Improved Majority Identification by the Coarsened Majority Automaton

The initial majority identification task is a fundamental test problem in cellular automaton research. To pass the test, a two-state automaton has to attain a uniform configuration consisting of only the state that was initially in the majority. It does so solely through its local, internal dynamics—i.e., success in the task is an example of emergent computation. Finding new, better-performing automata continues to be of interest for what additional clues they might reveal about this form of computation. Here we describe a novel, coarsened version of one of the standard majority identifiers. We show that this coarsened system outperforms its parent automaton while significantly reducing the number of computations required to accomplish the task.