This work examines the use of graph grammars for learning the building blocks of complex networks and building additional instances of a given observed network. The focus is on the use of vertex replacement grammars and their ability to build instances of an observed graph that are much larger than what was seen. Parallel implementations of graph generation are examined as are more complex generation models that allow for more growth in the generated instance. Additionally, methods for adding node metadata into the generation model are explored.
BlueJeans Link: https://bluejeans.com/959267792/3856