Timezone: »

 
Poster
Equivariant Graph Hierarchy-Based Neural Networks
Jiaqi Han · Wenbing Huang · Tingyang Xu · Yu Rong

@

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-UnPool. 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-UnPool leverages the high-level information to update the dynamics of the low-level nodes. As their names imply, both E-Pool and E-UnPool 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.

Author Information

Jiaqi Han (Tsinghua University)
Wenbing Huang (Tsinghua University)
Tingyang Xu (Tencent AI Lab)
Yu Rong (Shenzhen Tencent Computer System Co., Ltd.)

More from the Same Authors