The collective intelligence of asymmetric learning promotes cooperation on collaborative networks

Abstract The learning activities in collective intelligence have inspired many collective behaviors, such as self-organization, which is extremely important for human society. Most learning relations are unilateral or asymmetrical, depending on social status. In particular, the status involving asymmetric learning, which is characterized by nodes with different degrees in social networks, affects how the collective intelligence responds to the evolutionary environment, especially its collective cooperation behavior. In order to figure out how both high degree ( H ) and low degree ( L ) individuals behave, we introduce an asymmetric learning method, where individuals respond to the environment in the opposite way characterized by an asymmetric parameter. It is found that there exists a range of asymmetric parameters with the optimal promotion of cooperation. A conspicuous cluster has emerged by dividing all individuals into four different clusters according to their strategies at the given asymmetric parameter. This cluster consists of individuals who devote their utmost resources to investment. Remarkably, the preponderant majority of these individuals possess high levels of connectivity and, driven by the cumulative payoff effect, display a pronounced propensity to engage in cooperative behaviors. By contrast, within small clusters, a substantial quantity of individuals, notwithstanding their relatively high payoff coefficient, frequently encounter cooperation predicaments. A particularly salient finding is the vulnerability of H individuals positioned on medium-connected nodes to the influence of asymmetric learning modalities. The triggering and subsequent diffusion of cooperative behavior throughout the population is contingent upon the fulfillment of two cardinal conditions: the existence of inborn altruistic behavior exhibitors on super hubs and a diminished self-centered learning framework among the H individuals. This phenomenon holds significance as it could deepen our understanding of the system and offer potential ways to restructure its overall dynamics, leading to more efficient cooperative outcomes.