Jesse made a presentation about Dynamic State Networks - effectively graphs that get updated according to rules applied to a node depending on it's pattern of connections with other nodes.
The rule being applied on the nodes was an analogon of the Game of Life rule - nodes with too few neighbors die of loneliness, nodes with too many neighbors die of overcrowding, and so on. They defined classes of initial conditions according to the procedure by which the start network is constructed - vertex duplication, preferential attachment, grid exponential, ring exponential.
Then they defined two different measures of network entropy and showed how entropy varies along the time evolution of the rule. A clean computer experiment.
Comments