An Adaptive Witness Selection Method for Reputation-Based Trust Models
In distributed multi-agent systems where agents need to cooperate with each other in order to thrive, accurately estimating a potential partner’s trustworthiness is vital to an agent’s wellbeing. Many distributed reputation-based agent trust models have been proposed. It has been agreed that testimony sharing is a useful way for agents to gain knowledge about the reputation of potential interaction partners without having to expose themselves to the risk of actually interacting with them. However, the presence of unfair testimonies adversely affects an agent’s long term wellbeing and has been an important problem in agent trust research. Many testimony filtering methods have been proposed, but they often rely on assumptions about the characteristics of the witness agent population or supporting facilities in the environment. In addition, some methods involve highly iterative approaches that consume excessive amount of computational resources.
In this paper, we propose a witness selection method, based on the principles of reinforcement learning, for distributed reputation-based agent trust models. The proposed method innovatively formulates the witness selection problem to allow reinforcement learning technique to be leveraged to solve it without making limiting assumptions on the characteristics of the witness agent population or the existence of supporting facilities. Extensive simulations have shown that it significantly outperforms existing approaches in mitigating the adverse effect of unfair testimonies, especially under conditions where trustee agents and witness agents strategically collude to boost their chances of being selected for interaction by unsuspecting truster agents.
Keywordstrust reputation testimony reinforcement learning
Unable to display preview. Download preview PDF.
- 2.Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction, 342 p. MIT Press (1998)Google Scholar
- 3.Jøsang, A., Ismail, R.: The Beta Reputation System. In: Proceedings of the 15th Bled Conference on Electronic Commerce (2002)Google Scholar
- 4.Sabater-Mir, J., Sierra, C.: Reputation and Social Network Analysis in Multi-agent Systems. In: Proceedings of the 1st International Joint Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), pp. 475–482 (2002)Google Scholar
- 7.Yu, B., Singh, M.P.: Detecting Deception in Reputation Management. In: Proceedings of 2nd International Joint Conference on Autonomous Agents and Multi Agent Systems (AAMAS), pp. 73–80 (2003)Google Scholar
- 9.Whitby, A., Jøsang, A., Indulska, J.: Filtering out Unfair Ratings in Bayesian Reputation Systems. In: Proceedings of the Workshop on Trust in Agent Societies at the 3rd International Joint Conference on Autonomous Agents and Multi-Agent Systems, AAMAS (2004)Google Scholar
- 11.Weng, J., Miao, C., Goh, A.: An Entropy-based Approach to Protecting Rating Systems from Unfair Testimonies. IEICE Transactions 89-D(9), 2502–2511 (2006)Google Scholar
- 13.Fullam, K.K., Barber, K.S.: Dynamically Learning Sources of Trust Information: Experience vs. Reputation. In: Proceedings of 6th International Joint Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), pp. 1055–1060 (2007)Google Scholar
- 15.Marsh, S., Briggs, P.: Examining Trust, Forgiveness and Regret as Computational Concepts. In: Computing with Social Trust. Human-computer Interaction Series, Part I, pp. 9–43 (2009)Google Scholar
- 19.An, B., Lesser, V., Irwin, D., Zink, M.: Automated negotiation with decommitment for dynamic resource allocation in cloud computing. In: Proceedings of the 9th International Joint Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), pp. 981–988 (2010)Google Scholar
- 20.Liu, S., Zhang, J., Miao, C., Theng, Y.-L., Kot, A.C.: iCLUB: An Integrated Clustering-based Approach to Improve the Robustness of Reputation Systems. In: Proceedings of 10th International Joint Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), pp. 1151–1152 (2011)Google Scholar