Time: 5/4(Wed) 3.30-5:00 PM

Location: EC122

Topic:

Machine Learning in Network Anomaly Detection

Speaker:

Winston K.G. Seah, Professor of Network Engineering, School of Engineering and Computer Science, Victoria University of Wellington, New Zealand

Abstract:

Network anomaly detection is a broad research area that goes beyond security-related issues, like intrusion and malware detection, and encompasses detecting faulty network devices, usual traffic patterns, misconfigurations of network devices, etc. Very broadly, it refers to finding patterns in network data that do not conform to expected behavior. The statistics community has been studying the anomaly/outlier detection problem from as early as the 19th century. More recently, machine learning has taken an increasing significant role in anomaly detection in general, not just within the domain of networks. In this talk, we shall discuss two approaches in applying machine learning to detect network anomalies. Firstly, we derive a network graph from the individual routers' data and then apply machine learning to detect the occurrence of Border Gateway Protocol (BGP) hijacking. We validated this approach using real network data and demonstrated the ability to detect impending anomalous events prior to their occurrences. To deal with ever-changing normal network conditions, we propose a hybrid online-offline framework that enables online and offline methods to work together in order to deal with changing normal conditions. New anomalies that emerge can then be successfully detected while new normal conditions are not erroneously identified as anomalous.

Biography:

Winston K.G. Seah received the Dr.Eng. degree from Kyoto University, Kyoto, Japan, in 1997, and ME and BSc degrees from the National University of Singapore in 1993 and 1987 respectively. In 2009, he joined Victoria University of Wellington, New Zealand as a Professor of Network Engineering. Prior to this, he has worked for more than 16 years in mission-oriented industrial research, taking ideas from theory to prototypes, most recently, as a Senior Scientist (Networking Protocols) in the Institute for Infocomm Research (I2R), Singapore. Being actively involved in research in the areas of mobile ad hoc and sensor networks, he developed the first Quality of Service (QoS) models for mobile ad hoc networks and first to study networking protocols for wireless sensor networks powered solely by ambient energy harvesting. His latest research focuses on networking protocols and the use of artificial intelligence and machine learning techniques to address the needs of 5G networks, the Internet of Things (IoT) and other machine-type communications (MTC). He is a Senior Member of the IEEE (S'90-M'97-SM'11) and a Professional Member of the ACM.