Computational perspectives on judgment aggregation
speaker: Umberto Grandi (Università di Padova)abstract: Societies of arti cial agents, i.e., multiagent systems, need procedures to take collective decisions. Researchers in Computer Science, in particu- lar in Arti cial Intelligence, have recently started borrowing techniques from Social Choice Theory to model and design systems of autonomous interacting agents. This multidisciplinary interest posed novel questions in collective decision making, from problems of compact representation of knowledge to the computational complexity of basic decision-making tasks. After providing a quick overview of the main research questions in the eld of Computational Social Choice, a research community that stems from the interaction between Arti cial Intelligence and Social Choice The- ory, I will focus on two on-going research topics in this area. First, I will show a characterisation of classical paradoxes in Social Choice Theory based on propositional logic, presenting a general model for the aggrega- tion of individual expressions which accounts for preferences, judgments and situations of voting in multi-issue domains. Second, by moving to the realm of voting theory, I will present the results of a computational simulation of iterated manipulation in electoral environments, modelling the response of voters to information from iterated polls.
timetable:
Mon 8 Jul, 16:00 - 16:30, Sala Azzurra
documents:
<< Go back