Abstract:
A major obstacle which limits the possibilities of a number of machine learning techniques like the genetic algorithm is the lack of mechanisms which allow for the dynamic construction of hierarchically organized solutions or artifacts. In the context of the GA, such mechanisms would require some form grouping or social behavior. Yet, this type of behavior is per definition maladaptive in the competitive context of the GA. Here, a biological model is analyzed and modeled which allows this social behavior to emerge. This model provides a first step for the computational synthesis of complex structures.