Speaker
Michael A Lomholt
(University of Southern Denmark)
Description
Recently, a competition (the AnDi challenge) was held where different research groups competed about analysing synthetic single particle tracking data from anomalous diffusion models, with the aim of inferring the model and anomalous exponent which was used to generate the data. In this talk, I will argue how this task is in principle solved optimally by Bayesian inference, but with the downside being computational efficiency for models with hidden variables.
