NFFDy Summer Programme 2024: Bayesian Data assimilation and Model Selection

Date:

This workshop is part of the National Fellowships in Fluid Dynamics Summer Programme 2024, funded by the UK EPSRC.

In this 1-day workshop, Prof. Matthew Juniper and I have gone through the basic idea of how to infer parameters in a model by assimilating data in a Bayesian framework, and how to select the best model among many possible models based on the evidence (marginalized likelihood), i.e. by a type-II maximum likelihood optimisation. Then, we introduced Bayesian-SINDy, a Bayesian recast of the classical SINDy algorithm.

For the next half of the workshop, I continued with some tutorial on how to use Turing.jl to infer parameters in a channel flow.

Material for the tutorial can be found in the following link:

Tutorial for Bayesian-SINDy

Tutorial for Turing.jl