Poster: Mobile Malware Detection using Multiple Detector Set Artificial Immune System

As mobile devices become increasingly more powerful and important in everyday life, the need for efficient and effective detection of mobile malware has become pressing. We developed a multi-detector set Artificial Immune System (mAIS) to classify apps into benign and malicious categories based upon information flows within the app. The performance of mAIS has been compared with the performance of a variety of conventional Artificial Immune Systems (AISs) using a featureset of information flows captured from malicious and benign Android applications. Our preliminary results show that the mAIS outperforms the conventional AISs in terms of accuracy and false positive rate as well as the computational complexity of the negative selection process. We plan to replicate the study on a large set of mobile applications. Keywords— Mobile Malware Detection; Artificial Immune System; Multiple Detector Set; Evolutionary Computing