Doctoral defense by Conor Heins

This doctoral thesis delves into the convergence of complex systems science, cognitive science, and statistical inference, to probe the phenomena of collective behavior—where autonomous entities such as cells, birds, or humans exhibit coordination that eclipses their individual capacities. Central to this inquiry is the application of active inference—a theoretical framework grounded in the principles of statistical physics and information theory, initially developed for understanding the human brain's functioning—to the domain of collective behavior. Encompassing a series of studies, this thesis employs the Free Energy Principle (FEP) to model group dynamics, portraying collective behavior as the emergent result of distributed action among active inference agents. This framework posits that systems at all scales engage in a form of inference that synthesizes perception and decision-making in the face of uncertainty, thereby offering a new lens through which collective behavior can be interpreted across scales of organization. Structured into six chapters, each representing a peer-reviewed publication, the dissertation establishes a basis for the application of active inference and FEP to collective phenomena. The scope of discussion ranges from the core tenets of Bayesian mechanics and the examination of stochastic chaos in coupled systems, to the development of computational tools for active inference. Further exploration includes simulation studies on opinion dynamics and collective motion, as well as theoretical models elucidating the nexus between individual cognition and collective computational processes.