Combining Cognitive with Computational Trust Reasoning

We propose a concept that combines the cognitive with the computational approaches to experience-based trust reasoning. We emphasize that a cognitive component is vital for computationally modeling trust. At the same time, we recognize the predictive nature of trust. This suggests a combination of the different approaches. The idea is to introduce a cognitive component that produces factual positive and factual negative experiences. These experiences can then be used in a predictive component to estimate the future behavior of an agent. The components are combined in a modular way which allows the replacement of them independently. It further facilitates the integration of already existing trust update algorithms. In this work, we analyze the chain of trust processing for the concept step by step. This results in a concise survey on challenges for experience-based trust models.