There has been tremendous interest in the development of formal trust models and metrics through the use of analytics (e.g., Belief Theory and Bayesian models), logics (e.g., Epistemic and Subjective Logic) and other mathematical models. The choice of trust metric will depend on context, circumstance and user requirements and there is no single best metric for use in all circumstances. Where different users require different trust metrics to be employed the trust score calculations should still be based on all available trust evidence. Trust is normally computed using past experiences but, in practice (especially in centralised systems), the validity and accuracy of these experiences are taken for granted. In this paper, we provide a formal framework and practical blockchain-based implementation that allows independent trust providers to implement different trust metrics in a distributed manner while still allowing all trust providers to base their calculations on a common set of trust evidence. Further, our design allows experiences to be provably linked to interactions without the need for a central authority. This leads to the notion of evidence-based trust with provable interactions. Leveraging blockchain allows the trust providers to offer their services in a competitive manner, charging fees while users are provided with payments for recording experiences. Performance details of the blockchain implementation are provided.
翻译:通过使用分析(如信仰理论和巴伊西亚模型)、逻辑(如理论和主观逻辑)和其他数学模型,人们对制定正式的信任模式和衡量标准非常感兴趣。选择信任衡量标准将取决于背景、情况和用户要求,没有一种在所有情况下都可使用的单一最佳衡量标准。如果不同的用户需要不同的信任衡量标准才能使用,则仍然应根据所有现有信任证据来计算信任分数。信任通常使用以往的经验来计算,但在实践中(特别是在集中化系统中),这些经验的有效性和准确性是理所当然的。在本文件中,我们提供了一个正式框架和基于供应链的实际执行,使独立信任提供者能够以分配的方式实施不同的信任衡量标准,同时仍然允许所有信任提供者根据一套共同的信任证据进行计算。此外,我们的设计允许将经验与互动联系起来,而不需要中央当局。这导致以证据为基础的信任概念与可验证的互动(特别是在集中化系统中),这些经验的有效性和准确性是理所当然的。我们提供了一个正式框架和基于供应链的实用性。我们提供了一种正式框架和基于供应链的执行,使信任提供者能够以分散的方式执行不同的方式执行不同的信任指标,同时向用户收取费用。