Trust and Reputation in Social Networks

Trust is a statement (or prediction of reliance) about what is otherwise unknown or uncertain -- for example, because it is far away, cannot be verified, or is in the future. Trust is pervasive and beneficial in complex social systems. It can be built from direct interactions between the source party (truster) and the target (trustee). However, in large open systems, it is infeasible for each party to have a direct basis for trusting another party. Therefore, the participants in an open system must share and use information about trust (reputation). When using shared trust information, trusters have to decide how to use and trust such shared information. Different approaches for computing trust in multiagent systems and social systems have been developed and address issues such as modeling, propagating, inferencing, and using trust in such systems. The talk will be focused on trust in the context of social networks ---social trust. We will look at existing modeling trust formalisms, inference and propagation algorithms, as well as the use of social trust in recommmender systems.