Computing Trust in Social Networks

Web-based social networks like mySpace and Facebook are hugely popular,
and provide an open, public source of social information that can be used
in a wide range of applications. Understanding relationships between users
can lead to intelligent systems that take advantage of this social context
to improve performance. One of the most useful relationships for sorting
and filtering information is trust. In this talk, I will present several
methods for understanding and computing trust in social networks.

Specifically, I will present an algorithm for inferring trust between two
individuals who are not directly connected in the network. Then, I will
present results of a recent study that compares the impact that changes in
trust values have on inferred values in the network.