InfoSec Seminar: MaMaDroid: Detecting Android Malware by Building Markov Chains of Behavioral Models.

Speaker: Enrico Mariconti

Date/Time: 09-Feb-2017, 15:30 UTC

Venue: Gordon Square (16-18) 101

Details

Abstract


Android malware is a rising threat. New malware samples are increasing day by day and app markets are facing challenges in keeping up and preventing the distribution of these apps.
In this talk I will present MaMaDroid, a system able to detect up to 98% of existing Android malware samples.  MaMaDroid works by building a behavioral model of Android apps, based on the sequence of API calls that these apps make, and then automatically distinguishing between benign and legitimate apps through machine learning. We show that MaMaDroid is not only effective in detecting current malware, but it also does not require frequent retraining, by being able to still detect a high fraction of malware one year after being trained.

Bio


Enrico Mariconti is a PhD student of the UCL Department of Security and Crime Science. He is part of the SECReT DTC and his research focus is mainly on Malware Detection. His studies at UCL started in 2014 under the supervision of Dr Gianluca Stringhini and Dr Gordon Ross. He has a background in Telecommunication Engineering (BSc and MSc at Università degli Studi di Genova) and, as part of the DTC, an MRes in Security Science.

Add to Calendar

This page was last modified on 27 Mar 2014.