Seminar by Anthony Couthures

Analysis of a model of opinion dynamics coupled with dynamics of the external environment

Why can public opinion influence behaviour – ranging from consensus to resilient ‘echo chambers’ – or fluctuate over time? This lecture unpicks the mathematical mechanism underpinning these macroscopic social dynamics using a non-linear communication model. 

By varying the non-linearity of the amplification of the communication signal as a function of the graph’s connectivity, we can analytically predict the emergence of polarisation within a social network.

Furthermore, by coupling the social network to a shared environment via a feedback loop, the dynamics exhibit new behaviours. 

Depending on whether agents prioritise their social neighbours or the environmental reality, the system undergoes Pitchfork- and Hopf-type bifurcations. 

Ultimately, this leads to a unified mathematical framework that successfully accounts for four distinct behaviours observed in the real world: constant and oscillating consensus, as well as constant and oscillating polarisation.

Anthony Couthures is a postdoctoral researcher at the ICTEAM Institute of the Catholic University of Louvain (UCLouvain) in Belgium. 

Previously, he obtained his PhD in automatic control, dynamical systems and applied mathematics at the Nancy Centre for Research in Automatic Control (CRAN), a joint research unit of the CNRS and the University of Lorraine, in France.

His research lies at the intersection of game theory, dynamical systems, distributed optimisation and multi-agent systems, with applications to socio-environmental systems, opinion dynamics and decarbonisation strategies aimed at mitigating climate change.