Who: Nicolas Maudet
Where: Univ. of Padova (Francesca Rossi)
When: 28th of March — 1st of April

The purpose of this STSM was to study models of influence among agents in combinatorial domains. Influences are typically represented as networks capturing potential modifications in an agent's view when observing choices made by fellow agents.
The work carried over during this week has been very promising and involved Francesca Rossi, as well as Maria Silvia Pini and Brent Venable. Our starting point was that in settings involving preferences over combinatorial domains, there is a need to express in more detail what variables of the domain may be influenced. Focusing on CP-nets, we were able to show what it takes to integrate such influences within this model, and proposed a first approach to interleave influence propagation with the computation of social outcomes. A first draft document based on these ideas was produced.