Proactive breach detection with security analytics

Abstract

Security breaches are on the rise as attackers continuously evolve their strategies and traditional defensive methods are largely ineffective at keeping up. In this talk, I will discuss how machine learning techniques can be used to develop more effective defenses in face of increasingly powerful adversaries. I will describe the design of an analytics platform that processes large volumes of log data with the goal of building practical tools for proactive breach detection. The techniques have been successfully used in operational settings and transferred to the industry. I will also describe some of the challenges of making machine learning models resilient against adversarial manipulation.

Bio: Alina Oprea joined Northeastern University's College of Computer and Information Science as an Associate Professor in August 2016. Alina is interested in extracting meaningful intelligence from different data sources for various security applications, designing rigorous machine learning techniques for learning and predicting sophisticated attackers' behavior, and protecting cloud infrastructures against emerging threats. Prior to her position at Northeastern, Alina was a consultant research scientist at RSA Laboratories (2007-2016), where she performed research in cloud security, applied cryptography, foundations of cybersecurity, and security analytics. Alina received a BS in mathematics and computer science from the University of Bucharest in Romania. She also earned M.Sc. and Ph.D. degrees in computer science from Carnegie Mellon University in 2003 and 2007, respectively.