distances and similarities as edge weights

when performing inference on a graph using SBM including edge weights, does it make any difference if weights are intended as distances or similarities between the vertexes?


The weighted SBM is a model for edge covariates taken from a set of
possible distributions. Other than the nature of the support (discrete
or continuous, bounded or not, etc), and the shape of the distribution,
the model makes no assertion on how the weights are to be interpreted.