基于联邦学习的分布式数据安全融合方法研究

Research on Distributed Data Security Fusion Method Based on Federated Learning

为满足在可用性、实用性的前提下实现大数据安全融合共享,设计了基于联邦学习的分布式数据安全融合模型,并运用差分隐私、安全多方计算、同态加密、函数加密四种方法进行实现,以保证分布式数据融合的安全性和可行性。

In order to achieve big data security sharing under the premise of availability and practicability,the distributed data security fusion based on federated learning is studied.It mainly uses four methods,namely differential privacy,secure multi-party computing,homomorphic encryption and functional encryption,to encrypt and ensure the security of data.