Advection Diffusion Reaction Graph Neural Networks for Spatio-Temporal Data

Moshe Eliasof, Eldad Haber, Eran Treister

Keywords: Advection, Diffusion, Reaction, Temporal, PDE, ODE

TL;DR: We present a novel architecture based on the Advection-Diffusion-Reaction PDE that allows greater expressiveness compared to existing models.

Abstract:

Graph neural networks (GNNs) have shown remarkable success in learning representations for graph-structured data. However, GNNs still face challenges in modeling complex phenomena that involve advection. In this paper, we propose a novel GNN architecture based on Advection-Diffusion-Reaction systems, and demonstrate its efficacy on real-world spatio-temporal datasets.

Submission Type: Extended abstract (max 4 main pages).

Agreement: Check this if you are okay with being contacted to participate in an anonymous survey.

Submission Number: 7