Parallel and distributed computing for Bayesian graphical models

The goal of this project is to develop and analyze parallel and distributed algorithms for statistical inference in Bayesian graphical models and especially Bayesian state-space models. The primary target architectures of the project are graphics processing units (GPUs), internet of things, and cloud systems, but the algorithms can be applied to many other parallel computing systems. Within the project, the developed algorithms are applied to medical imaging, biomedical signal processing, and electromagnetic field mapping for indoor positioning. The impact of the project is ensured by implementing these applications in close business sector collaboration with companies.