We are honored to have Prof. Yevgeniy Vorobeychik as our invited speaker at CoopMAS; talk details below.
Collective decision making is a problem of fundamental importance in multi agent systems. This problem exhibits a multitude of conceptual and technical challenges. In this talk, I focus on three of these:
Considering the issue of robust social choice, we investigate the problem of group-level election control and protection. We model optimal protection of elections as a Stackelberg game in which the defender, who aims to preserve election integrity by deploying limited resources to protecting voter groups, while an adversary aims to subvert the election by attacking these groups. We show how this problem can be addressed in a scalable manner using linear programming combined with a double-oracle method, with oracles representing the best-response sub-problems for both the attacker and defender.
Our final foray is into hedonic coalition (or team) formation, where we focus on maintaining incentives to honestly reveal preferences while achieving a series of important properties, such as efficiency and equity. Our first contribution is in defining a novel property of IMS (iterative matching of soulmates), which we view as a fundamental criterion in team formation. We show that implemeting IMS has several important side-effects, some economic (such as strong incentive compatibility and core uniqueness on restricted preference domains), others computational. We also exhibit a novel mechanism implementing IMS which is also Pareto efficient and individually rational, and has remarkably good empirical incentive and equity properties.
The talk by J. Low and T. Rahwan - Efficient Synergy Computation under MC-Nets, is cancelled due to unforseen circumstances