A Trustworthy Service Transaction Framework for Privacy Protection

Servitization is one of the important trends in reshaping the information world in recent years. With the development of division of labor in today’s service-oriented society, all parties involved need to gather data and collaborate on training. However, it is difficult to gather data from all parties involved, and each party in the transaction is fighting on its own, forming a complex digital service network. In this network, trading parties need to collaborate with multiple parties while engaging in multi-party games. The key issue faced by this complex service network is how to achieve coordination of overall interests, that is, to achieve multi-party cooperation among all parties involved in the entire transaction process, and to accurately trace problems. Therefore, this paper proposes a trustworthy service transaction framework for privacy protection. To address the differences in service content openness, degree, and standards among different service providers in digital service networks, a service sharing model training system based on federated learning is constructed. By combining deep neural network algorithms and large language models, service recommendation and risk assessment can be implemented to safeguard and regulate service transaction behavior while ensuring data and model privacy. Distributed verification of data and service chains in the service transaction process is carried out through blockchain technology for various transaction records stored in multiple service entities and service terminals. A case study on credit services in a large state-owned bank is given to demonstrate the application of the framework.