Understanding and Improving Belief Propagation

Een wetenschappelijke proeve op het gebied van de Natuurwetenschappen, Wiskunde en Informatica Proefschrift ter verkrijging van de graad van doctor aan de Radboud Universiteit Nijmegen, op gezag van de rector magnificus prof. mr. Contents Title page i Table of Contents iii 1 Introduction 1 1.1 A gentle introduction to graphical models 1 1.1.1 The Asia network: an example of a Bayesian network 2 1.1.2 The trade-off between computation time and accuracy 7 1.1.3 Image processing: an example of a Markov random field 9 1.1.4 Summary 16 1.2 A less gentle introduction to Belief Propagation 17 1.2.1 Bayesian networks 17 1.2.2 Markov random fields 18 1.2.3 Factor graphs 19 1.2.4 Inference in graphical models 20 1.2.5 Belief Propagation: an approximate inference method 21 1.2.6 Related approximate inference algorithms 24 1.2.7 Applications of Belief Propagation 24 1.3 Outline of this thesis 25