Our work on multi-probabilistic decision assessment for Android Malware detection has been accepted at 9th ACM Workshop on Artificial Intelligence and Security (AISec 2016). The paper is titled “Prescience: Probabilistic Guidance on the Retraining Conundrum for Malware Detection”.

We will be presenting a poster on our Android  detection and classification framework titled “Barometer: Sizing Up Android Applications Through Statistical Evaluation” at S&P 2016. The differentiating factor of our work compared to run-of-the-mill classifiers would be that we assess each decision by the classifier statistically and use fitting measures to improve our classification model over time.

Our work on dynamic analysis of Android malware for classification of malware has been accepted at Mobile Security Technologies (MoST’ 2016).

Marco Cova of Lastline gave a great talk on building industrial scale malware analysis systems at the Information Security Group seminars on 04/12/2014. I enjoyed the talk as well as talking to Marco about the challenges in building malware analysis systems that scale to large feature sets.