Significant Ties Graph Neural Networks for Continuous-Time Temporal Networks Modeling
This addresses the challenge of modeling dynamic systems like social networks and epidemiology, which often rely on discrete snapshots, by focusing on significant interactions, though it appears incremental in improving aggregation methods.
The paper tackles the problem of modeling continuous-time temporal networks by proposing Significant Ties Graph Neural Networks (STGNN), which captures significant ties through a novel aggregation mechanism, and experimental results on four real networks demonstrate its effectiveness.
Temporal networks are suitable for modeling complex evolving systems. It has a wide range of applications, such as social network analysis, recommender systems, and epidemiology. Recently, modeling such dynamic systems has drawn great attention in many domains. However, most existing approaches resort to taking discrete snapshots of the temporal networks and modeling all events with equal importance. This paper proposes Significant Ties Graph Neural Networks (STGNN), a novel framework that captures and describes significant ties. To better model the diversity of interactions, STGNN introduces a novel aggregation mechanism to organize the most significant historical neighbors' information and adaptively obtain the significance of node pairs. Experimental results on four real networks demonstrate the effectiveness of the proposed framework.