Graph-Centric Machine Learning: Algorithms, Systems, and Cybersecurity Applications

H. Huang
George Washington University,
United States

Keywords: graph analytics, cybersecurity, machine learning


Today we live in a highly connected world. Our society, commerce, and national security all depend on a myriad of networks that range from transportation networks to power grids, from biological networks to healthcare networks, from wearable devices to autonomous vehicles, and from computer networks to social networks. In Graph Computing Lab at the George Washington University, we are developing novel graph-centric machine learning algorithms and systems to manage big data generated by these networks, to understand the contextual and causal relationships within entities and events, and to deliver actionable knowledge to stakeholders in real time. In this talk, I will share our experiences in designing and developing high-performance graph systems, and discuss our techniques for addressing the algorithmic, computational, and I/O challenges in graph computing. In addition, I will present our ongoing work on utilizing these graph systems for understanding and analyzing complex network data for cyber threat hunting.