Automated malware detection using supervised machine learning

Malware is one of today’s biggest computer security problems. Traditionally, malware is detected by inspecting its signature, which is a unique identifier of the software’s binary. Signature-based malware

detection is becoming less effective as cyber criminals mutate the malware they design, where the malware is programed in such a way that simply comparing its signature to known malware signatures

will not identify it as being malicious. Thus, new detection techniques inspect the behavior of the software instead, where an algorithm is used to learn the patterns of malware activities. Usually, this

is achieved by using supervised machine learning, where the malware detection system, called the classifier, is trained using already identified malware samples. In this research project, we propose to

design, implement and diagnose a supervised machine learning architecture for automating malware detection at Sophos Inc., one of the leading anti-malware software vendors.

Faculty Supervisor:

Matei Ripeanu

Student:

Partner:

Sophos Inc

Discipline:

Engineering

Sector:

University:

The University of British Columbia

Program:

Accelerate