Automated Malware Classification/Analysis Through Network Theory and Statistics presented at BlackhatUSA 2006

by Daniel Bilar,

Summary : Automated identification of malicious code and subsequent classification into known malware families can help cut down laborious manual malware analysis time. Call sequence, assembly instruction statistics and graph topology all say something about the code. This talk will present three identification and classification approaches that use methods and results from complex network theory. Some familiarity with assembly, Win32 architecture, statistics and basic graph theory is helpful.