“Anomaly Detection Tool for Software/Hardware Protection”

Location: LTS Auditorium, 8080 Greenmead Drive

Speaker:
Michael Azarian
Research Scientist at the Center for Advanced Life Cycle Engineering (CALCE)
Department of Mechanical Engineering

Abstract:
This presentation will describe proposed research targeting a weighted isolation forest-based method for the identification of anomalous behavior of workstations activated by software and/or hardware anomalies.

The method takes advantage of the fact that: (i) anomalies are the minority consisting of few instances and (ii) anomalies are characterized by attribute values that are very different from those of normal instances.

The proposed anomaly detection tool represents the first known attempt to detect both software and hardware anomalies in computer systems for security.

Speaker Bio:
Michael H. Azarian is a research scientist at CALCE at the University of Maryland.

Azarian’s primary research interests are detection, prediction and analysis of failures in electronic components and assemblies.

He has over 150 publications on electronics reliability and packaging, prognostics and health management, and tribology, and holds 5 U.S. patents.

Prior to joining CALCE, Azarian spent over a dozen years in the disk drive and fiber optics industries.

He is chair of the SAE G-19A Test Laboratory Standards Development Committee on detection of counterfeit parts, which is developing the AS6171 family of standards. Azarian is also co-chair of the Miscellaneous Techniques subcommittee of SAE G-19A. Azarian has led several standards committees on reliability for the IEEE.

He received his doctorate in materials science from Carnegie Mellon University.