We conduct research on Bayesian inference, developing its methodological, computational, and applied aspects.
Special topics of interest are: Monte Carlo methods including Markov chain Monte Carlo and sequential Monte Carlo (particle filters); likelihood-free methods for models with intractable likelihoods (such as approximate Bayesian computation aka ABC); the role of improper priors; scalable methods for large spatio-temporal models.
Application areas include: modelling of protein folding data; tumour growth in mice; modelling of virus capsid assembly, forensic statistics, stochastic models for shape deformations, FitzHugh-Nagumo modelling in neuro science, ice core data.
We have regularly speakers at the Statistics and Biomathematics Seminar.
We have also a Bayesian reading group.
Senior members:
David Bolin (Bayesian methods for spatial and spatio-temporal models)
Olle Häggström (fundamentals of Bayesian reasoning, applications in futurology, Markov chain Monte Carlo)
Rebecka Jörnsten (data integration and network modeling for multi-omics single cell and bulk data)
Petter Mostad (Bayesian forensic statistics, theory for improper priors)
Umberto Picchini (likelihood-free inference, ABC, inference for SDEs, general Monte Carlo methods)
Moritz Schauer (Bayesian inference for SDEs and Markov processes, data assimilation, Nonparametric Bayes)
Current PhD students:
Felix Held
Anders Hildeman
Sandra Eriksson Barman
Anton Johansson
Ongoing projects:
2014-01-01--31-12-2019 VR
Statistical inference and stochastic modelling of protein folding
(Project leader Umberto Picchini)
2017-01-01 - 2020-12-31 VR
Latent jump fields for spatial statistics
(Project leader David Bolin)
Past projects:
2012-01-01 - 2017-12-31 VR
Improving Bayesian Network computational and modeling methods, with applications in forensics.
(Project leader Petter Mostad)