Research On Social Credit Information Sharing Based On Blockchain And Federated Learning

As one of the important components of social credit system construction,social credit information has the requirements of completeness,accuracy,efficiency and security.With the expansion of the total amount of credit information,the increase of asset value,and the increase of privacy and security requirements,the difficulty of information sharing is gradually increasing.In the sharing process,problems such as centralized bottleneck,single decision,artificial efficiency and visual security appear,leading to low efficiency of credit information sharing,and information privacy cannot be protected.Therefore,it is of great significance to balance the power of the leading unit,clarify the authority of the member units,unify the credit information standard and control the plaintext information transmission to maintain the control right of credit information,locate the rights and responsibilities of the department and promote the timely and secret sharing of credit information.Firstly,through literature analysis,this paper makes an in-depth analysis of academic literature,laws and regulations,policy documents and standard catalogs,differentiates privacy protection and other concepts,and clarifies the essential goal that needs to be protected in the process of credit information sharing,namely the privacy protection of "electronic data".Accurately grasp the current situation and process of social credit information sharing,find out that the sharing problem is caused by the concentration of power,unclear authority,different standards,and uncontrollable transmission,and analyze and conclude that the distributed architecture of blockchain and the collaborative architecture of federated learning are feasible in solving the central bottleneck and single decision problem.The features of intelligent contract and local computing technology can also effectively solve the labor efficiency and visual security of credit information content.Secondly,it analyzes the technical characteristics of blockchain and federated learning through comparative analysis to explore the advantages and disadvantages of shared technological means.Finally,the internal demands of credit information sharing system and privacy protection system are analyzed through system analysis,and the social credit information blockchain data sharing model and the social credit information federal learning privacy protection model are constructed.In the sharing model,under the premise of not weakening the power of the central node,the weak center of the alliance chain is used to transform its function in the sharing,so as to solve the bottleneck problem of the center to some extent.Intelligent sharing of credit information can be realized by intelligent contract,and problems such as efficiency,accuracy and integrity caused by traditional manual transmission and sharing can be avoided.In the sharing process,IPFS storage is introduced to meet the need of mass storage of credit data.CP-ABE is used to control the access to credit data and ensure data security.In the privacy protection model,vertical federation learning is used to achieve a better fit between the technical framework and the credit information sharing system,so that all departments participate in the sharing process and avoid their own decision-making problems.Federal deep learning algorithm is used to ensure that the original credit data does not come out of the department in the training process,and at the same time to deal with the security problem of sharing gradient parameters,further protect the privacy of credit data,and solve the privacy security threat caused by data content visualization.The research shows that by integrating blockchain technology and federal learning technology into the process of social credit information sharing,credit information can be shared in the way of de-identification and anonymity,and at the same time,the privacy and security of credit information can be taken into account to meet the requirements of regulations and meet the development needs of social credit information sharing.