Learning Defenses in Computer Networks: Adversarial Machine Learning Approaches, Summer 2017

Computer systems are easy to attack if considered in a static scenario. The adversary has the advantage in time to study the system, find its vulnerabilities and choose the place to attack. To counter that, one can use the concept of moving target defense (MTD) by making the system dynamic and consequently more difficult for attacker to exploit since he also then has to deal with a great deal of uncertainty just like defenders do. This project aims at using adversarial neural networks concept in order to model the dynamics between the defender and attacker. The project will involve applying machine learning to investigate how to secure Peer-to-Peer networks against autonomous and adaptive adversaries. It is ideal for students planning on taking 8.857 and/or 6.858.


Contact: Erik Hemberg