Authors: Jelena Milosevic Andreas Dittrich Alberto Ferrante Miroslaw Malek Camilo Rojas Quiros Rubén Braojos Giovanni Ansaloni David Atienza 41st Computing in Cardiology Conference, CinC 2014, Cambridge, MA, USA, September 7-10, 2014 Download: author version, final published version

We present a methodology for identifying patients who have experienced Paroxysmal Atrial Fibrillation (PAF) among a given subjects population. Our work is intended as an initial step towards the design of an unobtrusive system for concurrent detection and monitoring of chronic cardiac conditions.

Our methodology comprises two stages: off-line training and on-line analysis. During training the most significant features are selected using machine-learning methods, without relying on a manual selection based on previous knowledge. Analysis is based on two phases: feature extraction and detection of PAF patients. Light-weight algorithms are employed in the feature extraction phase, allowing the on-line implementation of this step on wearable and resource-constrained sensor nodes. The detection phase employs techniques borrowed from the field of failure prediction. While these algorithms have found extensive applications in diverse scenarios, their application to automated cardiac analysis has not been sufficiently investigated.

Obtained results, in terms of performance, are comparable to similar efforts in the field. Nonetheless, the proposed method employs computationally simpler and more efficient algorithms, which are compatible with the computational constraints of state-of-the-art body sensor nodes.

	Authors: Jelena Milosevic Andreas Dittrich Miroslaw Malek Alberto Ferrante 3rd International Workshop on Security of Mobile Applications, IWSMA 2014, Fribourg, Switzerland, September 8-12, 2014 Download: accepted version, final published version

With explosive growth in the number of mobile devices, the mobile malware is rapidly spreading. Existing solutions, which are mainly based on binary signatures, are no longer effective, making security one of the key issues. The main contribution of this paper is a novel methodology to design and implement secure mobile devices by offering a resource-optimized method that combines efficient, light-weight malware detection on the device with high precision detection methods on cloud servers. We focus on early detection of behavioral patterns of malware families rather than the detection of malware binary signatures. Together with the alarm about the device being attacked, damage that detected type of malware can cause is estimated. Furthermore, the database with behavioral patterns is continuously updated, thus keeping a device resistant to new malware families.