Event Detection on Dynamic Graphs

TL;DR

DyGED is a novel deep learning model that effectively detects events in dynamic graphs by capturing macro-level graph dynamics and integrating structural and temporal attention mechanisms, outperforming existing methods.

Contribution

Introduces DyGED, a new deep learning approach combining graph-level dynamics with attention mechanisms for improved event detection on dynamic graphs.

Findings

Outperforms competing solutions by up to 8.5% in accuracy

More scalable than top alternatives

Demonstrated effectiveness through case studies

Abstract

Event detection is a critical task for timely decision-making in graph analytics applications. Despite the recent progress towards deep learning on graphs, event detection on dynamic graphs presents particular challenges to existing architectures. Real-life events are often associated with sudden deviations of the normal behavior of the graph. However, existing approaches for dynamic node embedding are unable to capture the graph-level dynamics related to events. In this paper, we propose DyGED, a simple yet novel deep learning model for event detection on dynamic graphs. DyGED learns correlations between the graph macro dynamics -- i.e. a sequence of graph-level representations -- and labeled events. Moreover, our approach combines structural and temporal self-attention mechanisms to account for application-specific node and time importances effectively. Our experimental evaluation, using a representative set of datasets, demonstrates that DyGED outperforms competing solutions in terms of event detection accuracy by up to 8.5% while being more scalable than the top alternatives. We also present case studies illustrating key features of our model.