Invertible Neural Networks Versus MCMC for Posterior Reconstruction in Grazing Incidence X-Ray Fluorescence

Grazing incidence X-ray fluorescence is a non-destructive technique for analyzing the geometry and compositional parameters of nanostructures appearing e.g. in computer chips. In this paper, we propose to reconstruct posterior parameter distribution given noisy measurement generated by forward model an appropriately learned invertible neural network. This network resembles transport map from reference posterior. We demonstrate numerical comparisons that our method can compete with established Markov Chain Monte Carlo approaches, while being more efficient flexible applications.