BriefGPT.xyz
Oct, 2019
物理结构嵌入的变分积分网络
Variational Integrator Networks for Physically Meaningful Embeddings
HTML
PDF
Steindor Saemundsson, Alexander Terenin, Katja Hofmann, Marc Peter Deisenroth
TL;DR
利用深度神经网络与微分方程系统之间联系的最新进展，我们提出了一种名为“变分积分器网络”的神经网络架构，旨在保留物理系统的几何结构，并且能够对复杂的动态系统进行精确的长期预测、解释和数据效率学习。
Abstract
Learning workable representations of
dynamical systems
is becoming an increasingly important problem in a number of application areas. By leveraging recent work connecting deep
neural networks
to systems of diffe
→