• Home
  • Page : 'emploi' inconnue (menus.php)

 Thesis : Polarization in large socio-informational networks: models and measures

PhD school thesis
The prevalence of algorithmic recommendations in social media platforms has raised concerns about undesired societal effects. A central threat is the risk of polarization. Despite the numerous recent studies that investigate polarization, no clear answer has emerged as to what polarization is and how it should be measured, let alone what is the role of Recommender Systems in the phenomenon and what algorithmic design principles could be adopted to manage it. As many emerging phenomena in complex systems, polarization is hard to conceptualize, measure, and operationalize in Recommender Systems research. This doctoral project proposes to build on graph models for networks representing digital traces on large web platforms, and the development of measures from information theory, to understand and redefine polarization as a formal property of large complex networks, and the effect that recent Recommender Systems have on it. This highly interdisciplinary project will draw from recent advances but also from applied mathematics, graph formalisms, and empirical geometrical opinion space embeddings for social networks developed in the emergent field of computational social science.

Ce projet de recherche doctoral fait l’objet d’une demande de financement auprès de « Ecole Doctorale d‘Informatique, Télécommunication et d‘Electronique (EDITE) », le candidat retenu par son porteur devra donc participer au concours correspondant (prévoir un dossier et une audition) en vue d’obtenir le financement effectif.

More details here

Contact :Lionel Tabourier, Pedro Ramaciotti Morales