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 2102.11026cs.LGcs.GR
收藏

用于非线性模型降阶的高阶可微自编码器

High-order Differentiable Autoencoder for Nonlinear Model Reduction

State Key Lab of CAD&CG, Zhejiang University(浙江大学CAD&CG国家重点实验室);Clemson University(克莱姆森大学);University of Leeds(利兹大学);University of Pennsylvania(宾夕法尼亚大学)

Siyuan Shen, Yang Yin, Tianjia Shao, He Wang, Chenfanfu Jiang, Lei Lan, Kun Zhou

摘要:

本文为利用深度神经网络改进基于物理的仿真提供了一条新途径。具体来说,我们将经典的拉格朗日力学与深度自动编码器相结合,以加速可变形固体的弹性模拟。由于惯性效应,在不评估深度自编码器网络的二阶导数的情况下不能建立动态平衡。这超出了现成的自动微分包和算法的能力,其主要集中在梯度评估上。如果使用标准牛顿法,求解非线性力平衡甚至更具挑战性。这是因为我们需要计算网络的三阶导数来获得变分海森。我们攻击这些困难,利用复杂的步骤有限差分,加上反向自动微分。该策略在充分利用复步长有限差分的方便性和精确性的同时,尽可能集中地部署复值扰动,以节省过多的网络通道。通过基于GPU的实现,我们能够使用深度自动编码器(例如,$10+$层)与相对高维的潜在空间。沿着这条管道,我们还设计了一个采样网络和一个加权网络,使\emph{weight-varying} Cubature积分,以纳入非线性模型简化。我们相信这项工作将启发和有益于未来的研究工作,在非线性减少物理模拟问题。

英文摘要:

This paper provides a new avenue for exploiting deep neural networks to improve physics-based simulation. Specifically, we integrate the classic Lagrangian mechanics with a deep autoencoder to accelerate elastic simulation of deformable solids. Due to the inertia effect, the dynamic equilibrium cannot be established without evaluating the second-order derivatives of the deep autoencoder network. This is beyond the capability of off-the-shelf automatic differentiation packages and algorithms, which mainly focus on the gradient evaluation. Solving the nonlinear force equilibrium is even more challenging if the standard Newton's method is to be used. This is because we need to compute a third-order derivative of the network to obtain the variational Hessian. We attack those difficulties by exploiting complex-step finite difference, coupled with reverse automatic differentiation. This strategy allows us to enjoy the convenience and accuracy of complex-step finite difference and in the meantime, to deploy complex-value perturbations as collectively as possible to save excessive network passes. With a GPU-based implementation, we are able to wield deep autoencoders (e.g., $10+$ layers) with a relatively high-dimension latent space in real-time. Along this pipeline, we also design a sampling network and a weighting network to enable \emph{weight-varying} Cubature integration in order to incorporate nonlinearity in the model reduction. We believe this work will inspire and benefit future research efforts in nonlinearly reduced physical simulation problems.

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。

↑