Back to publications
On The Temporal Domain of Differential Equation Inspired Graph Neural Networks
Paper Details
Published: 2024/01/20
Pages: 1792-1800
DOI: 10.48550/arXiv.2401.11074
ARXIV ID: 2401.11074v1
Container Title: International Conference on Artificial Intelligence and Statistics
Graph Neural Networks (GNNs) have demonstrated remarkable success in modeling complex relationships in graph-structured data. A recent innovation in this field is the family of Differential Equation-Inspired Graph Neural Networks (DE-GNNs), which leverage principles from continuous dynamical systems to model information flow on graphs with built-in properties such as feature smoothing or preservation. However, existing DE-GNNs rely on first or second-order temporal dependencies. In this paper, the authors propose a neural extension to those pre-defined temporal dependencies through a new model, TDE-GNN, which can capture a wide range of temporal dynamics that go beyond typical first or second-order methods, and provide use cases where existing temporal models are challenged.
Authors
E. Haber
E. Treister