Managing Trust in Virtual Communities with Markov Logic

This paper describes advances in the development of a trust management system for web-based virtual communities. We review current research on computational trust and report on a proposal for trust management with Markov Logic, a recently developed probabilistic logic language that aims at unifying statistical relational procedures. This work addresses two complex questions: how to learn trust metrics for a group of interacting agents upon amachine learning perspective, and how to use pairwise trust metrics structured as a trust network to compute personalized trust metrics.

1