Abstract
This paper describes a probabilistic model of trust and reputation that learns to distinguish reliable sources from unreliable ones. It is demonstrated to guard against the effects of inaccurate opinions, even when 50% of sources are intentionally misleading, and to outperform the most similar models. The work has served as the foundation for two PhD theses (J. Patel and W. T. L. Teacy) and is one of the component technologies used in the entry that won the First Agent Reputation and Trust (ART) Competition at the Fifth International Conference on Autonomous Agents and Multiagent Sytems in 2006.
Original language | English |
---|---|
Pages (from-to) | 183 - 198 |
Number of pages | 16 |
Journal | Autonomous Agents and Multi-Agent Systems |
Volume | 12 |
Issue number | 2 |
DOIs | |
Publication status | Published - Mar 2006 |