Bridging across scales in animal behavior

Recent years have seen an explosion in our ability to measure the movement of animals in naturalistic environments. However, such data poses tremendous challenges for quantitative analysis, as behavior is generally partially-observed, unpredictable, multiscale and far from equilibrium. I'll discuss how we can leverage first principles to draw simple yet predictive models of animal behavior across scales: from posture to navigation; from single animals to collectives. Modelling the dynamics through transfer operators, I'll discuss how we can coarse-grain the dynamics into slow collective modes that capture long lived behavior. Finally, I'll introduce a picture of behavior as evolving in potential landscapes that change in time, and show how such non-stationary dynamics captures the observation of heavy-tailed statistics in foraging animals.