Probabilistic graphical models: Factor graphs and more

Probabilistic graphical models together with dedicated algorithms such as belief propagation and other message passing algorithms allow a unified approach to a multitude of problems arising in signal processing, coding theory as well as in statistical inference in general. The following talk will briefly cover Bayesian networks and Markov random fields before focussing on factor graphs and related algorithms such as sum-product message passing. Various toy examples as well as an input estimation and a Bayesian inference algorithm using the factor graph description language will be shown.