Am 06.02.20 um 12:09 schrieb Davide Cittaro:

Hi, I'm running the nested version of nSBM, I'm collecting the group marginals using the code from gt documentation, basically counting the number of non empty blocks for each hierarchy level for each iteration:

group_marginals = [np.zeros(g.num_vertices() + 1) for s in state.get_levels()] def _collect_marginals(s): levels = s.get_levels() for l, sl in enumerate(levels): group_marginals[l][sl.get_nonempty_B()] += 1 […]

At the end of the equilibration I look at the distributions and, in general, the most probable number of blocks at each level is not the one that is stored in the final state, although the final number of blocks is typically the second most probable. I may be naive, but I expected the two to be the same.

There is no guarantee that the mode of a distribution needs to be equal to the mean. Indeed, posterior averages often diverge from point estimates with the maximum likelihood. I talk about this in this paper (look at Fig 6 which shows exactly what you see): https://arxiv.org/abs/1610.02703 Best, Tiago -- Tiago de Paula Peixoto <tiago@skewed.de>