The Bayesian Superorganism I: collective probability estimation

摘要: Abstract Superorganisms such as social insect colonies are very successful relative to their non-social counterparts. Powerful emergent information processing capabilities would seem contribute abundance, they explore and exploit environment collectively. In this series of three papers, we develop a Bayesian model collective processing, starting here with nest-finding, then examining foraging (part II) externalised memories (pheromone territory markers) in part III. House-hunting Temnothorax ants adept at discovering choosing the best available nest site for colony. Essentially, propose that estimate probability each choice is best, choose highest probability. Viewed way, behavioural algorithm can be understood sophisticated statistical method predates recent mathematical advances by some tens millions years. Here, finding incorporates insights from approximate computation estimation alternative choices; Thompson sampling, an effective regret-minimising decision-making rule viewing terms multi-armed bandit problem (Robbins, 1952). Our framework points potential further bio-inspired techniques. It also facilitates generation hypotheses regarding individual movement behaviours when decisions must made.