This section covers the main GBML research communities. These communities are more disjoint than the methods they use and the lines between them are increasingly blurring. For example, LCS often evolve NNs and fuzzy rules, and some are powered by GP. Nonetheless, the differences between the communities and their approaches is such that it seemed most useful to structure this section by community and not e.g. by phenotype or learning paradigm; such integrated surveys of GBML are left to the future. Many communities have reinvented the same ideas yet each has its own focus and strengths and so each has much to learn from the others.