Neural Trust Model

Abstract

The problems found in the existing models push the researcher to look for a better solution for computational trust and computational reputation. According the problem exposed earlier, the newly proposed model should be a systematic model which supports both trust and reputation. The model should also take the learning capability for agents into consideration because agents cannot quickly adapt to the changes without learning. The model also needs to have the ability to make decisions according to its recognition of trust. Before actually building the model, it is necessary to analyze the concept of trust. Usually when people say trust they mean human trust, however, in this research trust refers to computational trust. How human trust is different from computational trust is a very interesting question. The answers to the question helped the researcher recover many features of computational trust and built a solid theoretical foundation for the proposed model. The definitions of trust in different disciplines such as economy, sociology and psychology will be compared. A possible definition of computational trust will be made and such trust from several different perspectives will be analyzed. The description of the model is important. As a whole, it is represented as a framework that defines components and component relationships. As the concrete components, the purposes and responsibilities of the specific component are explained. This is to illustrate the static structure of the model. The dynamic structure of the model is described as the process of executing the model.