Event Detection on Dynamic Networks

Period: September 2018 - Present
Location: UC Santa Barbara - Dynamo Project Type: Research

We proposed an event detection framework for dynamic graphs Graph Convolutional Networks. A very simple figure below illustrates, we are creating time series data from the dynamic graphs and trying to learn a pattern which can predict/detect an event. We implemented and tested the framework on multiple datasets. It is currently on submission (2020).

Event Detection Summary Figure