A synopsis across the results from the three next stages in the Fundamental Universal Learning Patterns Adaptive Cellular Automata application. These span across new Moran’s I insights, identifying and reorienting the cell's rules, and implementing a mutation engine for better rule adaptation. The latest experiments continued to build upon the framework as well, with the seventh stage being implemented as a Thompson Sampling stochastic process, to switch between rules for each individual cell. A set of specific rules were defined in order for the cells to identify if there was a best rule or if the cell should switch between them to optimize survival, depending upon what it is currently experiencing per timestamp. This alone proved to only be marginally effective, as cells began to have “favorite rules” which they would latch on to, regardless if their internal rules were too big and noticeably superior. Due to this a Mutation Engine was created - borrowing from evolutionary cellular automata’s major adaptive mechanism - in order to better transition between states and goal. As this was also biologically aligned, it fit well with the original theoretical direction of the framework. The most recent results showed that there is still room for improvement and adjustment in future work; while maintaining healthy momentum in overall improved adaptation from the original mechanisms alone.
William Fullerton (2026) studied this question.