Giulia De Santis Email & Phone Number

After a Master Degree in Mathematics and Applications at the University of Camerino (Italy), and a Master Degree in Applied Mathematics at the Technical University of Clausthal (Germany) with a thesis in Operation Research (MAT/09) entitled "Analysis of Social Networks using Exponential Random Graph Models", I have obtained a Ph.D. in Computer Science at the University of Lorraine (France), with a working contract with the center of research INRIA Nancy - Grand Est, with a thesis entitled "Modeling and Recognizing Network Scanning Activities of the Ipv4 space through Hidden Markov Models".Thread in common of the thesis of the two Master Degrees and of the Ph.D. is the leveraging of models obtained empirically from real data. These obtained models are able to describe the observed phenomena. MOre precisely, the Master Thesis of the two Master Degrees leveraged ERGMs (Exponential Random Graph Models) to understand for what reasons the edges connecting nodes of a graph are present. In the study case, the presence of such an edge (or link) indicates the presence of a relation between two people, represented by nodes of the graph. It is thanks to ERGMs and leveraging features of nodes and edges that it was possible to determine motivations of choices made by individuals. The work done during the Ph.D. instead was perpetrated in a different context: more precisely, in Network Security. The Statistical Models (Finite Mixture Models and Hidden Markov Models) have been obtained from Big Data sets, collected by a Darknet. After a preliminary analysis, such Data has been leveraged to obtain Hidden Markov Models able to model Network Scanners, leveraging Machine Learning techniques. The so-obtained models have been then leveraged to recognize what Network Scanners generated freshly collected Big Data sets of logs.