Group Decision Making on Social Networks: Computational Models, Algorithms, and Applications

Speaker: Dr. Salehi-Abari (FBIT)

Abstract: With the explosive growth of choices and information, we have more freedom to choose at the expense of dealing with many overwhelming decisions. Intelligent Decision Systems helps us to navigate decisions and information spaces. Of practical import are group decision problems, which address the problem of choosing an outcome or a decision, from a set of alternatives, for a group of individuals. These decision problems, in the physical and online worlds, are prevalent: for instance, when a group of friends chooses a movie to watch, a resort at which to vacation, or a restaurant at which to eat; when a company decides on assigning various advertisements to its users based on the relevance of ads and user preferences; selecting a policy for an online system or a nation given individuals’ preferences, etc. One of the main challenges in any group decision problem is learning the individual preferences upon which decisions are based. In this talk, I argue that social networks—by capturing preference correlations across individuals induced by social interactions—provide a natural platform for preference learning. By modelling empathy, preference dynamics, and correlations over social networks, I introduce efficient learning algorithms for group decision making which require less user data. Since these algorithms produce better decisions with less information collected from users, they help solve the preference learning problem, which in turn reduces user’s cognitive and communication burden and protects their privacy.