Title: "Data-driven Modeling of Collective Behavior in Animal Groups"
Abstract: This talk will describe data-driven modeling of collective behavior in two animal groups: fish and mosquitoes. Fish-behavior experiments reveal the role of intra-school interactions in rheotaxis and startle response. We model these behaviors using tools from physics and epidemiology, namely the Kuramoto model of coupled oscillators and the Susceptible-Infected-Removed (SIR) epidemic model. Swarming and pursuit behaviors in wild malarial mosquitoes were reconstructed to reveal three-dimensional flight patterns. By studying the auto- and cross-correlation velocity statistics, these patterns are well-modeled using a mechanistic approach based on Newtonian particles.