估计,以解释模型预测的准确性,并在低计算成本下提供准确的误差图,用于验证PINN。

摘要:

物理信息神经网络(PINN)是一种灵活的深度学习方法,用于求解偏微分方程(PDEs),可建模从热传导到量子力学系统等现象。尽管灵活,PINN在预测与真实解的偏差方面缺乏洞察力,阻碍了对其预测质量的信任。我们提出了一种轻量级的后处理方法,通过生成点误差估计来填补这一空白,为这些模型提供自然的解释,不仅确定预测是否错误,还确定错误的位置和程度。对于线性偏微分方程,PINN近似与真实解之间的误差满足与原问题相同的微分算子,但驱动项是PINN的PDE残差。我们通过有限差分法数值求解这个误差方程,无需了解真实解。在多个基准PDE上评估,我们的方法在低计算成本下产生准确的误差图,使PINN的定向和可解释验证成为可能。

英文摘要:

Physics-informed neural networks (PINNs) constitute a flexible deep learning approach for solving partial differential equations (PDEs), which model phenomena ranging from heat conduction to quantum mechanical systems. Despite their flexibility, PINNs offer limited insight into how their predictions deviate from the true solution, hindering trust in their prediction quality. We propose a lightweight post-hoc method that addresses this gap by producing pointwise error estimates for PINN predictions, which offer a natural form of explanation for such models, identifying not just whether a prediction is wrong, but where and by how much. For linear partial differential equations, the error between a PINN approximation and the true solution satisfies the same differential operator as the original problem, but driven by the PINN's PDE residual as its source term. We solve this error equation numerically using finite difference methods requiring no knowledge of the true solution. Evaluated on several benchmark PDEs, our method yields accurate error maps at low computational cost, enabling targeted and interpretable validation of PINNs.

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。

↑