arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2202.10643cs.LG
收藏

基于等变图层次结构的神经网络

Equivariant Graph Hierarchy-Based Neural Networks

Tsinghua University(清华大学);Renmin University of China(中国人民大学);Tencent AI Lab(腾讯人工智能实验室)

Jiaqi Han, Wenbing Huang, Tingyang Xu, Yu Rong

摘要:

等变图神经网络(EGN)是描述多体物理系统动力学的一种有效方法。现有的EGN进行平面消息传递,这,然而,是无法捕捉的空间/动态层次结构的复杂系统,特别是,限制子结构发现和全球信息融合。在本文中,我们提出了基于等变层次图网络(EGHN),它包括三个关键组件:广义等变矩阵消息传递(EMMP),E-Pool和E-UpPool。EMMP能够提高传统等变消息传递的表达能力,E-Pool将底层节点的数量分配到高层簇中,E-UpPool利用高层信息更新底层节点的动态。正如它们的名字所暗示的,E-Pool和E-UpPool都保证是等变的,以满足对称性。大量的实验评估验证了我们的EGHN的有效性,包括多目标动力学模拟,运动捕捉,蛋白质动力学建模的几个应用程序。

英文摘要:

Equivariant Graph neural Networks (EGNs) are powerful in characterizing the dynamics of multi-body physical systems. Existing EGNs conduct flat message passing, which, yet, is unable to capture the spatial/dynamical hierarchy for complex systems particularly, limiting substructure discovery and global information fusion. In this paper, we propose Equivariant Hierarchy-based Graph Networks (EGHNs) which consist of the three key components: generalized Equivariant Matrix Message Passing (EMMP) , E-Pool and E-UpPool. In particular, EMMP is able to improve the expressivity of conventional equivariant message passing, E-Pool assigns the quantities of the low-level nodes into high-level clusters, while E-UpPool leverages the high-level information to update the dynamics of the low-level nodes. As their names imply, both E-Pool and E-UpPool are guaranteed to be equivariant to meet physic symmetry. Considerable experimental evaluations verify the effectiveness of our EGHN on several applications including multi-object dynamics simulation, motion capture, and protein dynamics modeling.

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。

↑