Graphical Models and Network Science

Networks are an important paradigm for modelling various real world processes, such as genetics, finance and social phenomena. Mathematically, the network process can either live on the vertices – such as genes activities – or on the edges – such as friendship relationships. To model these phenomena we can use graphical models and random graph models, respectively. In this course, we will focus on both forms of probabilistic network modelling to describe large scale networks. For random network models, we describe p1, stochastic block models and exponential random graph models and focus on inference methods, such as generalized linear models and extensions thereof. In the second part of the course, we describe graphical models, the underlying Markov properties and inference methods, such as the graphical lasso.