arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2402.07465cs.LGcs.AIcs.NAmath.DSmath.NAstat.ML
收藏

基于分数的物理信息神经网络求解高维 Fokker-Planck 方程

Score-Based Physics-Informed Neural Networks for High-Dimensional Fokker-Planck Equations

National University of Singapore(新加坡国立大学);Worcester Polytechnic Institute(伍斯特理工学院);Brown University(布朗大学);Pacific Northwest National Laboratory(太平洋西北国家实验室)

Zheyuan Hu, Zhongqiang Zhang, George Em Karniadakis, Kenji Kawaguchi

AI总结:

针对高维 Fokker-Planck 方程的维度灾难问题,提出基于分数的 SDE 求解器,通过 SM、SSM 或 Score-PINN 拟合分数函数并经 ODE 求解对数似然,在多种 SDE 与分布上验证了其稳定性、速度与性能。

摘要:

Fokker-Planck(FP)方程是随机过程中的基础偏微分方程(PDE)。然而,在处理高维 FP PDE 时,维度灾难(CoD)带来了挑战。尽管 Monte Carlo 和原始的 Physics-Informed Neural Networks(PINNs)已展现出应对 CoD 的潜力,但在处理与 Brownian motion 相关的概率密度函数(PDF)时,这两种方法在高维情况下均会出现数值误差。随着维度的增加,逐点 PDF 值呈指数级下降,超出了数值模拟的精度,导致显著误差。此外,由于需要大量采样,Monte Carlo 无法提供快速采样。通过原始 PINNs 对对数似然(LL)进行建模,会将 FP 方程转化为一个困难的 HJB 方程,其误差随维度迅速增长。为此,我们提出了一种新颖的方法,利用基于分数的求解器来拟合 SDE 中的分数函数。分数函数定义为 LL 的梯度,在推断 LL 和 PDF 中起基础作用,并支持快速 SDE 采样。我们引入了三种拟合方法:Score Matching(SM)、Sliced SM(SSM)和 Score-PINN。所提出的基于分数的 SDE 求解器分两个阶段运行:首先,采用 SM、SSM 或 Score-PINN 获取分数;其次,利用所得分数通过 ODE 求解 LL。这些方法的对比评估展示了不同的权衡。所提方法在多种 SDE 上进行了评估,包括各向异性 OU 过程、几何 Brownian 以及具有不同特征空间的 Brownian。我们还测试了多种分布,包括 Gaussian、Log-normal、Laplace 和 Cauchy。数值结果证明了该基于分数的 SDE 求解器在不同设置下的稳定性、速度和性能,巩固了其作为高维 FP 方程 CoD 解决方案的潜力。

英文摘要:

The Fokker-Planck (FP) equation is a foundational PDE in stochastic processes. However, curse of dimensionality (CoD) poses challenge when dealing with high-dimensional FP PDEs. Although Monte Carlo and vanilla Physics-Informed Neural Networks (PINNs) have shown the potential to tackle CoD, both methods exhibit numerical errors in high dimensions when dealing with the probability density function (PDF) associated with Brownian motion. The point-wise PDF values tend to decrease exponentially as dimension increases, surpassing the precision of numerical simulations and resulting in substantial errors. Moreover, due to its massive sampling, Monte Carlo fails to offer fast sampling. Modeling the logarithm likelihood (LL) via vanilla PINNs transforms the FP equation into a difficult HJB equation, whose error grows rapidly with dimension. To this end, we propose a novel approach utilizing a score-based solver to fit the score function in SDEs. The score function, defined as the gradient of the LL, plays a fundamental role in inferring LL and PDF and enables fast SDE sampling. Three fitting methods, Score Matching (SM), Sliced SM (SSM), and Score-PINN, are introduced. The proposed score-based SDE solver operates in two stages: first, employing SM, SSM, or Score-PINN to acquire the score; and second, solving the LL via an ODE using the obtained score. Comparative evaluations across these methods showcase varying trade-offs. The proposed method is evaluated across diverse SDEs, including anisotropic OU processes, geometric Brownian, and Brownian with varying eigenspace. We also test various distributions, including Gaussian, Log-normal, Laplace, and Cauchy. The numerical results demonstrate the score-based SDE solver's stability, speed, and performance across different settings, solidifying its potential as a solution to CoD for high-dimensional FP equations.

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。

↑