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 2203.13648cs.LG
收藏

动力系统不动点在训练物理信息神经网络中的作用

On the Role of Fixed Points of Dynamical Systems in Training Physics-Informed Neural Networks

Know-Center GmbH – Research Center for Data-Driven Business & Big Data Analytics(知识中心有限公司——数据驱动业务与大数据分析研究中心);LEC GmbH – Large Engines Competence Center(大型发动机竞争力中心有限公司)

Franz M. Rohrhofer, Stefan Posch, Clemens Gößnitzer, Bernhard C. Geiger

摘要:

本文实证研究了动力系统上物理信息神经网络(PINNs)的常见训练困难。我们的研究结果表明,这些系统固有的不动点在PINNs嵌入物理损失函数的优化中起着关键作用。我们观察到,损失景观表现出局部最优的形状的存在下的固定点。我们发现,这些局部最优有助于物理损失优化的复杂性,这可以解释常见的训练困难和由此产生的非物理预测。在某些设置下,例如,初始条件接近固定点或长的模拟时间,我们表明,这些最优甚至可以变得比所需的解决方案更好。

英文摘要:

This paper empirically studies commonly observed training difficulties of Physics-Informed Neural Networks (PINNs) on dynamical systems. Our results indicate that fixed points which are inherent to these systems play a key role in the optimization of the in PINNs embedded physics loss function. We observe that the loss landscape exhibits local optima that are shaped by the presence of fixed points. We find that these local optima contribute to the complexity of the physics loss optimization which can explain common training difficulties and resulting nonphysical predictions. Under certain settings, e.g., initial conditions close to fixed points or long simulations times, we show that those optima can even become better than that of the desired solution.

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。

↑