Collective Behavior from Surprise Minimization

Explore a 52-minute conference talk examining how collective behavior emerges in animal groups through the lens of surprise minimization. Delve into an innovative active inference framework that moves beyond traditional "social forces" models to understand group dynamics in fish, birds, and insects. Learn how individual agents can be modeled as probabilistic decision-makers maximizing evidence for their internal world models, generating natural patterns of cohesion and movement without predetermined rules. Discover how uncertainty beliefs at the individual level influence collective dynamics, and examine how real-time belief adaptation enables groups to better encode information and respond to changes. Gain insights into the intersection of cognition and emergent group behavior, with applications spanning biology, physics, and engineering of multi-scale intelligent systems.