Towards a Provenance-Preserving Trust Model in Agent Networks

TLDR: This work proposes a new model in which trust values are derived from a bilattice that preserves valuable trust provenance information including partial trust, partial distrust, ignorance and inconsistency and presents initial results on the first learning step, namely trust propagation through trusted third parties (TTPs).

Abstract: Social networks in which users or agents are connected to other agents and sources by trust relations are an important part of many web applications where information may come from multiple sources. Trust recommendations derived from these social networks are supposed to help agents develop their own opinions about how much they may trust other agents and sources. Despite the recent developments in the area, most of the trust models and metrics proposed so far tend to lose trust-related knowledge. We propose a new model in which trust values are derived from a bilattice that preserves valuable trust provenance information including partial trust, partial distrust, ignorance and inconsistency. We outline the problems that need to be addressed to construct a corresponding trust learning mechanism. We present initial results on the first learning step, namely trust propagation through trusted third parties (TTPs).