Bayesian inference and experimental design for large generalised linear models [Elektronische Ressource] / vorgelegt von Hannes Nickisch

Bayesian Inference and Experimental Design forLarge Generalised Linear Modelsvorgelegt vonDipl.-Inf. Hannes Nickischaus LeipzigVon der Fakultät IV - Elektrotechnik und Informatikder Technischen Universität Berlinzur Erlangung des akademischen GradesDoktor der Naturwissenschaften– Dr. rer. nat. –genehmigte Dissertation.Promotionsausschuss:Vorsitzender: Prof. Dr. Klaus-Robert MüllerBerichter: Prof. Dr. Manfred OpperBerichter: PhD. Matthias W. Seeger PhD. Carl E. RasmussenSachverständiger: Prof. Dr. Klaus Obermayer Prof. Dr. Felix WichmannTag der wissenschaftlichen Aussprache: 17. September 2010Berlin 2010D83AcknowledgementsThe most important ingredient for this work was the open, cordial and scientific atmosphereat the Empirical Inference Department of the Max Planck Institute for Biological Cybernetics,Tübingen. I am thankful to Bernhard Schölkopf for creating this stimulating as well as produc-tive environment and giving input and feedback whenever necessary.Thanks to Carl Rasmussen, I learned about Gaussian processes, approximate inference, dif-ferentiating between the “big picture” and “technical details” and concise scientific program-ming, experimenting and writing.Matthias Seeger’s enthusiasm, experience in probabilistic modelling and knowledge onmathematics and numerical optimisation where a constant driving force and source of ideas.Without his supervision, this thesis wouldn’t have been possible.