Marked Neural Spatio-Temporal Point Process Involving a Dynamic Graph Neural Network

TL;DR

This paper introduces a Marked Neural Spatio-Temporal Point Process that uses a Dynamic Graph Neural Network to model and predict events in fully dynamic graph streams, capturing both structural and attribute changes.

Contribution

It presents the first neural TPP model capable of handling fully dynamic graphs with changing structure and attributes using a dynamic graph neural network.

Findings

Successfully models complex dynamic graph events

Enables prediction of future events in evolving graphs

Handles both structural and attribute changes

Abstract

Temporal Point Processes (TPPs) have recently become increasingly interesting for learning dynamics in graph data. A reason for this is that learning on dynamic graph data is becoming more relevant, since data from many scientific fields, ranging from mathematics, biology, social sciences, and physics to computer science, is naturally related and inherently dynamic. In addition, TPPs provide a meaningful characterization of event streams and a prediction mechanism for future events. Therefore, (semi-)parameterized Neural TPPs have been introduced whose characterization can be (partially) learned and, thus, enable the representation of more complex phenomena. However, the research on modeling dynamic graphs with TPPs is relatively young, and only a few models for node attribute changes or evolving edges have been proposed yet. To allow for learning on fully dynamic graph streams, i.e., graphs that can change in their structure (addition/deletion of nodes/edge) and in their node/edge attributes, we propose a Marked Neural Spatio-Temporal Point Process (MNSTPP). It leverages a Dynamic Graph Neural Network to learn a Marked TPP that handles attributes and spatial data to model and predict any event in a graph stream.