Belief and Opinion
Dynamics and Aggregation
in Multi-Agent Systems

The objective of the tutorial is to present a set of formal tools for modeling (i) how agents’ beliefs change and are merged in dyadic or group interaction through communication and joint observation, and (ii) how agents’ opinions spread in a social network through influence processes and aggregation mechanisms.
The toolbox includes epistemic logic (EL), dynamic epistemic logic (DEL), judgment aggregation, social-theoretic and game-theoretic models of social influence and opinion diffusion. The tools presented in the tutorial are useful for understanding a large variety of social phenomena and for several applications in AI and in the domain of multi-agent systems (MAS). This includes diffusion of fake news, gossips and rumours in social networks and media; modeling Theory of Mind (ToM) in human-robot interaction; modeling persuasion mechanisms in human-agent interaction.

Program & Slides

  • Belief Dynamics and Aggregation [slides]
    * Epistemic logic (EL)
    * Dynamic epistemic logic (DEL)
    * The logic of distributed belief and distributed knowledge
    * Compact semantics for EL and DEL exploiting belief bases
    * Distributed belief vs. belief merging

  • Opinion Dynamics and Aggregation [slides]
    * Review of opinion diffusion and social influence models
    * Judgment and goal aggregation
    * Propositional opinion diffusion
    * Diffusion of preferences and constrained opinions
    * Strategic opinion diffusion

Lecturers & Contact

Umberto Grandi (IRIT, University of Toulouse) | umberto.grandi[at]irit[dot]fr

Emiliano Lorini (CNRS, IRIT, Toulouse) | emiliano.lorini[at]irit[dot]fr

Arianna Novaro (ILLC, University of Amsterdam) | a.novaro[at]uva[dot]nl