Type: | kursus på phd-niveau, udbydes under åben uddannelse Sprog: engelsk |
|
|
|
|
|
|
Vejledende placering: Sidst i studiet.
|
Undervisningsform: Lectures and exercises
|
Evalueringsform Rapportaflevering
|
Karakter: bestået/ikke bestået
|
|
|
Kontaktperson: | Carl edward Rasmussen |
|
|
Institut: Informatik og Matematisk Modellering
|
|
Kursusmål: The aim of this course is to discuss various topics that are relevant to non-linear modelling in situations where no parametric model of the system is available. We will address fundamental issues involved, and describe methods including both Bayesian and non-Bayesian approaches. The choice of topics is not exhaustive but rather governed by our personal experiences and inclinations. Full description including schedule is available.
|
Kursusindhold: Probabilistic model fitting, Generalisation estimation, Regularisation, Bayesian learning, Markov Chain methods, Flexible metric kernel methods, Neural networks, variational methods, Gaussian Processes, Infinite Mixtures.
|