Data Rights Verification in the Industrial Internet: A Securing Progressive Scheme with Locked-NFTs and Adaptive Federated Learning

Within the context of the Industrial Internet, data has emerged as an increasingly critical asset in the domains of enterprise production, operation, and management, thereby solidifying its role as a pivotal determinant of competitive advantage for corporate entities. Nonetheless, the intrinsic characteristics of Industrial Internet data, characterized by its diverse origins, substantial volumes, and intricate privacy considerations, have given rise to issues surrounding the unambiguous establishment of data rights and inadvertent disclosure of sensitive information throughout the processes of data propagation and value transmission. To address these multifaceted challenges, this paper introduces a systematic framework for data rights verification, one that incorporates the utilization of Non-Fungible Tokens (NFTs) equipped with a secure locking mechanism in conjunction with adaptive federated learning. This proposed framework bifurcates the data rights verification process into two distinctive phases: data ownership verification and validation of data usage rights. Through the meticulous implementation of smart contracts and the adept utilization of the locking mechanism inherent within NFTs, the framework effectively dissects the intricacies of data rights, thereby ensuring contemporaneousness in data utilization and enabling controlled access. Furthermore, in the phase dedicated to the validation of data usage rights, we introduce an approach firmly rooted in the principles of adaptive federated learning. Leveraging the FedMGDA+ algorithm for distributed storage and computation, this approach successfully validates data usage rights. Empirical findings substantiate the significant advantages encapsulated within this framework, particularly in the domains of data and identity security. Notably, this framework maintains robust precision and scalability throughout the process of data rights verification for distributed model training, thereby exhibiting commendable performance in this intricate domain.