来自今天的爱可可AI前沿推介

[LG] Learning Vortex Dynamics for Fluid Inference and Prediction

Y Deng, H Yu, J Wu, B Zhu
[Dartmouth College & Stanford University]

面向流体推理与预测的涡旋动力学学习

要点:

  1. 提出一种基于可微涡旋粒子的新型机器学习方法，从单一视频中推断和预测流体动力学；
  2. 一个基于粒子的潜空间，来捕捉流体动力学的拉格朗日涡流演变；
  3. 一个新的可微涡旋粒子系统，有效地表示复杂的流动特征。

一句话总结:
提出了一种新的、可微涡粒子系统，从单个视频中推断和预测流体动力学；设计了一个端到端的训练管道，从数据中学习和合成模拟器，以可靠地提供未来的视频扩展。

摘要：
本文提出一种基于可微涡旋粒子的新型机器学习方法，从单一视频中推断和预测流体动力学。该系统的关键设计是一个基于粒子的潜空间，以封装隐藏的、拉格朗日的涡流演化，支撑可观察的、欧拉的流动现象。设计了一个新的可微涡流粒子系统，结合其可学习的涡流到速度的动力学映射，以有效地捕捉和表示减缩空间的复杂流动特征。进一步设计了一个端到端的训练管道，直接从数据中学习和合成模拟器，可以在有限的观察基础上可靠地提供未来的视频推广。该方法有两方面的价值：首先，学到的模拟器能纯粹从视觉观察中推断出隐藏的物理量(如速度场)，用于运动分析；其次，其支持未来预测，构建输入视频的序列及其未来的动态演化。通过在合成视频和真实世界视频上与一系列现有方法进行定量和定性比较来证明所提出方法的有效性，显示出更好的数据对应性、视觉可信性和物理完整性。

We propose a novel machine learning method based on differentiable vortex particles to infer and predict fluid dynamics from a single video. The key design of our system is a particle-based latent space to encapsulate the hidden, Lagrangian vortical evolution underpinning the observable, Eulerian flow phenomena. We devise a novel differentiable vortex particle system in conjunction with their learnable, vortex-to-velocity dynamics mapping to effectively capture and represent the complex flow features in a reduced space. We further design an end-to-end training pipeline to directly learn and synthesize simulators from data, that can reliably deliver future video rollouts based on limited observation. The value of our method is twofold: first, our learned simulator enables the inference of hidden physics quantities (e.g. velocity field) purely from visual observation, to be used for motion analysis; secondly, it also supports future prediction, constructing the input video's sequel along with its future dynamics evolution. We demonstrate our method's efficacy by comparing quantitatively and qualitatively with a range of existing methods on both synthetic and real-world videos, displaying improved data correspondence, visual plausibility, and physical integrity.

论文链接：https://arxiv.org/abs/2301.11494
图片
图片
图片
图片

内容中包含的图片若涉及版权问题，请及时与我们联系删除