Discovering Network Structures from Data by Information-Theoretic Methods

[b]Speaker:[/b] Teemu Roos (HIIT - Univ. of Helsinki) [h2]Abstract[/h2] The interesting structure in data can sometimes be summarized as a graphical model. The nodes of the graphical model correspond to individual objects, or features of the objects, and edges between the nodes represent direct dependencies between them. Well-known instances of graphical models include, among others, Bayesian networks and phylogenetic trees. We present some recent applications of the information-theoretic Minimum Description Length (MDL) principle in discovering Bayesian networks and phylogenetic trees from data. [h2]Short Bio[/h2] Teemu Roos obtained his MSc and PhD degrees, both in computer science, from the University of Helsinki in 2001 and 2007, respectively. He is currently working at the Helsinki Institute for Information Technology HIIT as a postdoctoral researcher. His research interests include statistical and information-theoretic methods in data analysis and machine learning, and their applications.