Hey! I'm investigating the label propagation algorithm as implemented in Neo4J's graph data science library. I think it's notable that it's possible to set concurrency parameters, even though the original (Raghavan et al.) paper on Label propagation does not specify explicitly a parallel algorithm. Is there any documentation how the parallel implementation works? Is the Staudt&Meyerhenke (PLP) version of the algorithm used, or some sort of coloring preprocessing?
Thanks in advance!