Statistical Modeling and Prediction for Dynamic Networks of Social InteractionsDr. Kevin Xu AbstractSignificant efforts have gone into the development of statistical models for analyzing data in the form of networks, such as social networks. While most existing work has focused on modeling static networks, there has been recent interest in modeling dynamic networks, which are observed at multiple points in time. In this talk, I present a state-space model for dynamic networks of social interactions as well as a computationally efficient approximate inference procedure, which I use to reveal the temporal evolution of an email network and a network of physical proximities between individuals. I then turn to the task of predicting future interactions—can it be done, and if so, what implications could it have? BiographyKevin S. Xu received the B.A.Sc. degree in Electrical Engineering from the University of Waterloo (2007) and the M.S.E. (2009) and Ph.D. (2012) degrees in Electrical Engineering: Systems from the University of Michigan. He is currently a researcher at the Technicolor Palo Alto Research Center. From 2012-2013, he was a Senior Research Scientist at the 3M Corporate Research Laboratory, St. Paul, MN. Kevin Xu was a recipient of the Natural Sciences and Engineering Research Council of Canada (NSERC) Postgraduate Master’s and Doctorate Scholarships. His main research interests are in statistical signal processing and machine learning with applications to network science and human dynamics. |