Machine Learning for Detection of Malicious Advertising

Project Title:

Machine Learning for Detection of Malicious Advertising

Supervisors:

Dr Iqbal Gondal (SoEITPS), Dr Taiwo Oseni (SoEITPS)

Contact Person:

Dr Iqbal Gondal Iqbal.gondal@federation.edu.au

Project Brief

Malicious advertising is an increasingly prevalent vector of malware infection. OCSC funded research has produced a proof of concept system for the detection of malicious advertising. This research detects malware by taking a memory dump of the virtual machine after loading each advertising page, the memory dump is then scanned to detect any installed malware. This memory dumping approach is slow, taking an average of 15 minutes to perform each memory dump, and anti-virus detection rates of new malware are low and could potentially fail to detect webpages loaded with new malware variants. This research calls for the design of a virtual machine introspection (VMI) based technique using machine learning for the detection of malware infections resulting from loading of webpages containing potentially malicious advertising.