矩阵奇异值和矩阵范数的保密计算服务协议

Secure computation service protocols for matrix singular value and norm

鉴于目前尚未研究矩阵奇异值和范数的保密计算,提出了矩阵奇异值和范数的保密计算服务协议,将矩阵作变换后,求出矩阵特征值,进而保密地求出矩阵奇异值和范数。通过广泛接受的模拟范例证明了协议的保密性。协议中接受计算服务的一方可用很少的计算资源解决复杂的计算问题,保证较低的计算复杂性和通信复杂性。

Because there are no researches on secure computation of matrix singular value and norm, this paper proposed secure computation service protocols to securely compute the singular values and norm of a private matrix. It transformed the original matrix into another, computed its eigenvalues, singular values and norm. Using the well-accepted simulation paradigm, it proved that the protocols were private. In the protocols, the service receiver can solve complicated computation problems with limited computation source. The protocols have lower computation and communication overheads.