SEGNO: Generalizing Equivariant Graph Neural Networks with Physical Inductive Biases

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Yang, Cheng, Jiashun, Zhao, Haihong, Xu, Tingyang, Zhao, Peilin, Tsung, Fugee, Li, Jia, Rong, Yu
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913261416349696
author Liu, Yang
Cheng, Jiashun
Zhao, Haihong
Xu, Tingyang
Zhao, Peilin
Tsung, Fugee
Li, Jia
Rong, Yu
author_facet Liu, Yang
Cheng, Jiashun
Zhao, Haihong
Xu, Tingyang
Zhao, Peilin
Tsung, Fugee
Li, Jia
Rong, Yu
contents Graph Neural Networks (GNNs) with equivariant properties have emerged as powerful tools for modeling complex dynamics of multi-object physical systems. However, their generalization ability is limited by the inadequate consideration of physical inductive biases: (1) Existing studies overlook the continuity of transitions among system states, opting to employ several discrete transformation layers to learn the direct mapping between two adjacent states; (2) Most models only account for first-order velocity information, despite the fact that many physical systems are governed by second-order motion laws. To incorporate these inductive biases, we propose the Second-order Equivariant Graph Neural Ordinary Differential Equation (SEGNO). Specifically, we show how the second-order continuity can be incorporated into GNNs while maintaining the equivariant property. Furthermore, we offer theoretical insights into SEGNO, highlighting that it can learn a unique trajectory between adjacent states, which is crucial for model generalization. Additionally, we prove that the discrepancy between this learned trajectory of SEGNO and the true trajectory is bounded. Extensive experiments on complex dynamical systems including molecular dynamics and motion capture demonstrate that our model yields a significant improvement over the state-of-the-art baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2308_13212
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle SEGNO: Generalizing Equivariant Graph Neural Networks with Physical Inductive Biases
Liu, Yang
Cheng, Jiashun
Zhao, Haihong
Xu, Tingyang
Zhao, Peilin
Tsung, Fugee
Li, Jia
Rong, Yu
Machine Learning
Artificial Intelligence
Graph Neural Networks (GNNs) with equivariant properties have emerged as powerful tools for modeling complex dynamics of multi-object physical systems. However, their generalization ability is limited by the inadequate consideration of physical inductive biases: (1) Existing studies overlook the continuity of transitions among system states, opting to employ several discrete transformation layers to learn the direct mapping between two adjacent states; (2) Most models only account for first-order velocity information, despite the fact that many physical systems are governed by second-order motion laws. To incorporate these inductive biases, we propose the Second-order Equivariant Graph Neural Ordinary Differential Equation (SEGNO). Specifically, we show how the second-order continuity can be incorporated into GNNs while maintaining the equivariant property. Furthermore, we offer theoretical insights into SEGNO, highlighting that it can learn a unique trajectory between adjacent states, which is crucial for model generalization. Additionally, we prove that the discrepancy between this learned trajectory of SEGNO and the true trajectory is bounded. Extensive experiments on complex dynamical systems including molecular dynamics and motion capture demonstrate that our model yields a significant improvement over the state-of-the-art baselines.
title SEGNO: Generalizing Equivariant Graph Neural Networks with Physical Inductive Biases
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2308.13212