Abstract: Aggregations of social animals, such as flocks, schools, herds, and swarms, are beautiful examples of self-organized behavior far from equilibrium. Such collectives have been the focus of a significant research effort in recent years, from many different perspectives. Biologists aim to understand the evolutionary benefits of acting together; physicists treat aggregations as examples of active matter; and engineers see them as potential templates for designing robust autonomous distributed systems. All of these goals require modeling, which typically assumes that every individual in the group is an identical agent playing by the same immutable rules. In reality, however, interactions are likely to be influenced by many factors, both internal to the group (such as social relationships) and external (such as ecological context).

I will discuss the impact of heterogeneity on flocking behavior using field data from flocks of jackdaws, a highly social corvid species that forms lifelong pair bonds. Using 3D stereoimaging, we measured the flight trajectories and kinematics of flocking jackdaws during the winter roosting season in Cornwall, UK. We found that mated pairs were indeed present inside these larger flocks, with significant consequences. Paired birds on average interacted with fewer of their neighbors, translating to an individual-level cognitive and energetic savings. However, the responsiveness of the flock as a whole, as measured by its correlation length, monotonically decreases with the proportion of paired birds, indicating a global cost balancing the local benefit. In contrast, measurements made in the summer nesting season in the presence of a model predator show wholly different behavior, with the jackdaws following a metric rather than a topological interaction rule. Our results have intriguing implications for the interplay and evolution of cognition and flocking behavior, as well as for the design of interacting engineered systems of heterogeneous autonomous agents.

Bio: Dr. Nick Ouellette is broadly interested in the behavior of complex systems far from equilibrium. In particular, a running theme in his research is dynamical self-organization. He seeks both to understand the physical principles that govern the spontaneous emergence of low-dimensional structure in high-dimensional systems and to harness this self-organization for engineering applications. His current research includes studies of turbulent flows in two and three dimensions, in both simple and complex fluids; the transport of inertial, anisotropic, and active particles in turbulence; the erosion of granular beds by fluid flows and subsequent sediment transport; quantitative measurements of collective behavior in insect swarms and bird flocks; and emergent, self-organized structure and dynamics in cities.

Before coming to Stanford in 2015, Ouellette spent seven years on the faculty in Mechanical Engineering and Materials Science at Yale University. He has won awards for his teaching at both Yale and Stanford. Before beginning his faculty career, he held postdoctoral positions at the Max Planck Institute for Dynamics and Self-Organization and in the Physics Department at Haverford College