A Survey on Trust Modeling from a Bayesian Perspective

TL;DR

This survey reviews Bayesian trust models in networked systems, highlighting a generic Bayesian trust framework that unifies various models and discusses its potential and limitations for future trust analysis.

Contribution

It provides the first comprehensive review of Bayesian trust models, introducing a generic Bayesian trust framework that unifies existing models and identifies open research questions.

Findings

All surveyed models can be unified under the GBT framework.

The GBT framework has both capabilities and limitations for trust modeling.

Open questions remain for advancing Bayesian trust analysis.

Abstract

In this paper, we are concerned with trust modeling for agents in networked computing systems. As trust is a subjective notion that is invisible, implicit and uncertain in nature, many attempts have been made to model trust with aid of Bayesian probability theory, while the field lacks a global comprehensive analysis for variants of Bayesian trust models. We present a study to fill in this gap by giving a comprehensive review of the literature. A generic Bayesian trust (GBT) modeling perspective is highlighted here. It is shown that all models under survey can cast into a GBT based computing paradigm as special cases. We discuss both capabilities and limitations of the GBT perspective and point out open questions to answer, with a hope to advance GBT to become a pragmatic infrastructure for analyzing intrinsic relationships among variants of trust models and developing novel tools for trust evaluation.