Graphical representation of the two levels in the hierarchical Bayesian inference framework. The method of inference for individual galaxy parameters (see Sect. 4.1) is contained within the main box (denoted Galaxy i), while the full figure includes the hierarchical extension to the Bayesian network incorporating the population parameters (see Sect. 4.2). The grey outlines of multiple boxes in the background indicate the conditional independence between individual galaxy observations. The filled squares denote free sampled parameters, while the rectangle with a solid border denotes a deterministic node (the value is unambiguously determined from the input parameters and Eq. 2). The rectangle with the dashed border denotes a probability distribution with which the likelihood of the observed data, denoted with a circle, is evaluated. The observational error on ℛ_obs, σ_ℛ is a fixed input to the model.