Inversion of hierarchical Bayesian models using Gaussian processes

Ekaterina I. Lomakina, Saee Paliwal, Andreea Oliviana Diaconescu, Kay Henning Brodersen, Eduardo A. Aponte, Joachim M. Buhmann, Klaas E. Stephan

Abstract: Highlights•The new method to optimise hierarchical Bayesian models•The Gaussian processes based method is able to avoid local minima.•The computational cost is comparable to conventional local methods.•The method was tested against MCMC and local methods.•The improved performance was demonstrated on both synthetic and real data.