论文标题
用哈密顿神经网络掌握高维动力
Mastering high-dimensional dynamics with Hamiltonian neural networks
论文作者
论文摘要
我们详细介绍将物理学纳入神经网络设计如何显着改善动态系统的学习和预测，甚至是许多维度的非线性系统。地图构建观点阐明了哈密顿神经网络比常规神经网络的优越性。结果阐明了数据，维度和神经网络学习绩效之间的关键关系。
We detail how incorporating physics into neural network design can significantly improve the learning and forecasting of dynamical systems, even nonlinear systems of many dimensions. A map building perspective elucidates the superiority of Hamiltonian neural networks over conventional neural networks. The results clarify the critical relation between data, dimension, and neural network learning performance.
