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 2303.15849cs.LGphysics.comp-ph
收藏

GAS:一种基于高斯混合分布的自适应采样方法用于PINNs

GAS: A Gaussian Mixture Distribution-Based Adaptive Sampling Method for PINNs

School of Mathematics and Statistics, Wuhan University(武汉大学数学与统计学院);Hubei Key Laboratory of Computational Science, Wuhan University(武汉大学湖北计算科学重点实验室)

Yuling Jiao, Di Li, Xiliang Lu, Jerry Zhijian Yang, Cheng Yuan

AI总结:

本文提出GAS方法,通过基于高斯混合分布的自适应采样改进PINNs训练,利用残差信息采样额外点并与历史数据联合训练,在2D和10D问题上达到深度求解器中最优精度,且与传统数值求解器相当。

摘要:

随着深度学习在科学计算中的近期研究,物理信息神经网络(PINNs)方法因求解偏微分方程(PDEs)而受到广泛关注。与传统方法相比,PINNs能够高效处理高维问题,但精度相对较低,尤其是对于高度不规则的问题。受自适应有限元方法和增量学习思想的启发,我们提出了GAS,一种基于高斯混合分布的自适应采样方法用于PINNs。在训练过程中,GAS利用当前残差信息生成高斯混合分布,用于采样额外点,然后将这些点与历史数据一起训练,以加速损失收敛并实现更高精度。在2D和10D问题上的若干数值模拟表明,GAS是一种有前景的方法,在深度求解器中达到了最先进的精度,同时与传统数值求解器相当。

英文摘要:

With the recent study of deep learning in scientific computation, the Physics-Informed Neural Networks (PINNs) method has drawn widespread attention for solving Partial Differential Equations (PDEs). Compared to traditional methods, PINNs can efficiently handle high-dimensional problems, but the accuracy is relatively low, especially for highly irregular problems. Inspired by the idea of adaptive finite element methods and incremental learning, we propose GAS, a Gaussian mixture distribution-based adaptive sampling method for PINNs. During the training procedure, GAS uses the current residual information to generate a Gaussian mixture distribution for the sampling of additional points, which are then trained together with historical data to speed up the convergence of the loss and achieve higher accuracy. Several numerical simulations on 2D and 10D problems show that GAS is a promising method that achieves state-of-the-art accuracy among deep solvers, while being comparable with traditional numerical solvers.

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。

↑