Fletcher, P. (2001) Connectionist learning of regular graph grammars. Connection Science, 13, no. 2, 127-188.
Each example involves a population of patterns describable by a graph grammar. The examples of are of steadily increasing complexity, beginning with a simple string-like example and progressing to larger examples with iteration in two dimensions. For each population the algorithm is able to learn a grammar that successfully represents the population.
In all these examples, the patterns are drawn on a rectangular image grid consisting of nodes, with horizontal and vertical neighbours connected.