Time-varying Interaction Graph ODE for Dynamic Graph Representation Learning

TL;DR

TI-ODE introduces a novel dynamic graph neural ODE model that decomposes interaction functions into basis functions with time-dependent weights, capturing diverse, evolving inter-node interactions for improved performance.

Contribution

The paper proposes TI-ODE, a new method that models time-varying inter-node interactions using learnable basis functions with dynamic weights, enhancing dynamic graph learning.

Findings

TI-ODE outperforms existing methods on six dynamic graph datasets.

TI-ODE achieves state-of-the-art results in attribute prediction tasks.

TI-ODE demonstrates improved robustness and interpretability.

Abstract

Graph neural Ordinary Differential Equations (ODE) combine neural ODE with the message passing mechanism of Graph Neural Networks (GNN), providing a continuous-time modeling method for graph representation learning. However, in dynamic graph scenarios, existing graph neural ODEs typically employ a unified message passing mechanism, assuming that inter-node interactions share the same message passing function at any time, which makes it challenging to capture the diversity and time-varying nature of inter-node interaction patterns. To address this, we propose Time-varying Interaction Graph Ordinary Differential Equations (TI-ODE). The core idea of TI-ODE is to decompose the evolution function of a graph ODE into a set of learnable interaction basis functions, where each basis function corresponds to a distinct type of inter-node interaction. These basis functions are dynamically combined through time-dependent learnable weights, enabling inter-node interaction patterns to adaptively evolve over time. Experimental results on six dynamic graph datasets demonstrate that TI-ODE consistently outperforms existing methods and achieves state-of-the-art performance on attribute prediction tasks, and experiments on the \textit{Covid} dataset further verify the interpretability and generalizability of our TI-ODE. Furthermore, we demonstrate both theoretically and empirically that TI-ODE exhibits superior robustness compared to models utilizing a unified message-passing mechanism.