P1: VLM TAAS0504-14 ACM-TRANSACTION November 10, 2010 18:47 Evidence-Based Trust: A Mathematical Model Geared for Multiagent Systems YONGHONG WANG Carnegie Mellon University and MUNINDAR P. SINGH North Carolina State University An evidence-based account of trust is essential for an appropriate treatment of application-level interactions among autonomous and adaptive parties. Key examples include social networks and service-oriented computing. Existing approaches either ignore evidence or only partially address the twin challenges of mapping evidence to trustworthiness and combining trust reports from imperfectly trusted sources. This article develops a mathematically well-formulated approach that naturally supports discounting and combining evidence-based trust reports. This article understands an agent Alice s trust in an agent Bob in terms of Alice s certainty in her belief that Bob is trustworthy. Unlike previous approaches, this article formulates certainty in terms of evidence based on a statistical measure de ned over a probability distribution of the probability of positive outcomes. This de nition supports important mathematical properties ensuring correct results despite con icting evidence: (1) for a xed amount of evidence, certainty increases as con ict in the evidence decreases and (2) for a xed level of con ict, certainty increases as the amount of evidence increases.
ACM Transactions on Autonomous and Adaptive Systems (TAAS) – Association for Computing Machinery
Published: Nov 1, 2010