A principled approach to characterize the hidden structure of networks is to
formulate generative models, and then infer their parameters from data. When
the desired structure is composed of modules or "communities", a suitable
choice for this task is the stochastic block model (SBM),