Toward a Practical Method for Introducing and Evaluating Trust Learning Models in Open Multi-agent Systems

In multi-agent systems, agents often interact with each others to achieves their own goals. In open dynamic systems, trust between agents become a critical challenge to make such interactions effective. Many trust models have been proposed to formalise this concept. These models are such good for dealing with trust by proposing components that present a computational form of this concept and a learning strategies to manage it. Components and learning strategies differs from one model to another. This diversity may influence the decision of a user about the best trust model to use in his system. A comparative study is needed to evaluate each trust model and to show the prediction quality of each one. Several testbeds for the evaluation of trust models have been proposed. However, those testbeds are not flexible enough to handle different scenarios in various contexts. In this paper we formulate a practical method based on a framework that introduce the trust concept into open distributed systems, and a testbed that can be used to evaluate trust models in different contexts.

[1]  Chris Cornelis,et al.  Trust networks for recommender systems , 2011 .

[2]  Abdeslam En-Nouaary,et al.  An Abstract Framework for Introducing Computational Trust Models in JADE-Based Multi-Agent Systems , 2015, UNet.

[3]  Sujeet Kumar,et al.  Java Agent Development Framework , 2014 .

[4]  Charles X. Ling,et al.  Constructing New and Better Evaluation Measures for Machine Learning , 2007, IJCAI.

[5]  Jordi Sabater-Mir,et al.  Repage: REPutation and ImAGE Among Limited Autonomous Partners , 2006, J. Artif. Soc. Soc. Simul..

[6]  Nicholas R. Jennings,et al.  A Probabilistic Trust Model for Handling Inaccurate Reputation Sources , 2005, iTrust.

[7]  Paolo Avesani,et al.  Trust-Aware Collaborative Filtering for Recommender Systems , 2004, CoopIS/DOA/ODBASE.

[8]  Audun Jøsang,et al.  AIS Electronic Library (AISeL) , 2017 .

[9]  Cesare Furlanello,et al.  A Comparison of MCC and CEN Error Measures in Multi-Class Prediction , 2010, PloS one.

[10]  Laurent Vercouter,et al.  The Agent Reputation and Trust (ART) Testbed Architecture , 2005, CCIA.

[11]  Robin Cohen,et al.  TREET: the Trust and Reputation Experimentation and Evaluation Testbed , 2010, Electron. Commer. Res..

[12]  Charles Parker,et al.  An Analysis of Performance Measures for Binary Classifiers , 2011, 2011 IEEE 11th International Conference on Data Mining.

[13]  Sriram Chellappan,et al.  Detecting Sybil attacks in vehicular networks , 2014, Journal of Trust Management.

[14]  Costin Badica,et al.  Reputation Model with Forgiveness Factor for Semi-competitive E-Business Agent Societies , 2010, NDT.

[15]  Nicholas R. Jennings,et al.  TRAVOS: Trust and Reputation in the Context of Inaccurate Information Sources , 2006, Autonomous Agents and Multi-Agent Systems.

[16]  Catholijn M. Jonker,et al.  A dynamic perspective on an agent's mental states and interaction with its environment , 2002, AAMAS '02.

[17]  Laurent Vercouter,et al.  A specification of the Agent Reputation and Trust (ART) testbed: experimentation and competition for trust in agent societies , 2005, AAMAS '05.

[18]  Jordi Sabater-Mir,et al.  Decision making matters: A better way to evaluate trust models , 2013, Knowl. Based Syst..

[19]  Jordi Sabater-Mir,et al.  Computational trust and reputation models for open multi-agent systems: a review , 2013, Artificial Intelligence Review.

[20]  Jie Zhang,et al.  Robustness of Trust Models and Combinations for Handling Unfair Ratings , 2012, IFIPTM.

[21]  Peter A. Flach,et al.  A Unified View of Performance Metrics: Translating Threshold Choice into Expected Classification Loss C` Esar Ferri , 2012 .