Generative Large Model Validation Mechanism for Blockchain Federated Learning Technology

With the rapid development of artificial intelligence (AI) technology, more and more difficulties have been encountered in solving “data silos” and realizing data value sharing. Due to the security and privacy issues of a large amount of data, it is impossible to conduct large-scale joint training, which also hinders the application of AI technology in various industries. In this paper, we propose a federated training scheme based on federated learning, by introducing the generative large model verification mechanism into federated learning and optimizing the encrypted communication technology, we design a set of efficient FedPT algorithms, which can effectively solve the problems of the existing federated learning technology in practical applications.